_{THE LARGEST THINGS YOU’VE NEVER HEARD OF}

Size Does Matter

Last week I went to Costco because I decided I needed my groceries measured in pallets rather than pounds for the next few weeks, and as I was walking around looking at all the members, it got me thinking about the revenue split between Costco’ membership and the actual in store business. While I wish I could tell you that I had a lively debate with the employees of Costco about their financial position, I did not. I did, however, decide to check their filings because I clearly need a hobby.

What I learned was fascinating, and it got me thinking about what other random parts of companies or pockets of capital are deceptively large. I mean in some capacity, size does matter after all. With that in mind, here are some random corners of the market that I found were much larger than any one of us gave them credit for. Except maybe their coverage analysts, but if you think I need to find a hobby go check in on your research friends now that Q3 just ended.

1. The Marlboro brand is worth more than Ford and GM combined | Value: $33–38 billion. Not Philip Morris as a company… just the trademark. Brand Finance estimates the Marlboro brand as a standalone asset at approximately $33–38 billion. Ford's market cap is roughly $45 billion and General Motors is around $50 billion. The two companies that built the American automobile industry, employ hundreds of thousands of workers, and have collectively produced over a billion vehicles are each worth roughly what one tobacco logo would fetch in an auction.

2. The federal government collects more in mineral royalties than American Airlines makes in revenue | Value: $17.4 billion a year. The Office of Natural Resources Revenue reported $17.4 billion in royalties, rents, and bonuses from oil, gas, and mineral production on federal land and offshore in fiscal year 2023. That is passive income, collected simply for owning the ground. American Airlines does about $13 billion in annual revenue and has to operate thousands of flights a day to get there. The federal government beats it by sitting still. This number appears nowhere in the public debate about the federal deficit.

3. Harvard's endowment earns more per year than the GDP of 50 countries | Value: ~$4.5 billion per year. In a good year it earns 9–10%, which is $4–5 billion in investment income. That is more than the entire annual economic output of Iceland, Kosovo, or Belize. A university in Cambridge, Massachusetts generates enough annual investment income to fund a mid-sized government's entire budget because apparently that’s what it takes to create people who turn going to your institution into their entire personality.

4. Ontario sold a highway for $3 billion that now generates $1.6 billion a year. | Value: ~$25–30 billion today. In 1999, the Ontario government sold a 99-year lease on the 407 ETR toll road across the top of Toronto for C$3.1 billion. The road now generates approximately $1.6 billion CAD per year in toll revenue. At current infrastructure multiples it is worth roughly C$25–30 billion, nearly 10x what Ontario received. Ontario sold a perpetual toll road at under 2x annual revenue and called it a good deal, and the buyer is writing up the best deal in infrastructure history. I am sure someone with deeper knowledge of Canadian history can tell me why this was a good deal at the time, but to play armchair forward (that’s hockey reference), the Ontario government should consider hiring Goldman instead of BMO next time they do a deal.

5. McDonald's rent and royalties generate more revenue than Chipotle | Value: $14.9 billion a year. If you’ve watched The Founder, great movie, you remember Ryan from The Office told Ray Kroc “You don’t build an empire on a 1.4% cut of a 15 cent hamburger. You build it by owning the land on which the hamburger is cooked”. And well, he cooked with that one. McDonald's collects $14.9 billion per year from franchisees in rent and royalties. Not from selling food, from owning the land under the restaurants and charging operators roughly 10% of gross sales. Chipotle - which McDonald’s actually owned at one point - generates about $11 billion in total annual revenue. Darden Restaurants, which runs Olive Garden and LongHorn Steakhouse combined, has a market cap of $18 billion, comparable to what McDonald's pulls in just as a landlord. Maybe that MBA really did pay off for Ryan. Wonder if he went to Harvard…oh wait, we’d know.

6. Costco's membership fee business alone is worth more than UBS | Value: $177 billion. Costco did $297 billion in sales last fiscal year, but the retail side runs on razor-thin margins: the $4.99 rotisserie chicken and $1.50 hot dog are famous loss leaders. The actual business is the annual membership fee: $5.9 billion in revenue at nearly 100% margin, and ~half of Costco's operating income. At a 30x multiple on that recurring stream, the fee business alone would be worth ~$177 billion, more than UBS's entire market value of ~$147 billion. Costco built a more profitable business than UBS and hid it inside a warehouse store, and doesn’t have staff working on 4 hours of sleep aligning footnotes.

7. The U.S. government is sitting on more oil than Harley-Davidson, Nordstrom, and Macy's are worth combined | Value: $24 billion. Okay, this one maybe isn’t a total shock, but I still found it fascinating. The Strategic Petroleum Reserve holds approximately 347 million barrels of crude oil stored in caverns along the Gulf Coast. At current prices that is roughly $24 billion in market value, exceeding the combined market caps of Harley-Davidson, Nordstrom, and Macy's. The caverns cost almost nothing to maintain. It gets mentioned in the news approximately once a year when someone proposes releasing some of it to bring gas prices down before an election. This excludes Marco Rubio’s personal collection of oil as President of Venezuela. More shocking than the value though is that the US consumes ~20 million barrels of oil per day, so the entire reserve will last only 14 days. Less than your Netflix or Spotify free trial.

8. Starbucks holds more in customer deposits than most U.S. banks | Value: $1.9 billion at zero percent interest. One of my accounting professors in college used to say that Apple is a mutual fund that sells iPhones. Well, he clearly hadn’t looked into Starbucks much. Starbucks reports having ~$1.9 billion sitting on its balance sheet in gift cards and app balances customers loaded and haven't spent. That is an interest-free loan from 35 million people, larger than the total deposits of hundreds of U.S. community banks. Starbucks collected it with no banking license, no regulatory oversight, and no interest payments. Starbucks is really a retail bank who also happens to sell overpriced lattes.

_{PRESENTED BY TOLTIQ}

Built for the Data Room, Not Adapted to It

Diligence is a compilation of documents that were never meant to be read together. There are thousands of files with no consistent format, and most data rooms are not built to handle the analysis of  a deal as it evolves. Most AI isn’t built to handle it either.

ToltIQ, however, was designed for private markets investment professionals. PE and M&A deal teams can upload the full VDR and work it as one record, with traceable links to its source document so verification is simple. Here are some of their premier features:

• Bulk Query pulls consistent, side-by-side answers from hundreds of documents in a single run. 

• Workflows turn your firm's diligence sequence into repeatable processes. 

• Blueprints convert your firm’s document templates into reusable formats. 

• Vaults keep research, templates, and market data available for every deal.

Their architecture is built to avoid redundant processing across large document sets, which keeps AI costs forecastable as usage scales. 

See these features in action; schedule a demo with their team today.

_{THE WORST CALLS EVER MADE}

Be Like A Goldfish

“One thing I can promise you, even in this market, is that I never ask my clients to judge me on my winners. I ask them to judge me on my losers, because I have so few.” - Leonardo DiCaprio as Jordan Belfort in the Wolf of Wall Street

When you go into PE, you hear that you are only as good as your last deal, and Wall Street can have a long memory at times. But sometimes, it helps to be a goldfish (short memory).

Here are some of the most expensive public calls ever made, by people who absolutely should have known better, and who you will still find quoted as experts in your morning briefing.

This entire list could have been Jim Cramer calls, but that felt like bringing a flamethrower to a knife fight, so we gave him two entries and moved on. You're welcome, Jim. Contact my people if you want me on Mad Money.

Jim Cramer told viewers Bear Stearns was fine. It collapsed six days later. March 11, 2008: "Bear Stearns is fine. Do not take your money out of them. Don't move your money from Bear. That's just silly. Don't be silly." JPMorgan bought it six days later for $2 a share, down from $170. He was technically right… it was fine, right up until it wasn't.

Jim Cramer called Silicon Valley Bank undervalued. It failed 30 days later. February 8, 2023, Cramer called SVB "a merchant bank with a deposit base that Wall Street has been mistakenly concerned about" and said it was oversold at $320. The FDIC shut it down March 10 in the second-largest bank failure in U.S. history. The Inverse Cramer ETF, which was a real product that bet against his picks, had launched the prior year. It did well.

Warren Buffett passed on Google while his own company was buying ads on it GEICO was paying Google $10–11 per click for car insurance leads. Buffett had the unit economics in his own P&L and still passed on the 2004 IPO. At the 2017 Berkshire annual meeting: "I was wrong on Google." 

Warren Buffett also passed on Amazon. Twice. 1994 and again at the 1997 IPO. At the same 2017 meeting: "I was too dumb to realize what was going to happen." Berkshire finally bought in 2019, at roughly 500x the price it passed on. To his credit, he eventually bought it. He just waited until it was worth more than the GDP of most countries first.

Warren Buffett paid $433 million in Berkshire stock for a shoe company that went to zero. Dexter Shoe, 1993. He paid in stock, not cash, which is the part that really stings. The business got wiped out by foreign competition. In his 2007 shareholder letter: "I gave away 1.6% of a wonderful business to buy a worthless business." 

Irving Fisher called a permanently high plateau in stocks two weeks before the crash. October 16, 1929, New York Times: "Stock prices have reached what looks like a permanently high plateau." The Dow fell 89% over the next three years and didn't recover until 1954. He was considered one of the most brilliant economists alive at the time, which tells you everything you need to know about economic forecasting.

Steve Ballmer said the iPhone had no chance of getting significant market share. USA Today, April 30, 2007: "There's no chance that the iPhone is going to get any significant market share. No chance." The iPhone generated over $1 trillion in cumulative revenue. Windows Mobile is dead. To Ballmer's credit, he said it with tremendous conviction, which is really the most important thing when you're wrong.

Hedge funds shorted Volkswagen and lost $30 billion in two days. In October 2008 shorts had roughly 12% of VW's float. Then Porsche disclosed it had quietly accumulated 74.1% of VW through cash-settled options, with Lower Saxony holding another 20%, leaving almost nothing available to borrow. VW went up 376% in two trading days, briefly making it the largest company on earth by market cap. Porsche made more money on that options trade than it had made selling cars in its entire history.

Jim Chanos shorted Tesla and watched it go up 743% in a single year.
Chanos called Enron. He started shorting Tesla around 2016, calling it a "walking insolvency." In 2020 Tesla rose 743% and short sellers collectively lost roughly $40 billion, the largest annual short loss on a single stock in history. Chanos was not wrong about the fundamentals. He was just wrong about whether that matters.

Sequoia Fund put 36% of its portfolio in one pharmaceutical company and watched it collapse. Sequoia was founded with Warren Buffett's personal endorsement. By 2015, it had 36% of the fund in Valeant Pharmaceuticals. When Citron published a fraud report, Valeant fell 90% from peak and Sequoia lost nearly half its portfolio value. Multiple board members resigned while management refused to cut the position. A fund with 65 years of Buffett's blessing put more than a third of it in one trade. Concentration risk is a hell of a drug. 

_{WARREN’S FINAL LETTER}

Father Time Comes For Us All

I had an article written that was honestly in pretty good shape, but then the greatest stock picker of our lifetime decided that he was going to drop his resignation letter, so here I am drafting something entirely new. It’s a tough life I live keeping everyone updated on the finance world.

If you didn’t already know, Warren Buffett just released a letter resigning as chairman of Berkshire Hathaway. It was a short, simple letter.

“Serving as your Chairman has been the privilege of a lifetime, and I have never taken your trust for granted. Father Time always wins. He has, however, been generous with me."

That's it. Just two sentences putting a period on the end of an incredible career.

Buffett took control of Berkshire in 1965. It was a failing textile mill worth about $18 million. He has since described the purchase as the worst investment he ever made.

He was so offended by the way the previous owner handled a price negotiation that he bought the company out of spite. The $18 million is now $1.1 trillion, so the spite was productive.

The idea now is conceptually obvious: use insurance float (premiums collected before claims are paid) to fund long-term investments in durable businesses.

You can buy shares in Coca-Cola, American Express, and Apple or you could buy BNSF Railway and GEICO outright.

The two things to remember? Hold everything forever and never try to be clever. The formula compounded at roughly 19.8% per year for 60 years versus 10.2% for the S&P 500.

That gap sounds small, but over 60 years it is the difference between $1 and $40,000, and he invested a bit more than $1.

He did all of this while eating McDonald's for breakfast, drinking Coke for reasons that are probably not unrelated to owning 9% of the company, and still living in the house he bought in 1958 for $31,500.

As we look back on his career, here are ten things he said that every investor should take to heart. Well, many more than ten, but that's all the column space we have today.

• "Price is what you pay. Value is what you get." (2008 shareholder letter) The market's price is always right as a reflection of sentiment. It is often wrong as a reflection of what something is worth in ten years. Buffett spent 60 years betting on the difference.

• "Be fearful when others are greedy and greedy when others are fearful." (1986 Berkshire shareholder letter at the height of a roaring bull market) Seven months later, Black Monday wiped 22% off the Dow in a single session and Buffett would be able to turn this quote into reality.

• "It's far better to buy a wonderful company at a fair price than a fair company at a wonderful price." (1989 shareholder letter) The lesson he drew from buying Berkshire out of spite in the first place. He got a wonderful price on a mediocre company and spent decades explaining why that was the wrong trade.

• "The stock market is a device for transferring money from the impatient to the patient." This is the entire Berkshire model in one sentence. While Keynes reminds us that the market is not always rational, if you look with a long enough time horizon, it is. Keep a longterm view if you can. If you’re trading quarters at Point72, I have nothing for you.

• "Someone is sitting in the shade today because someone planted a tree a long time ago." A quote Buffett has used across many shareholder meetings over the years. Berkshire is the tree, and the $365.5 billion Greg Abel inherited is the shade…maybe more like night at that point. 

• "Risk comes from not knowing what you're doing." This is one of Buffett's most repeated convictions across annual meetings spanning four decades. Risk is not from volatility, not from beta, not from a VaR model, but from ignorance. The real world exists outside of your spreadsheets. You won’t be able to control every variable, but you can control how prepared you are to respond to any situation.

• "Our favorite holding period is forever." (1988 Berkshire shareholder letter) This is not a strategy, but a philosophy. Of course you have to know when to cut your losses, but if you buy the right thing in the first place, you might just be the Oracle of Omaha then.

• "Forecasts may tell you a great deal about the forecaster; they tell you nothing about the future." (1980 Berkshire shareholder letter) When was the last time you had a model actually turn out to be correct? How many times do you look to justify assumptions to yourself because they make the returns work? The real world rarely matches your models, all your models tell you are what you needed to convince yourself at the time.

• "In the business world, the rearview mirror is always clearer than the windshield." (1991 Berkshire shareholder letter) Written while Buffett was temporarily running Salomon Brothers in the middle of a scandal. Sure, hindsight is always 20/20, but while the disclaimer says “past performance doesn’t guarantee future results” sometimes looking at the past can reveal details that are easily missed when you are just looking at the long road ahead.

• "Only when the tide goes out do you discover who's been swimming naked." (1992 Berkshire shareholder letter) Written about the aftermath of the 1980s credit boom. With the 10-year at 5% and leverage costs rising across every asset class simultaneously, the tide is going out right now. We will find out shortly who forgot their swimsuit.

Greg Abel is now running the company with $365.5 billion in cash, the largest cash position ever held by a single company. Berkshire stock is up 1% in 2026 versus 11% for the S&P 500. Abel has started buying back stock, $4.5 billion in Q2. It is a start. The first major deployment he makes will define how the market thinks about the post-Buffett era. Buffett could overpay slightly and be forgiven because of the float compounding underneath. Abel does not have that latitude yet.

This month also happened to be when the Fed raised rates for the first time in three years, the 10-year crossed 5%, and oil broke $100. The era of cheap capital that made patient compounding so powerful is definitively over. The formula still works. Buy good businesses at fair prices, hold them forever, don't do anything clever. The conditions are just harder now.

"Father Time always wins. He has, however, been generous with me."

He was, and from all of us here at Overheard on Wall Street - we wish you many years as Chairman Emeritus at Berkshire, Mr. Buffett, and we are cracking open a coke for you in the bullpen.

_{PRESENTED BY OPENSTOCKS}

AI Agents Are Now Competing with Wall Street on Earnings Forecasts

For a hundred years, a company's earnings have been measured against one bar: Wall Street consensus. 

Consensus has always been a human product that carries human problems. Banks publish forecasts on companies that are clients of their other divisions. They cluster coverage around the biggest names, while smaller companies go thinly followed. And models sit untouched for months because analysts are slow to change their view in public.

That consensus has a new competitor; one forecast entirely by AI agents.

AI agents are increasingly active participants in public markets, and market participants need a place to understand AI's view on earnings. OpenStocks is that place, with one rule: no human analysts. Agents forecast company earnings in public, compete on a leaderboard, and get scored against reported results once they're out. Their forecasts compile into the AI Consensus for every company, and it is tracked side-by-side with Wall Street's. Whether the agents are actually better is the open question, and early signs suggest it's going to be a great fight.

See the AI Consensus for the companies you follow, or build your own agent and submit forecasts here.

_{WHAT WE ARE WATCHING}

How Will it End?

This year has been quite the ride, but the S&P is still up 12% YTD even with rates also up and the AI trade waffling. With that in mind, I have had a few people ask what I think will happen in Q4. Honestly - I have no idea. If I did, I would be making millions on my Substack selling trading advice like Michael Burry, not giving it to you in a free newsletter. That being said, I like to think I am at least more correct than say, Jim Cramer, so here are a few things I am watching going into the end of the year.

The IPO window. Anthropic filed confidential IPO paperwork in early July. OpenAI is targeting a September/October filing. Both are expected to list before year-end. SpaceX listed in June and has since fallen below its $150 opening price. If Anthropic and OpenAI price well and trade well after listing, the IPO market opens broadly for companies that have been waiting since 2022 and the PE exit backlog starts to clear. If they don't, the window closes again and the dynamic gets complicated.

The Fed's next move. One more hike is roughly 60% priced. If inflation stays sticky, which oil above $90 makes more likely, the Fed moves again. If the Iran situation resolves and oil falls materially, the calculus changes. The market's reaction to the next inflation print will tell you a lot about how Q4 trades. Currently, prediction markets are pricing in another hike so if you are in the market for debt, I suggest looking to rate lock sooner than later. Though, it could be a risky move locking in that mortgage right now.

The election. Midterms are in November. Markets generally do not love uncertainty around elections and generally do fine regardless of the outcome. The specific Q4 variable is whether election uncertainty causes a pullback in deal activity in October as boards wait for clarity. In 2022 it did. In 2018 it did not. While it’s still early, most people are expecting a blue wave in the midterms to divide government and reduce Republican control in Washington. It’s unclear what that will mean for the market, but I am hopeful that whoever gets elected has Nancy Pelosi running their portfolio so we can piggyback some gains.

Credit. Investment-grade credit spreads are tight. High-yield spreads are tighter than the rate environment suggests they should be. Private credit is doing an enormous amount of work absorbing deal flow that the syndicated loan market used to handle. At some point, tighter financing conditions work their way through to deal multiples and deal count. Whether that point is Q4 2026 or sometime in 2027 is the most important question in PE right now.

Earnings. Q3 earnings season starts in mid-October. The AI capex story will be tested again. The hyperscalers are spending $725 billion combined this year and investors want to see it in revenue. If AWS, Azure, and Google Cloud all report strong acceleration, the market extends. If any of them disappoint on cloud revenue while flagging higher capex, the AI trade gets a serious re-rating conversation.

An ode to the Oracle and a look into the future, a perfect way to walk into fall.

_BACKGROUND

The Smartest Guys in the Room

In this edition, we are doing one more legacy finance story as we come off of the Labor Day holiday here in the states. Honestly, I was expecting the Anthropic S-1 to drop by the time this issue came out, but no dice yet so we are taking it back to Greenwich, CT.

In 1994, John Meriwether, the former head of bond arbitrage at Salomon Brothers, founded a hedge fund in Greenwich, Connecticut staffed by two Nobel laureates, a former Federal Reserve vice chairman, and the sharpest quantitative minds Wall Street had ever produced.

The fund was - ironically - called Long-Term Capital Management.

For four years, it was the most successful hedge fund in history. Annualized returns of 21% in year one, 43% in year two, 41% in year three. (Net of fees)

The minimum investment was $10 million and there was a three-year lockup. People lined up down the harbor to get in.

In August and September of 1998, the fund lost $4.6 billion in less than five months.

Its equity fell from $4.7 billion to $600 million. Its notional exposure at peak was $1.25 trillion, roughly 5% of global GDP. The Federal Reserve Bank of New York convened an emergency meeting of fourteen Wall Street institutions on a Sunday afternoon and told them they needed to put in $3.6 billion by the end of the day or global credit markets would seize up on Monday morning. They put up the money.

So what went wrong? The trade, the timing, or the leverage? Keep reading.

_{PRESENTED BY TOLTIQ}

Uber Burned Through Its Full-Year AI Coding Budget by April

Uber exhausted its full-year 2026 AI coding-tool budget by April, with its COO saying it is "very hard to draw a line" from AI usage to consumer-feature output. ServiceNow drew down its annual AI coding budget early over the same stretch. And, this is becoming a common theme.

79% of finance leaders reported AI cost overruns in the past twelve months. The major vendors, like OpenAI and Anthropic, moved enterprise AI to metered billing this spring. The default enterprise pricing model is now variable with usage, not headcount.

Optimal model architecture has become vital for usage scalability across active deals, and ToltIQ is built for it. Their research team continually tests frontier models to optimize the platform, but more importantly, their system is designed to retrieve information efficiently to help PE firms (and their PortCos) avoid unanticipated costs.

Some of the most respected firms use ToltIQ for their diligence. Will yours?

Schedule a demo with them today.

_{THE PEOPLE}

The Top Minds on Wall Street

Listen, the book is called "When Genius Failed" for a reason. The people behind LTCM were the smartest guys in any room, at least if you're counting book smarts. The partner roster was an all-star list, and sometimes too many geniuses in one room doesn't play out the way you'd expect.

John Meriwether had built Salomon Brothers' bond arbitrage desk into the most profitable trading operation on Wall Street in the 1980s. He left under a cloud in 1991 after a Treasury auction scandal involving one of his traders. LTCM was his comeback.

Myron Scholes and Robert Merton shared the 1997 Nobel Prize in Economics for developing the Black-Scholes options pricing model, the foundational framework for pricing derivatives that the entire industry still runs on. They were not figureheads. They were active partners who contributed directly to the fund's trading strategies.

David Mullins was the former vice chairman of the Federal Reserve. He had the relationships, the regulatory knowledge, and the institutional credibility that let LTCM borrow from counterparties on terms no other fund could access.

The reputation of the partners was a trading asset in itself. LTCM could borrow at near-sovereign terms because no bank believed this particular group of people could be wrong enough to default. That belief was the foundation of the leverage. When it broke, everything broke with it.

_{THE TRADE}

When it Rains, it Pours

LTCM's core strategy was convergence arbitrage. The basic idea is straightforward: if two similar instruments are mispriced relative to each other, buy the cheap one, short the expensive one, and wait for the spread to close. The trade makes money when the world returns to normal. It loses money when the world moves further away from normal.

The specific trades took many forms. The one that defined LTCM was the on-the-run versus off-the-run Treasury spread. A freshly issued 30-year Treasury bond ("on the run") trades at a small premium to a Treasury issued six months earlier with the same maturity ("off the run") because the newer bond is more liquid.

The premium is irrational. Both bonds are backed by the same government and mature to the same value. LTCM bought the off-the-run bond, shorted the on-the-run, and waited for the liquidity premium to compress.

The trade worked consistently. The problem was that the spread was small, just a few basis points, so the returns were also small unless you applied enormous leverage…yes, enormous.

LTCM's balance sheet leverage ran at approximately 25:1 on disclosed positions and significantly higher on a risk-adjusted basis. For every dollar of equity, the fund controlled roughly $25 of assets. On notional derivatives exposure the ratio was north of 250:1.

By 1997 the fund had grown to $7 billion in equity. The partners concluded the fund was too large relative to the opportunities available and returned $2.7 billion to outside investors. This allowed them to keep the capital they had invested themselves, reducing the outside equity base while maintaining the full portfolio size.

The leverage ratio rose from approximately 25:1 to 28:1 as a result.

They made the fund riskier by giving investors their money back. The portfolio did not change, but the cushion did.

_{THE RUSSIAN DEFAULT}

The Straw That Broke the Camel’s Back

On August 17, 1998, the Russian government defaulted on its domestic ruble-denominated debt and devalued the ruble.

Russia itself was a small part of the portfolio and in isolation would not have taken LTCM down. The fallout reached Chernobyl-levels quickly, melting down the LTCM portfolio in a manner of months.

When Russia defaulted, global investors panicked and fled to safety simultaneously. Every player in every market made the same decision at the same time: sell risky, illiquid, complex positions and buy safe, liquid, simple ones. The flight to quality was not a gentle drift, it was a stampede.

Every spread LTCM had bet would converge instead diverged, dramatically and simultaneously. 

The off-the-run Treasury spread blew out instead of compressing. European swap spreads blew out. Mortgage-backed security spreads blew out. Emerging market spreads blew out. Equity volatility positions that should have been uncorrelated with credit moved in lockstep.

LTCM's risk models were built on historical correlations between markets. Those correlations assumed that different asset classes moved somewhat independently. In a crisis, they do not. In a crisis, everything correlates to one.

The diversification that the model showed as risk reduction vanished precisely when it was needed most.

The fund lost $1.9 billion in the last two weeks of August alone. By late August its equity was under $2.5 billion. By September 22, it was $600 million against a portfolio that had barely shrunk. The leverage ratio was now in excess of 100:1.

_{THE BAILOUT}

A Sunday Morning at the Fed

On September 23, 1998, William McDonough, the President of the Federal Reserve Bank of New York, convened a meeting at 33 Liberty Street in lower Manhattan.

In the room were the senior executives of fourteen Wall Street institutions: Goldman Sachs, Merrill Lynch, JPMorgan, Morgan Stanley, Lehman Brothers (RIP), Bear Stearns (RIP), Deutsche Bank (a bulge bracket at the time), Société Générale, and others.

McDonough's message was not complicated. LTCM had approximately $1.25 trillion in notional derivatives exposure across virtually every major financial market in the world.

If LTCM defaulted, its counterparties would be forced to immediately close out those positions. Every major bank in the room was a counterparty. 

The resulting fire sale would hit markets already in crisis. Spreads that had blown out in August would blow out further. Liquidity would disappear. The cascade could not be modeled with confidence.

The Fed was not going to use taxpayer money. It was explicit about that, but McDonough made clear that the alternative (doing nothing) was not obviously better for the institutions in the room.

Fourteen firms put in ~$3.6 billion for a 90% stake in the fund. Bear Stearns, LTCM's clearing broker, refused to contribute a dollar, a decision the street remembered ten years later when Bear needed help of its own.

The existing partners kept 10%. Meriwether kept his job managing the wind-down. The positions were unwound over the following months at prices that recovered most of the capital.

That same morning, Goldman Sachs, AIG and Warren Buffett had offered to buy the partners out for $250 million, essentially nothing, inject ~$3.75 billion of new capital, and take over the fund entirely. The offer carried a fuse of about an hour. Meriwether questioned whether he had the authority to accept it. By the end of the day he had a worse outcome and no choice.

That $250 million offer would have been the best deal of John Meriwether's life.

_{THE POST MORTEM}

What Went Wrong

The models assumed normal correlations. LTCM's framework was built on the assumption that different markets move with historical degrees of independence. In August 1998, every market moved in the same direction simultaneously. Correlation went to one. The diversification the model showed as safety disappeared at exactly the moment it was needed. LTCM's models had assigned a 1-in-6.4-trillion-year probability to losing 40% in a month. They lost 44%.

The leverage left no room for being early. The trades were correct. Spreads did eventually converge. The fund ran out of capital before they did. At 28:1, a 4% adverse move eliminates all equity. Being right on a six-month delay is indistinguishable from being wrong.

They returned $2.7 billion of outside capital at the worst possible moment. They kept the positions. They reduced the cushion. The effect was to concentrate risk in the partners' own money at precisely the moment that risk was about to be tested. In late 1997, with the Asian crisis building and early credit stress visible, LTCM's smartest people in the room concluded the fund was too large and gave investors their money back. It was the most expensive act of modesty in financial history.

Counterparty risk became a real concept. Before LTCM, banks did not think systematically about simultaneous default across hundreds of derivatives positions with a single counterparty. After LTCM, they did. For a while.

The Fed set a precedent it could not unset. No public funds were used in 1998, but the principle that the Fed would intervene to prevent the disorderly collapse of a systemically important private institution was established explicitly. It was cited during Bear Stearns, AIG, and the 2008 banking system interventions. The LTCM Sunday was the rehearsal; 2008 was the show.

Nobody changed their models. The lesson was written up in academic papers, Congressional testimony, and the President's Working Group on Financial Markets report in 1999. Ten years later, the same failure mode at ten times the scale produced the 2008 crisis. The models continued to assume diversification would protect institutions in exactly the conditions under which diversification stops working.

RJR Nabisco failed because leverage amplified a macro outcome nobody fully modeled. AOL-Time Warner failed because a bubble currency bought real assets at the peak. LTCM failed because the model was excellent and the market was not the model. The trades were right, but the timing went wrong and the leverage left no room for the gap between those two things.

The LTCM partners were not stupid. Obviously, they were the smartest people in the room. 

That was part of the problem. When you are certain enough in your model to borrow 28 dollars for every dollar of equity, you have stopped asking what happens when the model is wrong. 

The market is not interested in how good your model is. It will take your equity while you wait to be proven right.

If you want to learn more, check out When Genius Failed by Roger Lowenstein. If you work in credit, maybe read it twice.

_{2026 BLIND DATES}

Who is Swiping Right on Who?

Last year around this time I did an M&A wishlist that turned out to have a pretty good call on it with WBD x Netflix. This year, I am doing something different by focusing on firms that have never publicly disclosed talks (except for one potential option) and seeing if these firms would be a good match. Once the summer slowdown is over, let’s see if any of these predictions can play out to be true going into year end.

Apple x Whoop

Apple Watch is a great product for people who want to know how many steps they took and hate any other stats because the battery will die in 15 minutes. Whoop is a great product for people who actually train.

The Whoop 5.0 tracks strain, recovery, sleep quality, HRV, and blood oxygen with a level of physiological depth that Apple Watch does not come close to, has a subscription model that generates recurring revenue, and has built a devoted following among serious athletes and military personnel who wear it as their primary health wearable. There are no buttons and there is no screen. It does one thing and it does it better than anything else on the market.

Apple has spent a decade trying to make Apple Watch a serious health device, and it keeps running into the same problem: it is also a notification machine and a payment terminal, which means it will never be trusted as a medical-grade sensor by anyone who cares about signal integrity.

Apple acquires Whoop, the hardware and sensor stack gets Apple silicon, and suddenly the Apple Health ecosystem has a product that your cardiologist and your marathon coach both take seriously.

While it doesn’t solve the Apple Watch’s biggest flaw - its battery life - it would provide some additional credibility to the Watch’s tracking metrics.

Whoop raised at a $3.6 billion valuation in 2021. It has not raised since. The founders have been running the company for thirteen years. The math on an Apple acquisition at $6-8 billion writes itself.

Whoop’s Dating Profile: Looking for someone with distribution, an App Store, and a decade of failing to build this in-house. Willing to overlook a history of dying at 3pm.

Odds of Successful Marriage: ~70%

SpaceX x Tesla

This one isn’t really a blind date. It’s more like matching with someone on Hinge that you grew up with and hoping they’ve been hit with the glow up stick. It’s also less of a marriage-by-choice and more an arranged marriage, but let’s take a look anyways.

Musk founded SpaceX in 2002 with his PayPal money, then led Tesla's Series A in early 2004 and took the chairman seat. By December 2008 both were weeks from running out of cash and he split what was left of his personal fortune between them. SpaceX later bought a quarter of a billion dollars of SolarCity bonds; Tesla bought SolarCity outright in 2016. Engineers have moved between the two for years, and they now share the Terafab project in Austin.

Elon Musk owns approximately 42% of SpaceX and roughly 13% of Tesla. He has spent the last three years building xAI, absorbing Cursor, buying stakes in Delivery Hero through X, and generally demonstrating that his preferred M&A strategy is to acquire things adjacent to things he already owns rather than build them from scratch.

SpaceX and Tesla are already adjacent in ways the market has not fully priced. Tesla’s energy storage business (Megapack, Powerwall) is the ground infrastructure for the same grid that SpaceX’s satellite network is powering from above.

Tesla’s manufacturing expertise in battery cells and electric drivetrains is directly applicable to SpaceX’s ambitions in electric propulsion and Starship recovery systems. Tesla’s Dojo supercomputer and its autonomous driving AI sit in the same compute stack as Grok.

The combination at current market caps would be worth somewhere north of $3 trillion. The governance is the obvious problem since Musk controls both, which makes any merger technically a related-party transaction that requires independent board approval on both sides. That process would be contentious. It would also be a near-certainty.

Tesla’s Dating Profile: $1.4 trillion market cap, autonomous driving AI, global manufacturing footprint, founder who also runs the rocket company. Looking for a formal relationship rather than continued informally shared management. Will require special committee approval.

Odds of Successful Marriage: ~80%. While this is an arranged marriage, it does require someone other than Elon to approve it. There is always a chance that someone with a morsel of independent thought decides to look at the numbers rather than just giving Elon the thumbs up.

Microsoft x Figma
Adobe tried to buy Figma for $20 billion in 2022 but ran into some regulatory problems. Figma went public instead in early 2026 and now trades at roughly a $16 billion market cap.

Microsoft has Copilot, Teams, PowerPoint, Word, and approximately $80 billion in cash. It does not have a design tool.

Every product team inside every enterprise that runs Microsoft 365 uses Figma to design things they then present in PowerPoint. That is an absurd workflow that exists only because Microsoft does not own the thing happening one step before PowerPoint.

Figma’s Dating Profile: Recently listed, design category leader, no enterprise sales team of its own. Looking for someone who already has 345 million commercial Microsoft 365 seats and wants to sell something into all of them.

Odds of Successful Marriage: ~50%. Adobe tried and the EU said no. Microsoft will argue it is a different buyer with a different competitive footprint, which is true. Whether Brussels agrees is the bet. It’s like having to ask your girlfriend’s estranged uncle for permission but he lives in a trailer park and you aren’t actually getting an inheritance benefit from being on his good side. If the EU clears it, this is a 10/10 strategic fit. If it doesn’t, Microsoft builds a worse version for five years and Figma stays independent.

Spotify x SiriusXM
Spotify has 777 million monthly users and 300 million subscribers, and after years of losses it finally prints money. What it doesn't have is growth in developed markets or an advertising business that works: ad-supported revenue grew 1% last quarter. SiriusXM has 33 million paid subscribers, a captive automotive audience, and has been bleeding subscribers since podcast listening moved to phones.

Together it’s the largest audio platform on earth with a paid subscriber base, an advertising network, podcast infrastructure, and a proprietary in-car distribution channel that Spotify has never had.

Separately, it’s one company with too many users and no profit and one with a contracting user base and a subscription model that belongs to a different era of commuting.

The deal structuring is not simple - Liberty Media owns the majority of SiriusXM - but Spotify has done harder things.

SiriusXM’s Dating Profile: Owns the car. Losing the phone. Looking for a partner with 600 million users and no opinion about Howard Stern.

Odds of Successful Marriage: ~40%. The strategic logic is sound, but Liberty Media’s ownership of SiriusXM makes the structure a nightmare; and Spotify has historically preferred to build rather than buy. If SiriusXM’s subscriber decline accelerates, this gets easier. Right now the price is too high for what is essentially a distribution asset.

JPMorgan x Robinhood
Robinhood has 28.4 million funded customers, the best retail brokerage app ever built, a gold subscription product growing at 40% year-on-year, and a market cap of ~$94 billion.

JPMorgan spent $1.2 billion building Chase Invest, which has fewer users and a worse app. It also spent $175 million acquiring Frank, the college financial aid startup, which turned out to be largely fabricated.

Robinhood is the only real entry point JPMorgan has into a generation of investors who have never walked into a branch and have no intention of starting. The problem is the price. HOOD trades at ~$94 billion, ~32% off its October 2025 peak, and a shareholder base that watched the stock hit $153.86 isn't tendering at $104.

Getting them there probably means a number north of $150 billion, nearly three times what JPMorgan paid for Bank One ($58 billion in 2004), and the largest deal in US banking since Citicorp and Travelers.

Jamie Dimon has called crypto “a pet rock” and Robinhood “a casino.” This deal would require him to own both. He would do it anyway, maybe.

Robinhood’s Dating Profile: 28 million users, great app, crypto exposure, no branches. Looking for a balance sheet. Not interested in being compared to a pet rock.

Odds of Successful Marriage: ~30%. JPM has the money (and advisory skills) to get the deal done, but there would likely be a culture clash between Robinhood and the traditional JPM teams. Additionally, Dimon would be forced to reverse his stance on the aforementioned issues, which isn’t impossible but isn’t likely. I would take more caution in integrating a start-up culture with an established banking culture.

Moonshot: SpaceX x TikTok (through X)
Elon Musk is a great builder. Look at PayPal, Tesla, and SpaceX. However, he is also not a stranger to buying into companies when he thinks he can improve them rather than build from scratch. He bought Twitter rather than building a social platform. He bought Cursor rather than training Grok to code.

X already has a payments infrastructure, a content delivery network, a 600 million user base, and a CEO who has been very public about wanting X to become the everything app, payments, commerce, video, AI, communication. The one thing X does not have is short-form video at scale. TikTok has 170 million US users, $100 billion in global GMV through TikTok Shop, and a forced divestiture from ByteDance that has been sitting in legal limbo for two years.

The national security argument that has blocked every other TikTok acquirer does not apply to Musk in the same way. He holds DoD contracts through SpaceX and Starshield. He ran DOGE. He has the kind of relationship with the current administration that makes a TikTok acquisition politically achievable in a way it has not been for any other buyer. ByteDance gets a clean exit north of $100 billion. X gets the short-form video engine it cannot build. SpaceX funds it, with maybe a little help from Daddy Elon.

The synergies between a satellite internet network, a social media platform, and a short-form commerce engine are real enough that you can model them. The probability that Musk is not already thinking about this is low.

TikTok’s Dating Profile: 170 million US users, $100 billion in GMV, legally complicated family situation, needs American owner urgently. Interested in someone with government connections, a satellite network, and a history of buying things rather than building them. Previous applicants need not reapply.

Odds of Successful Marriage: ~25%. Everything about the thesis is right and everything about the execution is chaotic. ByteDance will not sell cheaply. The DOJ will have opinions and Elon will have to stay on Trump’s good side through the review process. X’s balance sheet cannot absorb $100 billion without SpaceX writing the check. And Musk is already running five companies. But he has done crazier things and succeeded.

_{PRESENTED BY ALUMNI VENTURES}

Invest in High-Potential Startups Like These

Alumni Ventures is giving readers early access to high-potential startup opportunities across today’s most active sectors, co-invested alongside name-brand VC firms like Andreessen Horowitz (a16z), Bessemer, & Y Combinator.

You get:

• Curated deal flow of high-potential startups

• Invest alongside well known lead venture firms in these deals

• No cost to see deals

• No obligation to invest

See Current AV Deals →

_{ADVICE CORNER}

Things I Wish I Knew as an Analyst

Look, I know most of you come here for deal rundowns and deep dives, but I am also cognizant that it is now the start of a new banking season on Wall Street. Training is likely nearing a close, or over already, and new first years are hitting the desk. Despite them saying they have always dreamed of being investment bankers in the interview, they probably have no idea what is going on.

The good news is I didn’t either, and now I write a newsletter, so anything is possible! On a more serious note, here are some things I wish I knew as a first year banker. I recommend pinning these next to your color pallet on your cubicle.

Use your brain.
It is easy to fall into the trap in banking of being a process monkey. Turn comments, quick review, send to associate, repeat. There will be a lot happening all at once and falling into a trap of just grinding blindly is easy. Using your brain is the single easiest way to avoid all of the issues listed below. If something seems directionally incorrect, spend an extra 5 minutes thinking about it. A great example of this is in the First Brand’s deal. Apollo passed on the deal because it couldn’t figure out how the SOCF and Income Statement tied out for the Company, something all of the lenders should have done, but didn’t. Taking the extra 5 minutes to think before doing saved them millions and actually allowed them to make some money buying CDS’s. 

The model is not the work.
Every analyst thinks the model is the deliverable. It is not. The model is the research tool. The deliverable is the judgment you form from running it. The most dangerous analysts I have seen are the ones who conflate precision in the model with correctness in the conclusion. A model that is technically flawless and built on a wrong assumption is worse than a rough model built on the right one, because the flawless model will be trusted and the rough one will be questioned. Question the assumptions before you polish the formatting. You can have the most technically perfect model, but if it doesn’t line up with the back of the envelope analysis your Director or MD will run, then it’s wrong… and chances are it’s not just because yours is more dynamic. The best deals are often not the over analyzed ones, but the ones that are able to pass the smell test early.

Your MD’s reputation is on the line every time you send an email.
When you send something out, it goes with your MD’s name on it, whether their name is on it or not. The culture of checking everything twice before it leaves the building is not bureaucracy. It is the only way people can afford to delegate, and being someone who can be delegated to without supervision is worth more than any technical skill you will develop in your first three years.

The first number out of your mouth becomes the anchor.
In any negotiation, any IC presentation, any client call where a valuation gets mentioned before you are ready, that number becomes the floor or the ceiling of the entire conversation regardless of what the analysis actually shows. Know your number before you walk in. If you don’t have one yet, say you don’t have one yet. “I’m still working through the analysis” is a complete sentence. “Somewhere around six to eight times” when you actually mean six is a number you will spend the rest of the meeting defending. If anything changes between versions of a model in any material way, then you will be constantly explaining why that delta exists.

Most deals fail for reasons that were visible on day one.
The diligence process is supposed to surface new information. It usually surfaces information that was already there but that nobody wanted to say out loud before the process started. The seller knew the customer concentration was a problem. The buyer suspected the management team was not as strong as the deck suggested. Both sides wanted the deal to happen, so nobody said anything until the lawyers found it at page 400 of the data room. The questions worth asking on day one are not the obvious ones. They are the questions that would embarrass someone to answer. Again, refer to the First Brands Story.

Being right early is the same as being wrong
For those of you who studied economics, this should be obvious, but it is hard to learn. As Keynes said: “the market can stay irrational longer than you can stay solvent.” You can have the correct view on a deal, a market, a company, or a trade and be completely wrong about the timing. In finance, those two things produce the same outcome. Being early on a thesis that eventually plays out is a story you tell at dinner, not a return you report to your LP. Time in a position costs capital. Conviction without a catalyst is a hobby.

Your network is the only thing that compounds faster than interest.
Not the LinkedIn kind. The kind where someone you worked with at 26 calls you at 36 because they trust you. Every person you treat with respect on the way up, every analyst you cover for, every associate you spend an extra hour helping at 11pm on a Thursday, that is the actual long-term investment portfolio. The deals change. The firms change. The people stay. There is nothing more valuable than being able to call a lender or a lawyer at the 11th hour and having them help you because they like working with you.

Nobody knows what they are doing as completely as they appear to.
The MD who walks into the room like he owns it has been winging significant portions of his career for thirty years and has simply gotten very good at looking like he is not. The VP who gives you feedback with perfect clarity worked out what she actually thought on the way to your desk. The partner who sounds certain has a model in his head built on assumptions he has never fully articulated. You are not uniquely underprepared. You are new. There is a difference.

Being 90% correct is the same as being 100% wrong.
This was something a professor told me in college, and I never forgot it. You spend most of your time in college fighting for partial credit on things, or hoping that not knowing the answer to one question is fine, because you know the rest. That’s not how the real world works. If your model is 90% correct but the mechanics in one spot are wrong, you are wrong. As scary as that sounds, the simplest way to fix it is just to remember to focus and double check everything before it goes out.

_{THE WORST DEAL EVER MADE}

The Short Version

On January 10, 2000, America Online (AOL) announced it would acquire Time Warner in an all-stock deal valued at approximately $350 billion; the largest corporate merger in history.

AOL, a 15-year-old dial-up internet company with $5 billion in revenues and an absurdly inflated stock price, used that stock as currency to buy Time Warner, the largest media company in the world, which had $27 billion in revenues and owned HBO, CNN, Warner Bros., and half of America’s cable infrastructure.

The Nasdaq peaked fifty-nine days later. AOL’s stock fell 80% over the following two years. The synergies never materialized. The accounting turned out to be fraudulent. 

The combined company wrote off $99 billion in a single year and posted a net loss of $98.7 billion for fiscal year 2002, the largest annual corporate loss in American history. Both CEOs were gone within two years. AOL was eventually spun off in 2009 at a market cap of approximately $3.4 billion, down from $164 billion at the time of the merger.

Total value destroyed: roughly $200 billion.

That is the story. What follows is how it happened, why it happened, and why nobody stopped it.

_{PRESENTED BY DILIGENCESQUARED}

Five of the Ten Largest PE Funds Use the Same New CDD Firm

DiligenceSquared is what happens when two former Principals, one from Blackstone’s PE team and one from BCG, decide the consulting model they grew up in deserves a rebuild. Their firm delivers investment-grade market research and commercial due diligence, with both senior ex-MBB consultants leading every engagement (and signing off on every finding) and AI agents gathering evidence at a scale no human team could staff.

The work runs the whole deal funnel: industry assessment, market sizing, competitive landscape, voice of customer, and live-deal diligence. DiligenceSquared’s reports land in one to two weeks, make every claim clickable back to the evidence behind it, and they do it at a fraction of what the big three charge.

Five of the ten largest PE funds in the world are already clients. Anyone who signs off on deals should know this firm exists.

See how it works.

_BACKGROUND

The Players

In January 2000, AOL had approximately 26 million dial-up internet subscribers, which made it the dominant internet company in America by reach. Its market capitalization was approximately $164 billion. It had revenues of roughly $5 billion and cash flow of approximately $1.5 billion.

That is a 109x cash flow multiple, at the peak of the biggest speculative bubble in stock market history, for a company whose primary product was the sound a modem makes when it connects to the internet. You know the one (maybe).

Your VP had an AOL account and thought it was the future. Your analyst was twelve. Both of them are now watching this story unfold from the same Bloomberg terminal (or your iPhone on the subway), which is the real convergence story of the era.

The stock had tripled in 18 months. AOL insiders knew it would not last. 

Time Warner, on the other hand, was the largest media and entertainment company in the world. It owned HBO, CNN, Warner Bros., Warner Music, Time magazine, People, Fortune, Sports Illustrated, and the second-largest cable system in the United States. It had revenues of approximately $27 billion, just five times AOL’s, and real, durable, cash-generating assets built over decades.

Its market cap was approximately $83 billion.

So a 15-year-old dial-up internet provider with $1.5 billion in cash flow and a $164 billion stock price was about to buy a century-old media conglomerate with $27 billion in revenues and an $83 billion market cap… using stock as currency. 

The company with the fake number bought the company with the real one.

This is what the bubble looked like from the inside.

_BACKGROUND

The Deal

The mega-deal was announced January 10, 2000 but took a year to close, ultimately closing on January 11, 2001.

The deal was all-stock, done at the peak of the dot-com bubble. 

AOL shareholders received 1 AOL share for 1 share in the new combined company. Time Warner shareholders received 1.5 shares of the new company for every Time Warner share.

AOL paid a 71% premium to Time Warner’s pre-announcement price.

The combined valuation at announcement was approximately $350 billion, making it the largest corporate merger in history.

AOL’s ownership stake was 55%. Time Warner’s was 45%. It was called a merger of equals.

Steve Case became Chairman. Jerry Levin became CEO. Both would be gone within two years, replaced by Richard Parsons, who had the unenviable task of explaining to shareholders how a deal misses even your most bearish downside scenarios.

The strategic rationale: AOL’s 26 million internet subscribers would access Time Warner’s content. Time Warner’s cable infrastructure would supercharge AOL’s broadband ambitions. Cross-selling, convergence, the future of media. Every investment banker in New York and San Francisco nodded along. The CD-ROMs practically mailed themselves.

The Nasdaq peaked on March 10, 2000. Fifty-nine days after the announcement.

_{THE FALLOUT}

What Went Wrong

Everything. But in a specific order.

(1) The currency evaporated. AOL’s stock was the consideration. Between announcement in January 2000 and close in January 2001, AOL’s share price fell roughly 50%. Time Warner shareholders who had agreed to receive AOL stock at a 71% premium watched that premium disappear before the ink was dry. The deal closed anyway because walking away would have been an admission that both boards had made a catastrophic mistake, and neither board was prepared to say that out loud.

(2) The synergies were fiction. Time Warner Cable refused to give AOL preferential broadband access on its systems. The two cultures were incompatible in ways that could have been identified before the merger if anyone had wanted to look. AOL was a scrappy, sales-driven internet company that used aggressive accounting.

Time Warner was a prestige media institution whose executives genuinely believed AOL was a fad. Putting them together was like your firm acquiring a WeWork. Everyone smiles at the press conference and then nobody knows whose desk is whose.

The accounting was fraudulent too. The SEC and DOJ investigated AOL for inflating its advertising revenue in the years before the merger. Time Warner ultimately paid $300 million to the SEC and $210 million to the DOJ to settle. Some of the growth that justified the deal’s valuation was not real. The CD-ROMs contained more value than the revenue figures.

(3) The goodwill writedown. When the company marked down the value of AOL’s assets to reflect reality, the result was a $99 billion goodwill impairment charge in 2002, the largest in corporate history at the time. AOL Time Warner reported a net loss of $98.7 billion for fiscal year 2002. In a single year. Larger than the GDP of most countries.

(4) Jerry Levin resigned in December 2001. Steve Case resigned as Chairman in January 2003. The company dropped “AOL” from its name the same year. That is how you know a merger has gone badly. In 2009, AOL was spun off as an independent company. Market cap at separation, approximately $3.4 billion. Down from $164 billion nine years earlier!

_{THE FALLOUT}

Why it Happened

The RJR Nabisco deal failed because leverage amplified a bad macro outcome. The AOL-Time Warner deal failed before it started, because the foundation it was built on was imaginary.

AOL’s management knew the stock was overvalued. A company that believes its stock is fairly valued does not rush to convert it into real assets. Case spent inflated internet currency to acquire durable media infrastructure before the market corrected. It was rational from AOL’s perspective. It was catastrophic for Time Warner’s shareholders, who traded their stake in HBO, CNN, and Warner Bros. for AOL stock that was about to fall 80%.

Levin’s decision is harder to explain. Time Warner’s assets were real. Its cash flows were real. It did not need to merge with anyone. The most generous reading is that Levin genuinely believed in the convergence thesis. The less generous reading is that he was flattered by the premium and excited by the idea of running the world’s largest media company, but neither instinct served his shareholders.

Both companies had advisors. Both advisors collected fees. Nobody told either board to stop.

_{THE FALLOUT}

Lessons Learned

A high stock price is not a competitive advantage. It is a liability looking for somewhere to land (SpaceX and Tesla merger could be similar, perhaps?)

AOL did not accidentally overpay for Time Warner. It deliberately converted an overvalued currency into a real asset before the market corrected.

The tragedy is that it worked, from AOL’s perspective. Case got out. It was Time Warner’s shareholders who held the bag.

The deal announced as the dawn of the internet age ultimately destroyed approximately $200 billion in shareholder value, produced the largest single-year corporate loss in American history, resulted in two federal settlements for accounting fraud, and ended both CEOs’ careers.

You've got mail. You destroyed $200 billion in shareholder value.

Situation Overview

RJR Nabisco was formed in 1985 when R.J. Reynolds Tobacco merged with Nabisco Brands, the packaged food company behind Oreo, Ritz and Chips Ahoy.

The logic was straightforward and, in hindsight, completely wrong: tobacco was facing regulatory pressure, so Reynolds would diversify into food. What they actually created was a conglomerate trading at a discount to its parts, run by a CEO whose lifestyle had become the headline, and generating enough cigarette cash flow that it looked, to the right kind of buyer, like a perfect LBO target.

F. Ross Johnson was that CEO. By 1988 he was legendary on Wall Street for what can be charitably described as an enthusiastic approach to corporate spending. RJR operated roughly ten aircraft and thirty-six pilots out of a facility employees called the “Taj Mahal” in Atlanta, ferrying executives and celebrity friends wherever they needed to go. Johnson had placed Don Meredith, Frank Gifford, and Dinah Shore on the company payroll as “consultants.”

The stock sat at $56, well below what any serious analyst thought the brands were worth. Johnson, watching it stagnate after the 1987 crash, had decided he wanted it all for himself.

Find out more below….

_{PRESENTED BY PRIMER}

Claude for Excel Only Scored 48% on its Analyst Exam

A leading analyst training firm put the top AI tools through a classic first-year assignment, a fully integrated three-statement model for Apple, and graded them like trainees. ChatGPT scored 43%, Claude managed 48%, and Shortcut topped the field at 50%. All of them failed. Yet, most investors still paste tickers into these chatbots and trust their modeling (and research).

Tier 1 Hedge funds run their work through Primer instead, which took the same test and scored 81%, a full 31 points clear of Shortcut.* Built by former hedge fund investors to help analysts do higher quality, more robust research. Until now, access was reserved for institutions. Today it's open to everyone.

Get your own hedge fund analyst.

The Bidding War

On October 20, 1988, Johnson and his team, backed by Shearson Lehman Hutton, proposed to take RJR private at $75 per share, roughly $17 billion. He wanted 20% of the equity for himself and his management team, with no personal capital required. He would run the company he was already running, own a fifth of it for free, and collect the upside.

The board made one decision that changed everything: instead of negotiating privately, they declared the company "in play" and opened a formal auction. This was the correct decision for shareholders. It was an absolute disaster for Ross Johnson's weekend plans.

Henry Kravis at KKR had reason to think RJR was his deal. He had raised the idea of an LBO with Johnson months earlier, walked him through how it would work, and waited while Johnson sat on it. When Johnson finally moved, he went to Shearson instead, a firm with almost no LBO experience but willing to hand him terms KKR never would have. Kravis found out the way everyone else did.

Henry Kravis took it personally, and he had the balance sheet to do something about it.

On October 26, KKR entered at $90 per share, approximately $20.4 billion, and explicitly excluded Johnson and his team from post-acquisition management. Kravis was telling the board, in writing, that he intended to clean house.

What followed was six weeks that would have seemed implausible as fiction.

KKR and Johnson briefly attempted a joint bid, reached a tentative agreement, and blew it up within 24 hours. Forstmann Little entered and withdrew. The special committee set a hard deadline of 5pm on November 18. Johnson's group came in at $100 per share. KKR, still working without access to inside information, bid a more cautious $94. First Boston missed the deadline entirely, then submitted a preliminary tax-deferral structure that the committee found interesting enough to reopen the auction and schedule a second round.

The deadline held for eleven days.

The final bids, submitted November 30, 1988:

1. KKR: $109 per share, approximately $25 billion, predominantly cash

2. Management group: $112 per share, approximately $26 billion, more complex and contingent

3. First Boston consortium: $118 per share, almost entirely contingent securities the committee didn't believe were financeable

The committee recommended KKR’s $109 over the management’s nominally higher $112. The lesson is permanently worth knowing: certainty of close is worth more than headline price.

KKR had cleaner financing, better execution credibility, and came without the self-dealing baggage of a CEO trying to buy his own company at a price he’d set himself.

The deal closed April 28, 1989, at a total enterprise value of approximately $31.4 billion. This was the Largest LBO in history and held that record for seventeen years.

The Financing

KKR’s capital structure was $12 billion in senior bank and bridge debt, short-term, needed to be refinanced, $11 billion in high-yield junk bonds, underwritten by Drexel Burnham Lambert at rates above 14%, and $1.5 billion in equity, the largest equity check KKR had ever written.

All in, the debt-to-equity ratio was 9:1, meaning for every dollar of equity, KKR borrowed nine.

The entire bet rested on RJR’s tobacco business generating enough cash to service $25 billion in debt.

Michael Milken made it possible. He had spent a decade building the junk bond market from scratch, and by 1988 could raise billions in days. The RJR underwriting was the peak of his influence.

The structure included PIK bonds on the subordinated tranches, meaning interest accrued and compounded rather than being paid in cash. This was clever at the time and reckless in retrospect: if anything went wrong, the debt load would be larger when the crisis hit, not smaller. Maybe it wasn’t so kind after all…

KKR received approximately $75 million in fees for its participation. Total fees across all parties ran close to $1 billion, roughly 4 cents on every dollar of deal value. There is a reason investment bankers remember 1988 fondly.

Post Mortem

KKR’s thesis: RJR’s tobacco business would generate reliable cash flows while the food brands (Oreo, Ritz, Grey Poupon) could be sold or retained for additional value. Clean. Sensible. Wrong.

The first crack came from the structure, not the business. Some of the original securities had reset provisions: if the bonds traded below issue price, the interest rate reset higher. When the junk bond market seized up in late 1989, RJR's paper fell and the resets triggered. As RJR struggled, its debt became more expensive, which made it struggle more. KKR had to inject an additional $1.7 billion in equity in 1990 and secure $2.25 billion in new bank loans just to avoid default. The structure that was clever going in became a trap on the way out.

Philip Morris finished the job in April 1993. In what became known as "Marlboro Friday," Philip Morris cut the price of Marlboro cigarettes by 40 cents per pack to fight back against generics. Its own market value fell roughly $13 billion in a single day. RJR's cash flow projections, built on pricing power that no longer existed, collapsed.

KKR took RJR public again in 1991 and unwound its position over the following years. By 2000, the food business was sold to Philip Morris for $14.9 billion. The tobacco business was spun off separately. KKR recovered most of its capital. Its LPs made approximately what they would have made in short-term bonds, without the nine-to-one leverage and five years of stress.

Ross Johnson walked away with $53 million, so maybe there was one winner in all this.

What people forget is RJR already had billions in investment-grade bonds outstanding when the LBO was announced, held by MetLife, Jefferson-Pilot, and other long-term institutional investors who had bought them as obligations of a conservatively managed company.

When KKR loaded $25 billion of new, higher-priority debt onto the structure, those bonds were instantly subordinated. Their credit quality collapsed from investment-grade to near-junk overnight.

MetLife sued, and lost. The court held that the implied covenant of good faith could not add a term the parties never bargained for, and that these were sophisticated investors who bought indentures containing no debt limitations. The contract was the contract.

So the market wrote a better contract. The lasting consequence is a covenant that now appears in almost every investment-grade bond indenture: the poison put. It allows bondholders to demand early repayment if a leveraging event occurs alongside a ratings downgrade.

Lasting Impact

LBO scale. Before RJR, the practical ceiling on an LBO was around $5-8 billion. After RJR, that ceiling was gone. The only constraint was how much debt the junk bond market could absorb, and Milken had demonstrated the answer was “quite a lot.”

CEO incentives. Johnson’s attempt to buy his own company below intrinsic value, reserving 20% of equity for himself for no personal investment, was the textbook illustration of the principal-agent problem. The 1990s wave of executive stock options, the shift toward aligning management and shareholder interests, the increased scrutiny of CEO pay all trace partly back to the disgust that greeted Johnson’s economics.

Credit covenants. The MetLife bondholders invented the poison put. The entire investment-grade credit market repriced event risk after RJR. A change in standard bond documentation that persists in every credit transaction today, because one deal proved the existing documentation was inadequate.

Private equity’s public image. Before RJR, LBOs were a niche strategy. After Barbarians at the Gate, KKR was the firm every young banker wanted to work for and every politician wanted to investigate.

The Lesson

The RJR Nabisco LBO is not a story about greed gone wrong. Afterall, I tend to agree with Gordon that greed is good.

The lesson is that leverage is not a strategy. It is a multiplier. At 9:1, it multiplied the upside if everything went right. When Philip Morris cut Marlboro prices by 40 cents in 1993, it multiplied the downside instead.

To put this in terms bankers from the ‘80s could understand, as my corporate finance professor once told me: leverage is like cocaine, it makes the great times great and the bad times way worse.

The thesis was good but price was too high and the structure was too fragile. The market changed in a direction nobody in the room in November 1988 was modeling.

The deal that remains the most famous in history, that gave private equity its public identity, that changed credit markets permanently, that inspired the best business book ever written, ultimately returned approximately nothing to the people whose money actually funded it.

That is the part nobody puts in the pitch deck.

_{DEAL OF THE MONTH}

Uber Takes Delivery Hero for a Ride

If you’ve been reading Buysiders since May, you already know the backstory. Uber had quietly built a 36.83% stake in Delivery Hero while publicly offering €33 per share. Delivery Hero’s shareholders told them where to go. Uber then paid €40 in the secondary market for Aspex Management’s block the same week it was officially bidding €33. Bold strategy. Even bolder when you own 37% of the company you’re trying to low-ball.

On July 16, it was over.

Uber announced a definitive agreement to acquire 100% of Delivery Hero at €41.50 per share in cash, representing a total equity value of $14.8 billion for the whole company, or $13.7 billion adjusted for Uber’s existing stake.

The premium to the unaffected share price is a modest 8.7% because Delivery Hero’s shares had already re-rated on months of speculation that Uber was going to do exactly this. Uber committed a bridge facility of ~€14 billion to fund the offer. So the company that spent years losing money on food delivery is now borrowing €14 billion to own all of it. Progress.

Both boards unanimously recommended the deal.

Prosus, which had been ordered by the EU to reduce its stake below 10% and therefore had every reason to want out, irrevocably committed to tender its 17% position, bringing Uber’s total economic interest to approximately 53% upon acceptance. Minimum threshold: 50% plus one share.

To manage the regulatory overlap, Delivery Hero simultaneously agreed to sell its business in 14 markets to New York-based SSW Partners for $1.6 billion. These are the markets where Uber Eats already operates.

Uber is lending SSW the funds to finance that transaction, with repayment over time. So Uber is financing the thing it’s selling to fix the regulatory problem caused by the thing it’s buying. We’ll give that one to the lawyers. They earned it.

The combined company:
• 99 markets with combined mobility and delivery operations
• $236 billion in combined pro-forma Gross Bookings (2025)
• $1.2 billion in run-rate synergies identified
• Non-GAAP EPS accretive at close, guided high-single-digit accretion by year three
• Transaction values Delivery Hero at approximately 8x EV/2027E Adjusted EBITDA (including synergies and Uber’s existing stake)

The strategic logic is simple enough. Uber and Delivery Hero have been burning capital competing with each other in markets where neither had a structural advantage.

One platform across 99 markets is better than two. Whether the integration actually works across 14 time zones with a German public company bureaucracy is the open question nobody in the investor presentation was asked to answer.

Uber CEO Dara Khosrowshahi: “Together, we’ll nearly double the number of markets where we offer both mobility and delivery services, scaling a proven platform that we believe will create significant long-term value for our customers and shareholders.”

_{PRESENTED BY DILIGENCESQUARED}

What Happens When You Give Senior ex-MBB Consultants CDD-Native AI

DiligenceSquared was built on a simple idea: keep the senior judgment, automate the rest. AI voice agents can now handle what armies of junior analysts used to do: sourcing experts, running interviews, and synthesizing transcripts. On the human side, senior ex-MBB consultants oversee and pressure-test every project, so the output holds up to an investment committee. The turnaround time becomes days instead of weeks.

It’s already in use by several of the world’s leading PE funds, representing over $2T in combined assets, across the full deal funnel: early-stage screens, pre-LOI diligence, active deals, and portfolio value creation.

See it in action today. Explore DiligenceSquared.

_{STRATEGIC DEAL OF THE MONTH}

Martin Marietta Hoists LHoist

You may not spend a lot of time thinking about lime and limestone, I do because I like rocks, but you probably don’t.

That’s pretty reasonable, because nobody has ever described a limestone deal as sexy. After this week, though, you might start paying attention.

On June 29, Martin Marietta Materials announced a definitive agreement to acquire Lhoist North America from Belgian parent Lhoist Group for $13.5 billion, structured as $7 billion in cash and $6.5 billion in Martin Marietta stock.

It is the largest deal in Martin Marietta’s history by a factor of roughly three, and represents 36% of its own market cap. MLM shares dropped roughly 3% on announcement day.

The shareholders naturally had questions. Specifically about the 17x EBITDA they just paid for rocks. I mean, I can buy limestone for less at my local quarry. 

Lhoist North America:
• 20 quarries and production facilities
• 45 distribution terminals across the US
• $1.8 billion in gross sales for the twelve months ended December 2025
• $786 million in Adjusted EBITDA
• Implied multiple: approximately 17.2x EBITDA

That multiple needs an explanation and, since I actually know a thing or two about this industry, I’ll do my best not to bore you too much with the details.

Lime is essential for steel manufacturing. It is essential for water and wastewater treatment. It is essential for flue gas desulfurisation at power plants. And it is increasingly essential for processing lithium from spodumene ore for EV batteries.

Every data center being built, every grid upgrade being planned, every new power plant going up needs lime at some stage of construction or operation. Behind every AI data center is a limestone quarry nobody is writing about.

That is why 17x EBITDA for a limestone company makes sense in 2026 in a way that it would not have in 2019. The infrastructure supercycle is real, and lime is load-bearing. Literally.

CRH announced its $8.5 billion acquisition of Arcosa the same week: building materials consolidation, same infrastructure thesis, different subsegment. When that happens, it usually means the thesis is right and everyone is racing to the same rock.

_{INTERNATIONAL DEAL OF THE MONTH}

Alcoa Goes to the South

Aluminum hit a four-year high above $3,760 per tonne this week, driven in part by supply concerns from the US-Iran conflict. Alcoa announced its deal on June 30. The timing is either brilliant or lucky. In commodity M&A, both look the same on the tombstone.

Alcoa agreed to acquire the bulk of South32’s global aluminum portfolio for an upfront consideration of $4.1 billion, consisting of $3.1 billion in cash and approximately 17 million newly issued Alcoa shares worth roughly $1 billion, plus assumed net debt of ~$750 million, bringing the implied enterprise value to approximately $4.7 billion.

South32 can also receive up to $750 million in additional contingent cash through 2030 via a CVR linked to alumina and aluminum prices. Total at the ceiling: $5.6 billion.

What Alcoa is buying: South32’s 86% stake in Worsley Alumina and Boddington bauxite mine in Western Australia. 100% of Hillside Aluminium and idled Bayside smelter in South Africa. 33% interest in the MRN bauxite mine in Brazil. 36% and 40% stakes in Alumar alumina refinery and aluminum smelter in Brazil.

What South32 is keeping: its $2.16 billion Hermosa development in Arizona expected to be the only US project capable of producing both zinc and manganese and its broader pivot toward copper and critical minerals. South32 sold its coal assets in 2024 and its ferronickel mine in Colombia in December. The Alcoa deal is the largest piece of its exit from aluminum, which it apparently no longer finds exciting enough. To each their own.

Alcoa expects ~$900 million in net present value synergies from the deal, and the transaction is expected to be immediately accretive to earnings.

Post-close, Alcoa’s bauxite production share rises to approximately 13% of global supply and its smelting capacity increases by roughly 26%.

It becomes the largest pure-play upstream aluminum producer outside China by a meaningful margin. China, for reference, produces about 57% of the world’s aluminum. Alcoa at 13% of bauxite is not catching up, but it’s a start.

The CVR structure is the clever part.

It gives South32 meaningful upside if aluminum prices stay elevated through 2030, while capping Alcoa’s exposure if they normalise. Both sides get a commodity hedge built into the deal, which is either smart structuring or a sign that neither side is completely sure where aluminum is going. Frankly, that’s a fair read.

Alcoa shares fell nearly 5% on announcement day.

Alcoa CEO William Oplinger “Together, these are world-class refineries that complement our existing footprint and enhance our ability to generate value through the cycle.” William did not mention the 5% share drop.

Buysiders’ H1 Recap

H1 is officially in the books, which means that Europeans are on vacation - especially with the World Cup ongoing. Although, I am not sure from a deal flow perspective we will miss their presence.

Those of you who have been reading for a while may have noticed that in H1’26 I have had a smaller number of interesting deals to cover. If you didn’t notice that, I accept Venmo or Zelle if you want to tip your waitress.

If you did notice this - you aren’t wrong. There was less quantity of deals to write about in H1, but that doesn’t mean that H1 was bad.

In fact, H1 2026 was the strongest first half for global M&A since records began. $2.85 trillion in announced transactions, up 50% from H1 2025. Q2 alone posted $1.6 trillion, a quarterly record.

There were 47 deals above $10 billion, totaling $1.3 trillion, meaning the 47 largest transactions - out of roughly 24,000 - accounted for almost half the entire market by value.

Now, I know I sound like your MD cherry-picking the league tables to justify ranking above Goldman, but genuinely, H1 is one for the record books.

Total deal count fell roughly 9% versus H1 2025, the weakest first-half count in six years. Fewer deals, sure, but they were dramatically bigger ones.

JPMorgan’s North America co-head of M&A, Charles Bouckaert, told Bloomberg: “The bigger, more high profile, more complex stuff is a very big part of the market, which is pretty unusual.”

Tom Miles, Morgan Stanley’s global M&A co-head, put it simpler: “People have just accepted the volatility and are investing through it as opposed to waiting until it’s over.”

So with the stage set, I give you Buysiders’ H1 wrapped. It’s like a Spotify wrapped, but worse because instead of a reminder of what brought you joy this year, it’s a reminder of all the deals you missed out on because you were pitching a 5-way merger that would be the largest fee event in bank history instead of working on the SpaceX IPO.

_{PRESENTED BY TOLTIQ}

PE Firms Are Making a Third of the Bets at Four Times the Size

Source: PwC’s US Deals 2026 midyear outlook.

The midyear numbers are in.

PE deal volume fell 34% in the first half while average deal size nearly quadrupled. Capital is crowding into fewer, higher-conviction deals. Diligence windows, meanwhile, haven't gotten any longer. Most deal teams still review the top contracts by value, sample the rest, and assume nothing important is buried in the remaining few hundred agreements. These new trends, however, are forcing firms to seek custom-built AI solutions for the full lifecycle of the deal.

That is why PwC recently embedded ToltIQ across its Deals practice. The platform handles the document-heavy front end of a deal - ingesting, classifying, and cross-referencing thousands of data room files - so teams get a searchable view of the entire room from day one and spend their time on judgment and client conversations instead of laborious review work. The beauty of the platform is that each finding is cited back to the underlying document, so it is fully auditable. Built by a former Partner and CIO at KKR, ToltIQ knows the industry.

Book some time to demo the platform and see it firsthand.

Riding the H1 Waves

Q1 was driven by mega-deals and a single extraordinary outlier. SpaceX’s acquisition of xAI in February 2026 at a revised combined valuation of approximately $250 billion accounted for nearly 30% of total Q1 deal value by itself. Remove that one transaction and Q1 was strong but not record-setting. S&P Global’s Q1 data showed only 7,924 deals announced globally.

That’s a 30% drop from the prior year, not an auspicious start for banks who make their money on volume and are cut out of larger deals. Interestingly though, the aggregate value of deals in Q1 was $861.1 billion, the strongest Q1 since 2021. The K-shape was already visible in January: megadeal values up 40% year-on-year, mid-market volumes down materially. Check in on your non-megafund PE friends who told you their carry was worth millions. 

Q1 also saw the announcement of two other landmark deals your Buysiders subscription covered in depth: McCormick’s combination with Unilever Foods in a $44.8 billion Reverse Morris Trust, and BlackRock/GIP and EQT’s definitive agreement to acquire AES Corporation for $33.4 billion. For those of you keeping score, that’s three transformational transactions in the first quarter alone.

Q2 was more of the same. NextEra’s $66.8 billion acquisition of Dominion Energy came in May. Shell’s $22 billion purchase of ARC Resources in Canada came in April. The Paramount-Warner Bros. Discovery combination at $111 billion, which had been announced late 2025, navigated its shareholder approval in April and is now in regulatory review. Given Trump has free time with the US out of the World Cup, he may turn his attention back to this one.

Additionally, SpaceX’s $60 billion all-stock acquisition of Cursor landed in June days after its IPO. Q2 2026 was, individually, the single largest quarter for M&A deal value in history at $1.6 trillion. For those of you wondering if you will reap the rewards in bonus season, I remind you of the Ray Dalio quote: “ever notice how the cashflow never seems to be flowing towards you.”

What drove both quarters was the same thing. In fact, it was so obvious that even PwC picked up on it, and we know those middle-to-back office guys have never done an actual deal. Granted, the trend is so obvious that if you missed it, you may as well join Jim Cramer on CNBC.

In their mid-year report PwC pointed out that “AI is redirecting capital towards data centers, power, grid infrastructure, and frontier platforms while forcing buyers to reassess which sectors and business models are most exposed to disruption.”

It’s hard to argue with that hard hitting analysis by PwC, I mean these are the direct competitors of the guys who told NYC that trash cans would solve its trash problem, so what can we really expect.

Ribbing them aside, the trend is there. Most of the deals of consequence in H1 2026 can be split into three categories: (1) companies buying AI capability, (2) companies buying the energy to run AI, and (3) companies restructuring themselves to compete in a world where their existing business model is getting disrupted by it (shoutout Allbirds for the all time pivot from shoe company to AI company in 2026).

Sector by Sector Review

Sure, the aggregate numbers are all well and good, but let’s roll up our sleeves and look across sectors. After all, we have seen a pretty wide breadth of deals already this year, and this is coming from the guy who typically writes about food deals. Though I have been forced to be more original in H1’26 and I thank you, the bankers, for giving me variety.

Technology pulled in $649 billion, up 90% year-on-year and was the largest sector by value.
Ninety percent growth year-on-year is sweet, but always read the footnotes. Have you never opened a CIM? The fine print is that a significant portion of that is SpaceX-related. Strip the SpaceX/xAI merger and the Cursor acquisition, and you still have strong underlying tech M&A, but the headline multiple is inflated. What’s real and structural is the AI coding and infrastructure subsector, which attracted more capital in the first six months of 2026 than it had in any prior full year.

The pattern in tech M&A is now reliably: large strategics buying AI tooling companies (Qualcomm/Modular for $4 billion to escape CUDA dependency, SpaceX/Cursor for $60 billion to shore up Grok…good luck), and AI labs receiving massive capital infusions that technically count as M&A when they involve equity.

OpenAI’s $122 billion Series D closed in Q1. Anthropic’s $65 billion Series H closed in Q2. Between those two transactions alone, roughly $187 billion flowed into frontier AI, an amount that rivals the entire tech M&A markets of some prior years. Some people say we are in an AI bubble, and if I could predict those things I probably wouldn’t tell you that in a free newsletter, but the better question is: are people prepared for what happens if it’s not a bubble?

Energy and Power: up 41% year-on-year.
The Energy and Power sector generated $146 billion in deal value in Q1 2026 alone, 8% of the entire market, driven almost entirely by the collision between the AI data centre buildout and an electrical grid that was not designed to power it.

Four companies — Alphabet, Amazon, Meta, and Microsoft — are expected to spend more than $700 billion in 2026 building AI infrastructure. That electricity has to come from somewhere, even if that somewhere is pulling power away from your home.

NextEra/Dominion at $66.8 billion and BlackRock-GIP/EQT/AES at $33.4 billion are the two clearest examples, but the theme runs deeper than two deals. Shell’s $22 billion ARC Resources acquisition is a gas supply deal for an LNG terminal positioned to serve Asian power demand. Every major energy deal this year traces back to the same underlying dynamic: AI compute needs power at a scale the existing infrastructure cannot deliver.

Paul Weiss partner Laura Turano told Bloomberg: “You can’t overstate how much better the regulatory environment is.” That is the other reason energy M&A ran hot in H1. Trump administration antitrust enforcement has been notably more permissive on strategic combinations than the prior four years, and the energy sector, particularly grid infrastructure, has been explicitly encouraged through executive order and NOAA permitting reform. I mean listen, the guy got FIFA to overturn a red card, surely he can get any deal over the line.

Industrials: up 57% year-on-year.
Industrials is less sexy these days than AI or Energy, but still consequential (or so every middle market PE firm tells me).

The driver is a similar grid and AI-adjacent infrastructure thesis applied further down the supply chain: KONE’s combination with TK Elevator, Safran’s €2.19 billion acquisition of Exail Technologies (autonomous naval systems), and a wave of mid-market industrials consolidation targeting companies with exposure to grid upgrades, EV supply chains, and defense manufacturing. This is where the institutional PE firms have been most active on a deal-count basis even if the individual transactions are smaller.

Pharmaceuticals and healthcare: Lilly injects life into the sector.
This is your reminder that I know nothing about healthcare which is why I typically don’t cover it. Let me know if you’d like to see more healthcare deals in future editions, and I can try to get up the curve.

Eli Lilly announced three acquisitions on the same morning in May — Curevo ($1.5 billion, shingles vaccine), LimmaTech ($780 million, gonorrhea/staph vaccines), and a third at $1.55 billion for Epstein-Barr virus platform.

AbbVie’s acquisition of Apogee Therapeutics at roughly $10.9 billion was the headline pharma deal of the quarter. Boston Scientific’s $14.5 billion acquisition of Penumbra is expected to close in H2.

From the little I know, the macro pattern across pharma M&A appears consistent with prior years but is accelerating rapidly. GLP-1 profits are funding acquisitions, patent cliff risk is forcing pipeline deals, and large strategics are paying premiums of 50-100% over recent share prices because they cannot afford to wait for assets to mature on their own timelines.

Biotech M&A tracking at 46 announced transactions through June 23 per BioBucks, led by Recordati/CVC at approximately $12.4 billion equity value.

Financial services: more deals, lower value. A two-speed market.
EY’s H1 financial services M&A report is one of the more interesting data points of the half. Banks, insurers and asset managers disclosed 1,137 deals globally, up 3% versus H1 2025, but total disclosed value fell from $191.3 billion to $134.5 billion, because the number of megadeals above $1 billion dropped from 37 in H1 2025 to 25 in H1 2026. Some niche boutique bank you’ve never heard of just dropped a pitch deck with it above GS by number of deals done in FIG despite earning ⅓ of the total fees of GS.

The volume increase is real: financial services is one of the few sectors where deal count actually rose in H1, driven by mid-market consolidation in insurance, RIA roll-ups, and accounting firm PE take-ins. The value decline reflects a specific composition shift: there was no Capital One/Discover-scale transaction in H1 2026. The KKR/Crowe deal at $3 billion is significant for what it signals about professional services consolidation, not for its contribution to the aggregate number.

Cross-border inbound M&A into the US from non-US financial services buyers rose from $4.8 billion in H1 2025 to $15.9 billion in H1 2026. The Santander/Webster Financial acquisition is the $12.2 billion elephant in the room when that comes up.

The Private Sector

The private markets picture in H1 2026 is genuinely contradictory until you understand the bifurcation. Global PE fundraising fell approximately 30% year-on-year to $287 billion. Global M&A hit a record. VC deal value set records. These things are all true simultaneously because private markets in 2026 are two distinct economies wearing the same label.

The large-platform economy is booming.
Blackstone, KKR, Apollo, Andreessen Horowitz, and General Catalyst are raising from sovereign wealth funds, pensions, and endowments at $5 billion minimum check sizes. The top five managers captured approximately 73% of all new LP capital in Q1.

They are deploying that capital into megadeals. PE-driven transactions totaled $583 billion in H1, up 54%. They are also deploying that capital into continuation vehicles that allow them to hold aging portfolio companies without forcing a traditional exit at depressed multiples.

Average deal size rose nearly 4x in H1 2026 versus H1 2025. That is not a signal of market health. It is a signal of extreme capital concentration.

The mid-market and emerging manager economy is contracting.
A year ago, it felt like everyone was opening up a PE fund named [Tree][Rock] focused on middle-market family owned businesses in the midwest with operational excellence.

Deal volume in H1 2026 declined 34% for PE broadly. There were fewer deals, but much bigger ones among the largest platforms. Mid-market GPs face valuation gaps between sellers anchored to 2021 marks and buyers priced to 2026 exit conditions.

The PwC Private Equity Midyear Outlook puts PE entry multiples in large deals at 11.2x EBITDA in H1 2026, with mid-market exits coming in at 9.8x. The spread between entry and exit multiples has compressed significantly, and that compression is the single biggest threat to returns in 2021-2024 vintage funds. Remember when IC said your entry and exit multiples should be the same? Yea, stop betting on multiple expansion.

The exit problem is real and the secondary market is trying to absorb it.
The exit backlog that has built since 2022 has not meaningfully cleared in H1 2026. IPO windows remain narrow outside the AI mega-listings. Sponsor-to-sponsor transactions face valuation scrutiny. LP distribution rates are low relative to historical averages.

In response, continuation vehicles and GP-led secondaries have become the primary liquidity mechanism and an important structural feature.

Total secondary market volume hit $233 billion in 2025, the largest year ever, and 2026 is tracking ahead of that pace. GP-led secondaries alone were $115-116 billion in 2025, up 53% year-on-year, with continuation vehicles representing 16% of all sponsor exit volume by Q3 2025. The discount-to-NAV on LP-initiated secondary sales has compressed from 15-20% in 2023 to 8-12% in 2026, as more capital chases the same secondary supply. High-quality single-asset continuation vehicles are clearing at or near par.

The VC picture has the same concentration problem as PE but even more extreme. Q1 2026 global VC deal value hit $330.9 billion, which is a record by all accounts. But OpenAI’s $122 billion Series D alone represented 41% of the quarter’s total global VC investment. Remove OpenAI and it was an unremarkable quarter (thanks, Sam!).

In H1, OpenAI and Anthropic together attracted more than 40% of all venture funding globally. The broader ecosystem is showing signs of life. 32 companies went public at values above $1 billion in Q2, led by SpaceX, Cerebras, and Quantinuum, but the headline numbers are being carried by two companies with historical precedent-setting fundraises.

H2 Predictions

Since I’m still riding my high from the NFLX x WBD prediction (even if that turned out to be a miss in the end) and from calling the OpenAI filing to the day, I am going to put out some trends I am watching for in H2. Let’s see if the Oracle can be right again.

1. The IPO window.
   SpaceX listed June 12. Anthropic filed in early July. OpenAI targets September. If all three price and trade well from listing, they become the signal that unlocks a broader IPO market for companies that have been waiting since 2022. The PE exit backlog directly depends on this: if the window opens properly, expect a wave of PE-backed company prospectuses in Q3 and Q4. If it doesn’t, and SpaceX’s post-IPO decline to below its $150 opening price is a cautionary data point, the secondary market will absorb more of the backlog and GPs will face another difficult conversation with LPs about distribution timelines.

2. European regulatory reform.
   The EU’s 98-page draft Merger Guidelines closed public comment on June 26. An economic study on the dynamic effects of mergers is expected in September. If the new guidelines streamline approval timelines, European cross-border M&A (which had its best H1 start since 2018) accelerates in H2. If the guidelines ultimately add complexity rather than reduce it, the deal activity that’s been building in EMEA will hit a regulatory wall in Q4. This will matter most for the UK-listed companies that have been acquisition targets all year. Remember the adage: America innovates, China replicates, and Europe regulates.

3. The PE liquidity reckoning.
   The continuation vehicle market is functioning as a pressure valve, but it is not a permanent solution. The firms with the cleanest exit track record — demonstrated distributions to paid-in capital, not paper marks — are raising quickly. Everyone else is facing longer fundraising timelines, reduced targets, and LP investment committees that have run out of patience with the “it’s still early days” framing. The bifurcation that characterized H1 - mega-platforms winning, everyone else grinding - will intensify in H2 if the IPO market doesn’t open broad enough to clear the backlog meaningfully. Once again, check in on your PE friends.

Regardless, this has been a stellar H1 for deals. We have seen incredible success stories and the largest IPO in history drop all in the start of the year. I hope H2 will be as successful as H1, even if that means I have my work cut out for me in December when I write my annual Buysiders awards.

_{NEWS ROUNDUP}

Top Reads

• Anthropic is up next

• Berkshire Hathaway crosses $1 trillion market cap

• KKR faces lawsuit over $500 Million payout to co-founders

• Private Equity’s hot deal summer

• Vista Equity partners transfers ownership of Pluralsight to lenders

• Scotiabank acquires 15% Stake in KeyCorp for $2.8 billion

• Woodside Energy acquires OCI Global Clean Ammonia Project for $2.4 billion

_{DEAL OF THE MONTH}

SpaceX Clicks on Cursor

SpaceX priced its IPO at $135 a share on June 12. It opened trading at $150 with many retail investors piling in driving the price even higher. Four days later it touched an all-time high of $225.64, briefly making it the fourth most valuable company in America and adding roughly $740 billion in market cap in less than a week.

Quite the first week on the job.

On Tuesday June 16, riding that high, SpaceX announced it would acquire Cursor, the AI coding tool, for $60 billion in an all-stock deal.

The math was, for about four days, beautiful. SpaceX’s stock had gained more than the entire cost of the acquisition before the acquisition was even announced.

The company didn’t touch its $86 billion IPO war chest. It didn’t take on debt. It paid for the whole thing with paper that had appreciated faster than the deal could be signed. Buying a $60 billion company felt, briefly, like finding money in an old coat.

Then the coat got a lot lighter.

By June 22, SpaceX shares had fallen 16.4% in a single session, their worst day since the IPO, closing at $154.60. 

The trigger was the company confirming a $20 billion bond offering (later upsized to $25 billion) to refinance the bridge loan it took out to fund the xAI merger back in February.

Translation, for anyone who hasn’t priced a corporate bond before: the company that just raised $86 billion eleven days earlier needed to go raise debt too, and the market read that as “hmmm, maybe this is not the money tree we were promised.”

The stock has kept falling faster than a SpaceX rocket en route back to Starbase.

SPCX has fallen more than 30% from its peak and was trading below its IPO opening price of $150, at one point dipping under $147. A stock that had, for one glorious week, made everyone who bought at the open feel like a genius, was now making them feel like they’d bought a timeshare.

The same people who complained about not getting their full allocation were probably thanking their brokers for limiting the exposure.

The $60 billion Cursor deal that looked like found money on June 16 now represents a meaningfully larger chunk of the company than it did when the ink dried. Funny how that works.

To be fair to SpaceX’s balance sheet, the bond sale itself went well. The $25 billion offering drew $89 billion in demand, a 3.5x oversubscription that is one of the largest investment-grade order books ever recorded.

Moody’s, Fitch, and S&P all assigned investment-grade ratings on June 18, the same day the stock fell nearly 4% anyway, which is the bond market and the equity market looking at the exact same company and reaching two completely different verdicts. Bond investors (and the ratings agencies) clearly believe Starlink’s cash flow can service the debt.

Equity investors are apparently still doing math on the xAI side of the ledger, where the operating loss alone in 2025 was $6.355 billion, enough to wipe out the entirety of Starlink’s $4.4 billion operating profit and then some.

Worth noting: every single one of xAI’s 11 co-founders left the company before the IPO even happened.

Musk himself said in March that xAI “was not built right the first time around” and that he’s rebuilding it “from the foundations up.” So the company decided the correct next move was to spend $60 billion buying a coding tool to bolt onto the AI division that just lost its entire founding team. Bold strategy.

The Cursor deal itself still represents just 15x forward revenue and 3.4% dilution at the IPO valuation, which by AI-deal standards is almost conservative.

Cursor’s market share in AI coding fell from 41% to 26% between June 2025 and May 2026 as Claude Code and OpenAI’s Codex ate its lunch, so SpaceX is effectively buying a company that’s losing a war in order to give Grok a weapon in it.

One analyst, with admirable restraint, said simply: “I’m just not sure what else he wants to add to the portfolio.” At this point neither are we, but we assume there’s a slide deck somewhere with the word “synergy” on it forty separate times. Btw - odds are the next bolt-on to SpaceX is Tesla.

PitchBook's Franco Granda described the underlying mechanic well, before the stock decided to make his quote slightly less impressive: "The IPO gave SpaceX a valuation and a premium currency. SpaceX can now buy a company that size without touching cash, debt, or IPO proceeds, and the higher the stock runs, the cheaper the deal feels."

The corollary, which nobody said out loud at the time but is now extremely relevant, is that the lower the stock runs, the more expensive the deal feels too. The flywheel works in both directions.

SpaceX just found that out the hard way, eleven days into being a public company. That said, Elon has a trillion more dollars at risk on SpaceX than I do, so I'm sure he'll figure it out.

_{PRESENTED BY DILIGENCESQUARED}

How to Skip the $500k McKinsey Bill

Most funds skip commercial due diligence early on. Not because it wouldn’t help, but because at $500K to $1M a pop, they can’t afford to. So they wait until after initial IC and first-round bids, go deep on a few late-stage bets, and pray the thesis holds.

DiligenceSquared runs AI-native CDD at a fraction of that cost. Built by ex-Blackstone and BCG operators, it delivers investment-grade market and customer insight, so funds screen more targets earlier, reach conviction before the competition, and save the heavy spend for deals that already cleared the commercial bar.

Help your fund win deals faster. Book a demo today.

_{SPONSOR DEAL OF THE MONTH}

KKR Counts on Crowe

In a week full of healthcare deals, I chose to do a FIG deal. God, I hate healthcare. Anyway, let's jump into KKR attempting to buy free audit services for life.

On June 11, Crowe LLP agreed to sell a majority stake to KKR and co-investors in a deal valuing the firm at nearly $3 billion. 

Crowe is the 12th-largest accounting firm in the US, with roughly $1.39 billion in annual revenue and 9,000 employees domestically. Crowe’s existing partners retain a minority stake.

For most of the last century, the accounting profession ran on one unbreakable rule: outside capital does not get to own audit firms.

State licensing laws generally require that the people performing audits also own the firm doing the auditing, which has historically kept private equity out of the industry’s core business entirely. That rule is now being worked around at scale, and Crowe just became the latest firm to find the workaround.

The mechanism is something called an alternative practice structure. Crowe is splitting itself into two entities: a regulated audit and attest arm that stays owned by its partners, and a separate advisory and consulting business that KKR’s money actually buys into. On paper, the audit function remains untouched by private capital.

In practice, the firm that does both is now substantially funded, resourced, and influenced by a private equity firm with $25.65 billion in annual revenue across its own platform. It’s the corporate equivalent of saying you didn’t let your roommate move in, you just gave him a key, paid his half of the rent, and let him redecorate.

Crowe joins a list that is getting long fast. Baker Tilly, Grant Thornton, CohnReznick, and EisnerAmper have all taken PE money in recent years. Among the top 20 US accounting firms, only CLA and Plante Moran remain structurally untouched by outside capital.

CEO Steven Strammello told the Wall Street Journal that the competitive environment for accounting firms has been “evolving rapidly,” which is the diplomatic way of saying that if you don’t take the AI investment dollars, the firm next door will, and then they’ll out-automate you on audit and tax work that used to require armies of associates.

KKR’s Chris Harrington framed the deal as being about funding “next-generation client capabilities,” which is private equity speak for AI tooling and bolt-on acquisitions.

The more interesting read, from analysts covering the deal, is that KKR isn’t pricing Crowe like a sleepy mid-tier accounting shop. They’re pricing it like a finance-and-accounting outsourcing platform, the same tier occupied by Genpact, WNS, and Cognizant.

If that thesis holds, every top-20 accounting firm that hasn’t yet taken outside capital just became a more obvious takeover target, whether they want to admit it or not.

The deal also lands on top of a broader trend of private credit quietly funding the entire wave. KKR is reportedly in talks to raise roughly $1 billion in debt financing to support the transaction, with pricing conversations happening around SOFR+550 basis points.

That’s the same private credit machine that has funded most of the accounting industry’s PE consolidation since 2021, working in the background while the press release talks about AI and client capabilities.

_{THE FIRST TRILLIONAIRE} 

What Elon accomplished by building SpaceX is rare in financial history. Rockefeller at the height of Standard Oil, Carnegie, the Medici, none of them reached this scale. It is worth understanding how he actually got there.

In 2002, Musk sold his PayPal stake for roughly $180 million after eBay acquired the company. He was 30. Most people in that position would have banked the money and moved on.

Instead he put about $100 million into SpaceX, $70 million into Tesla, and $10 million into SolarCity, to the point where he was reportedly borrowing money to cover rent. Both SpaceX and Tesla were widely expected to fail, including by people close to him.

Aerospace engineers argued that reusable rockets were not economically viable. Even Neil Armstrong publicly voiced concerns about the shift toward commercial spaceflight. Investors warned that consumer EVs were a graveyard. He wrote the checks regardless.

SpaceX failed three consecutive launches before Falcon 1 finally reached orbit on its fourth attempt in 2008. By then the company was weeks from bankruptcy. There is a photo from that evening at the Kwajalein launch facility: Musk standing alone on the floor, looking at the rocket that finally worked.

Those decisions, made alone, with his own money, against the consensus of most experts in both fields, are a large part of what the $1 trillion figure represents.

(IPO by the numbers - continued below)

_{PRESENTED BY TOLTIQ}

The Great Irony in Private Equity

PE firms are all about efficiency. The industry exists to reduce capex and find cost synergies. Their AI use, however, is anything but efficient.

Most firms run diligence through generic chatbots, asking one frontier model to do everything, and this is the great irony. A team of four running parallel workstreams over three weeks will re-process the same materials repeatedly, paying ingestion costs each time. Every follow-up, every refinement is a fresh round trip through a frontier model. The tokens consumed in re-processing often dwarf the original query.

ToltIQ is a PE firm's version of reducing capex. The platform is built across multiple LLMs and optimizes token spend by delegating each task to the model that handles it best. And it indexes every data room once, so each query draws from that index instead of paying to re-process the same files.

Start treating your AI usage with the same care as you do your PortCo's operations. Upgrade to ToltIQ today.

_{THE IPO BY THE NUMBERS} 

SpaceX priced at $135 Thursday night. It opened Friday at $150 and closed at $160.95, the largest IPO in history. The price has since surged past $200 as of writing.

On the run since the IPO, Musk has added more to his net worth than Warren Buffett accumulated over his entire career, and an investor who bought at the IPO price would have made more in three days than the S&P 500 returned from December 2023 to today.

The book was five times oversubscribed. Retail investors alone placed over $100 billion in orders, for a company still reporting a net loss.

Speaking at Starbase this morning before the opening bell, Musk said: "I never thought we would actually get to this point. I thought we had maybe a 10% chance of success. I'm somewhat in shock."

For scale: Musk's net worth now exceeds the GDP of Taiwan, and the total NASA has spent since its founding in 1958. The second-richest person on earth, Larry Page, sits at roughly $288 billion. Musk is worth more than three of him. Even after a $1 trillion loss, he would remain the richest person on earth.

(The American Dream - continued below)

_{THE AMERICAN DREAM} 

Juan Hernandez, the SpaceX welder who became a millionaire

Elon Musk was born in Pretoria, South Africa, and came to North America at 17 with almost nothing, first to Canada and then to the United States. He co-founded a company, sold it, and put the entire proceeds into two more that everyone said would fail.

But Elon didn't just make a fortune for himself. He shared the equity. Roughly 4,400 SpaceX employees became millionaires, and about 400 of them are worth more than $100 million. They are not all founders and executives.

They are the welders, machinists, technicians, ship engineers, and factory workers who built the rockets with their hands. For context, Google's 2004 IPO and Facebook's 2012 IPO each created around a thousand millionaires. SpaceX did roughly four times either of them in a single day.

Take Juan Hernandez. He had never heard of SpaceX when a friend told him to apply in 2015. It was just another contract welding job, and it came with $10,000 in stock he didn't think twice about, since no employer had ever offered him equity before.

He kept buying more with every paycheck for ten years and worked his way up to supervisor. When the stock closed at $160.95 on Friday, his 6,500 shares were worth a little over a million dollars. He still plans to keep working, and he's teaching his three kids how to invest. His 16-year-old daughter already owns shares in a handful of companies. Asked what he'd say to Musk, he said he'd thank him for making it possible "for somebody like us, the cook or the electrician."

He is one of thousands. Maryellen Musselman was 27 when she joined in 2022, working on the ship that recovers rocket parts after they splash down at sea. André Lavoie, a 63-year-old engineer who left for Italy five years ago, is sitting on more than $28 million and plans to renovate a hotel and help his town switch from burning wood to cleaner heat. "I don't want to just die with a pile of money in the bank."

Trevor Hise joined as an intern in 2011 and held on for twelve years, accumulating more than 100,000 shares now worth at least $13.5 million.

And it goes all the way down the building. The equity reached cafeteria workers, baristas, and janitors who took stock instead of higher pay. Real estate agents near SpaceX's California headquarters say longtime employees in their thirties and forties are quietly house hunting, some looking to buy homes for their parents, still in disbelief that they can.

An immigrant with nothing who built something by betting on himself and the people around him, and took thousands of ordinary people up with him. That is the American Dream.

Whatever you think of the man, there is only one country on earth where this story is possible, and we should all feel proud to live in it.

Hats off.

_{THE NEXT DOT-COM CYCLE}

The Three IPOs of the Apocalypse?

SpaceX — $1.75 trillion, listing Thursday
The company that makes rockets is raising $75 billion in the largest IPO in history, more than double Saudi Aramco’s record from 2019. Up to 30% goes to retail investors, which is three times the industry standard and will generate more SpaceX hype content than you thought was humanly possible.

2025: $18.67 billion in revenue, $4.94 billion net loss, $41.3 billion accumulated deficit. Starlink is the real business: $11.4 billion in revenue, 63% EBITDA margins, 50% growth. Everything else: xAI, Starship, orbital compute, eventually Mars, is optionality. At $1.75 trillion you’re paying full price for Starlink and a very large premium for Elon’s ambitions. Whether that’s a good deal is a question for your investment committee.

Musk controls 42% of equity and 79% of votes, but don’t worry the remaining votes will be held by whoever Reddit nominates.

Anthropic — $1+ trillion, filed Monday
Anthropic filed confidential IPO paperwork on Monday, jumping ahead of OpenAI. It raised $65 billion at a $965 billion valuation just last week, making it the most valuable private AI company in the world and passing OpenAI for the first time.

Revenue run rate: $47 billion annualized, up from $10 billion twelve months ago. This is driven by Claude Code, which one unnamed firm reportedly spent $500 million on in a single month. Burn rate projected to fall to 9% of revenue in 2027. First profitable quarter expected Q2 2026, which is this quarter.

The Pentagon has designated Anthropic a supply chain risk after CEO Dario Amodei refused to give the military unrestricted access to its models. The dispute is working through two federal courts. For a company trying to IPO, pissing off the current administration is a bold move.

OpenAI — $852 billion (current), filing imminent
Still preparing its prospectus, ouch.

OpenAI is targeting a September listing with its revenue run rate at ~$24-25 billion. Projected to lose $17 billion in 2026. Not expected to reach positive cash flow before 2029-2030, by which point Anthropic will have been profitable for two years and will have sent Altman a very polite note about it.

CFO Sarah Friar has reportedly cautioned the company may not be ready for a 2026 listing. Sam Altman has been less cautious about timelines than his CFOs for the entirety of his professional career, which should tell you something.

_{PRESENTED BY WALL STREET PREP}

Enroll by June 15: AI for Business & Finance Certificate Program

The 8-week AI for Business & Finance Certificate Program from Columbia Business School Exec Ed + Wall Street Prep is built for non-technical roles and helps you cut through buzzwords, apply AI to your work, and lead adoption in your org. 

Enroll by June 15 to get $200 off tuition + use code BUYSIDERS for an additional $300 off. Get the program brochure today.

_{INDEX FUND EXTRAVAGANZA}

But Mom, What About My 401(k)?

For those of us who know that a 401(k) is a retirement plan and not the race the worst VP you know is trying to run, how these companies will play into index funds matters.

Index funds account for roughly 20% of the total US stock market holdings and ~62% of US households own index funds, so figuring out if these stocks that may or may not be overvalued are going to be dumped onto your lap in a few months is kind of an important question.

Starting with the big one - the S&P 500 dropped a definitive no.

S&P Dow Jones Indices confirmed today it will not change its eligibility rules for any of the three.

To join the S&P 500, a company still needs 12 months of public trading, GAAP profitability in the most recent quarter and the preceding four, and a free float above 10%.

S&P’s statement: “exceptions to the financial viability, seasoning, and investable weight factor requirements should not be granted solely based on market capitalization.”

S&P really told Elon that size doesn’t actually matter.

The consequence is obvious: none of the $11.8 trillion in S&P 500 passive assets is forced to buy any of these companies for at least a year.

Bloomberg Intelligence estimated fast inclusion would have generated $14 billion in forced buying for SpaceX, $8 billion for OpenAI, and $4.6 billion for Anthropic. That demand has been deferred, not eliminated. Every passive S&P 500 investor got an accidental reprieve, unless they own QQQ.

The Nasdaq-100, on the other hand, said a very enthusiastic yes.

Nasdaq changed its rules in May. Any company in the top 40 by market cap can enter the Nasdaq-100 after just 15 trading days post-listing, with the prior free float requirement eliminated entirely.

SpaceX lists June 12. Fifteen trading days later is approximately July 7, which is when QQQ, QQQM, and the $1.4 trillion in Nasdaq-100 passive assets are mechanically forced to buy.

Here’s the problem. SpaceX’s float at IPO is only 3-5% of total shares, roughly $52-88 billion of stock actually tradeable in the market. But under the new Nasdaq methodology, a company with less than 20% free float is weighted at up to three times its actual float, capped at 15% of the index.

That means SpaceX could be weighted at 9-15% of the Nasdaq-100 despite having a float worth $88 billion at most, once again - mostly managed by Reddit.

QQQ alone manages $400 billion. The total Nasdaq-100 ecosystem is over $1.4 trillion. You are forcing hundreds of billions in passive buying into an $88 billion float. On day 15. With essentially no price discovery.

The closest analogue is Tesla’s S&P 500 inclusion in December 2020. Tesla ran 70% in the month before the inclusion date, then gave most of it back once the passive flow exhausted. SpaceX has a much smaller float and a much faster inclusion window. The front-run-and-exhaust pattern could be significantly more violent.

Add SpaceX, OpenAI, and Anthropic to the Nasdaq-100 at target valuations. You’re adding roughly $3.57 trillion into an index worth $28-29 trillion — a 12% increase in total index market cap.

Three companies that combined are set to lose more than $25 billion in 2025.

Current Nasdaq-100 top 10: accounts for 47.45% of the index. Post-inclusion, that concentration figure almost certainly rises. Three unprofitable companies could represent 10-15% of the index your retirement account tracks.

The passive investing revolution was supposed to protect ordinary investors from speculative concentration. Nasdaq just decided that for the biggest companies, those protections don’t apply.

The S&P 500 held the line this morning. The Nasdaq did not.

If you hold QQQ or QQQM, you are about to own significant positions in SpaceX, OpenAI, and Anthropic. Not because you chose to. Because Nasdaq changed its rules.

_{THE AI IPO RACE}

Anthropic Claims its Next Victim

After dismantling the entire SaaS industry in a matter of hours, Anthropic just claimed its next victim - OpenAI.

This shouldn’t come as a total surprise. I assume most of you reading use Claude at work - don’t lie - and not ChatGPT (and certainly not copilot). It appears that investors are finally seeing that too. Whether or not they are too late remains to be seen.

Since people have cared about the AI race, OpenAI has been the clear leader in terms of valuation. Until last week when Anthropic announced a $65 billion Series H at a $965 billion valuation, led by Altimeter, Dragoneer, Greenoaks, and Sequoia. The company also released Claude Opus 4.8 the same day.

For the first time, Anthropic is worth more than OpenAI. OpenAI’s last confirmed valuation was $852 billion, set when its $122 billion round closed in March. The gap is $113 billion. Both numbers are probably stale already but that’s life in the private markets.

Twelve months ago, OpenAI was at roughly $10 billion in annual revenue and Anthropic was at $1 billion. The idea that Anthropic would catch OpenAI seemed like something you told your IC to justify your outrageous mark to market.

OpenAI ended 2025 at $13 billion in actual annual revenue, reached a $24 billion run rate by late March 2026, and is now generating approximately $2 billion per month. That is one of the fastest revenue ramps in the history of enterprise software. It is also, it turns out, not fast enough.

Anthropic ended 2025 at $9 billion in annualized run rate. By February it was at $14 billion. March, $19 billion. April, $30 billion. This week: $47 billion. The company went from $10 billion in annual revenue to a $47 billion run rate in approximately twelve months. Dario Amodei told investors the growth felt “too hard to handle.” Normal numbers would return soon, he predicted. They have not. Dario is doing fine.

The driver is Claude Code, the AI coding assistant that has quietly become the product every engineering team’s budget is being routed toward. One firm reportedly spent over $500 million on Claude in a single month. We don’t know which firm. We do know their CFO had questions.

Revenue is an exciting number, but let’s not kid ourselves that either of these companies actually earn a profit. Burn rate is the number that really matters. After all, how long can $100 billion last? Longer than you with your Hinge date, hopefully.

OpenAI is projected to lose $17 billion in 2026. Its burn rate stays at 57% of revenue through 2027. The company has committed approximately $1 trillion to infrastructure over the next several years, because Sam Altman’s view is that the risk of underinvesting in compute is greater than the risk of overinvesting. He may be right, and he is certainly spending as though he is.

Anthropic’s burn rate should fall to one-third of revenue in 2026 and 9% by 2027. The company projects positive cash flow by 2028. OpenAI projects positive cash flow by 2029 at the earliest, by which point Anthropic should have been profitable for a year and will have sent Altman a very polite note about it.

This purely comes down to business strategy. Anthropic is an enterprise company with consumer products. OpenAI is a consumer company building enterprise products. 80% of Anthropic’s revenue comes from businesses. OpenAI has 900 million weekly active users, roughly 5% of whom pay anything at all. Serving 895 million free users is very expensive. It is also, arguably, an extraordinary distribution advantage. The two companies have made genuinely different bets and both are, for now, winning in their respective fields.

Anthropic is also deliberately not building image generation, video, or the other compute-hungry modalities that eat margin without clear enterprise ROI. This is either disciplined strategy or the AI equivalent of not getting a gym membership so you don’t have to pay for one. The revenue trajectory suggests the former.

It’s not all sunshine and computer generated rainbows for Anthropic though. Anthropic is fighting a Pentagon designation as a supply chain risk, the result of CEO Dario Amodei refusing to grant the military unrestricted access to its models. Defense Secretary Pete Hegseth made the designation in February after a public clash with Amodei. The dispute is working through two federal courts.

This is not ideal timing for a company trying to IPO.

OpenAI does not have this problem, partly because its relationship with the government is warmer and partly because nobody has tried to designate it a security risk yet.

Both are targeting a listing this year. Anthropic confidentially submitted its draft S-1 to the SEC yesterday, while OpenAI filed its own confidentially after beating Elon in court. The question now: who hits the market first?

The company that files first sets the narrative. The company that files second gets compared to it for the rest of the roadshow. Both could be doing this in the same quarter as SpaceX.

If Anthropic really is better than OpenAI financially, then going second could make answering roadshow questions easier and going first would make OpenAI’s roadshow brutal, so either way it may be a net win for Anthropic.

Q4 2026 is going to be a good quarter to be an ECM MD. A less good quarter to be an ECM analyst.

_{PRESENTED BY WALL STREET PREP}

The Best Investment You'll Make in Your Finance Career Takes 8 Weeks

The Wharton Online + Wall Street Prep Private Equity Certificate Program gives you the same hands-on training used inside Blackstone, Thoma Bravo, and Carlyle.

Earn your certificate in 8 weeks, online, without putting your career on hold.

• Live office hours with Wharton Business School faculty and guest speakers who have actually closed the deals you're reading about

• A 5,000+ network of finance professionals, plus lifetime access to events and meetups

• A Wharton Online certificate (proof that you've done more than skim the KKR annual report and call it research)

The June 8 cohort is filling up. Use code SQUEEZ to save $300 and secure your spot.

_{INTERNATIONAL DEAL OF THE MONTH}

DoorDash Goes International

Uber wants to buy Delivery Hero. It just doesn’t want to say so in writing.

Here is what has actually happened this week. Uber now holds a 36.83% economic stake in the Berlin-based food delivery company. Its voting rights are capped at 24.99%. German law requires any acquirer that crosses 30% of voting rights to make a mandatory cash offer to all shareholders.

Uber has deliberately structured its position to sit just below that threshold, owning more than a third of the company economically while keeping enough of the stake in instruments and derivatives to avoid triggering the mandatory bid.

This is not an accident. This is a very well-compensated German law firm earning its fee.

The timeline is worth understanding. A few weeks ago, Uber held roughly 7% of Delivery Hero. It quietly built to ~19.5% through open market purchases, becoming the company’s largest shareholder. The following weekend, Bloomberg reported Uber was exploring a full takeover.

Uber’s own shares fell 1.6% on that news, because markets were not thrilled about Dara Khosrowshahi paying $11.6 billion for a money-losing German food delivery company. With SpaceX getting ready to IPO at almost $2 trillion, money losing is irrelevant at this point.

Delivery Hero then confirmed it had received a formal offer from Uber at €33 per share, a price with literally zero premium to the prevailing share price. Uber offered to buy a company for exactly what it was already worth, apparently on the theory that the owners might not notice, but of course, they did.

Several institutional investors made clear they wouldn’t engage below €40 per share. Delivery Hero’s stock promptly traded up to a 52-week high of €39.79. The market had opinions.

Then on May 27, Uber bought Aspex Management’s stake at just under €40 per share, paying 21% more than its own public offer, in the secondary market, on the same day it was officially bidding €33. Uber’s public offer: €33. Uber’s actual transaction price: €40.

This is unusual. Most acquirers at least wait a few days before paying their own target price while publicly offering less.

The internal ceiling Uber has reportedly shared with advisors is €38. It has already paid €40 once. Nobody on Delivery Hero’s shareholder register is confused about where this settles.

Delivery Hero operates in 70+ markets (Middle East, Asia, Europe) under brands including Talabat, foodpanda, and Hunger Station. In most of those markets, Uber Eats either doesn’t exist or is losing. DoorDash has been expanding internationally and is circling the same assets.

If Uber doesn’t buy Delivery Hero, DoorDash might. And if DoorDash buys Delivery Hero, Uber’s international food delivery business gets substantially more complicated.

Prosus, a Dutch tech investment firm, has been ordered by the EU to reduce its Delivery Hero stake below 10%, which means a large block of shares is coming to market. Uber can buy that block and strengthen its position. So can DoorDash. Whoever moves faster controls the process. But it’s also European summer, so fast means by September.

Delivery Hero CEO Niklas Östberg is stepping down in Q1 2027 following pressure from major shareholders for a strategic review. A company in strategic review with a 36.83% anchor shareholder and a departing CEO has, functionally, already started a sale process and identified its own synergies.

The question is what the price is. The current answer from Delivery Hero shareholders: at least €40.

_{NEWS ROUNDUP}

Top Reads

• Likelihood of a SpaceX / Cursor acquisition

• Most Anticipated IPOs of 2026

• Dell Wins ~$10 Billion Government Contract

• Inflation at Highest Level in Years as Warsh Steps In

• Jamie Dimon Leans in On AI

_{NEWS ROUNDUP}

Which Bulge Bracket has the Best Mental Health?

Now you can find out. We just rolled out mental & physical health league tables on Wall Street 360, the first time the Street has ranked firms head-to-head on culture, not just comp.

Filter by:

• Bulge Brackets

• Elite Boutiques

• Middle Market

• Private Equity

• Country, city, or industry

The results are eye-opening. Some firms with the loudest reputations rank middle of the pack. Others you wouldn't expect are quietly outperforming. And a few elite boutiques are putting some bulge brackets to shame.

See where your firm stacks up.

_{DEAL OF THE MONTH}

The Largest IPO Ever

Listen, when the biggest IPO of all time hits the tape, you take a second and read the prospectus. It’s 300 pages with a decent amount of SEC nonsense and literally every bank other than Jefferies is on the cover, but it’s the first glimpse us normal people get behind the curtain of SpaceX.

SpaceX was founded in 2002, two years before Meta (which was then still the Facebook), a year before Tesla, three years before YouTube, and just two years after ICE (the company that owns the NYSE).

This means SpaceX has stayed private for almost 24 years. That’s longer than my analyst has been alive, but somehow still shorter than the number of years my intern has known he has wanted to be a banker for.

Before we dive into the details, this IPO could push Elon to become the world’s first trillionaire. This would make his net worth larger than the market cap of all but 10 public companies in the US and roughly equal to the market cap of Berkshire Hathaway.

Given 1 trillion feels like a made up number, here is some simple math for you to understand scale. 1 million seconds is about 12 days, 1 billion seconds is about 32 years, and 1 trillion seconds is 318 centuries.

Say what you will about Elon, but the man has realized the American dream and then some. Hats off to him no matter how this shakes out. Btw - if your family office needs a meme guy or a newsletter writer, the OWS team can start tomorrow.

Preamble aside, let’s dive into the biggest IPO in history.

Part I: The Prospectus Menace

Before we get into the numbers, you need to understand what you’re buying, because the S-1 does something that will confuse a lot of people.

SpaceX did not just file an IPO for a rocket company. It filed for a rocket company, a satellite internet business, an AI lab, and a social media platform, all rolled into one filing, all retroactively presented as if they’ve always been one company. I assure you, they have not.

The filing has been recast to include xAI (formally acquired February 2026) and X, the social media platform formerly known as Twitter (absorbed by xAI in March 2025).

Because Elon owned all of them, the accountants treat it as one entity across all historical periods, something I am sure they did not do in draft 1 which caused some headaches.

This means the 2024 and 2023 “SpaceX” numbers you see in the prospectus already include xAI and X, even for years when SpaceX didn’t own either.

The three segments are Space (rockets, Starship, NASA, defense), Connectivity (Starlink), and AI (xAI, Grok, X).

Every headline about SpaceX’s losses is a headline about the AI segment. What, you don’t use Grok? When you strip that out, the rocket and Starlink business looks very different. So just keep that in mind as we go forward.

Part II continues below…

_{PRESENTED BY TOLTIQ}

The Private Markets Outlook from Q1

Let’s get into the key AI developments and trends we saw in Q1, according to the ToltIQ research team.

• Tons of capital is flowing into the big name LLMs ($172B).

• As models improve, cost efficiency has also improved.

• Approximately 24% of enterprises are seeing ROI on their AI implementation.

Here are the consistencies of the 24%:

• Their implementation is CEO-led: Top-down AI accountability drives results.  

• They start specific: They select a use case, refine it, then scale. 

• They measure outputs: They track which deals used AI and how it drove value.

The simplest way to implement AI into your PE workflows? ToltIQ. Built by KKR’s former CIO, it is the tool leading firms are choosing for their AI. Through ToltIQ, deal teams are able to ingest entire data rooms, interrogate them in natural language, and get answers with every source cited and defensible at IC. It then drafts the tearsheet, memo, and LP update in your firm's format.

Explore the platform today.

Part II: Attack of the AI

The 2025 figures are astonishing. Revenue of $18.67 billion translating into Adj. EBITDA of $6.6 billion and a net loss of $4.9 billion with CapEx of $20.7 billion

The accumulated deficit as of March 31, 2026 is $41.3 billion, which is more than the combined accumulated losses of Uber, Airbnb, Rivian, and Didi at the time of their respective IPOs.

So you are probably asking - why on earth is a company that lost $4.9 billion worth over $1 trillion? Well, the answer is because Elon lives on Mars, obviously.

Assuming you are asking that question, which is fair, I would argue you have to dive a little deeper. Don’t ask why it’s worth $1 trillion, ask why it is losing so much money. The answer will harken back to what I mentioned earlier - SpaceX isn’t just rocket ships and satellites.

Part III: Revenge of the Profit

Starlink generated $11.4 billion in 2025 revenue, which was 61% of the entire company, with $4.42 billion in operating income and an adjusted EBITDA margin of about 63%.

That’s right - a satellite internet business that didn’t commercially exist before 2020 is now generating over four billion dollars in operating income with 63% EBITDA margins, still growing at 50% year on year.

Starlink ended 2025 with roughly 9.2 million subscribers across more than 9,500 satellites. In Q1 2026 alone it did $3.26 billion in revenue and $1.19 billion in operating income.

Now here’s where it gets interesting. The 9.2 million subscribers are mostly residential consumers paying somewhere between $120–$250 a month. That’s the established base. The growth story is in the segments where Starlink charges dramatically more.

Maritime customers pay roughly $34,000 per user per year. Aviation customers pay roughly $300,000 per user per year. Both segments have less than 1% market penetration today. If you were trying to find a market with room to run, you just found it.

The next phase of Starlink is V3 satellites, each carrying one terabit per second of downlink capacity, a generational jump from the current fleet. The S-1 confirms V3 deployment on Starship begins in the second half of 2026. This is why Starship matters to the Connectivity story. The two are not separate bets. One is the truck that delivers the other.

One real risk: average revenue per subscriber is declining as Starlink expands internationally and into lower-income markets. The path to $20 billion in Connectivity revenue requires high-ARPU enterprise and government segments to grow faster than consumer ARPU is falling.

Part IV: A Long-Term Hope

The launch business did $4.07 billion in 2025 revenue on a $657 million operating loss, driven almost entirely by Starship R&D.

SpaceX completed 165 orbital launches in 2025 (its sixth consecutive annual record) with 100% payload delivery success on Falcon 9. Since founding: roughly 7,400 metric tons to orbit, 99%+ mission success rate, 650+ orbital launches, more than 540 on flight-proven reused rockets.

Launch revenue growth has slowed. Volumes and prices were roughly flat between 2024 and 2025. SpaceX holds about 85% of the U.S. orbital launch market. There is not a lot of share left to take. The growth lever here is Starship, which is the million pound gorilla in this entire prospectus.

Starship consumed $3 billion in R&D in 2025 and $930 million in Q1 2026 alone. The filing says Starship is designed to reduce the cost of reaching orbit by 99% or more versus the historical average. SpaceX is launching Starship V3 the day after I am writing this, but if it works, that is the best-timed product demo in IPO history. If it doesn’t, the risk factors on page 26 are going to age very badly.

Starship delays are listed at the very top of the S-1’s 36 pages of risk factors: if Starship doesn’t achieve commercial scale, Starlink V3 slips, orbital AI compute doesn’t happen, the Artemis lunar contract is at risk, and the launch economics story stalls. The upper half of the $1.75 trillion valuation has Starship as load-bearing infrastructure.

Part V: Grok Strikes Back

xAI plus X generated $3.2 billion in 2025 revenue and an operating loss of approximately $6.4 billion. That means that xAI spent 3x its revenue in one year. I mean, no wonder Elon had to merge it in with SpaceX…that is not a viable operating model.

Most of that $3.2 billion in revenue comes from X’s advertising business, not from Grok.

The actual AI product — Grok — grew revenue about 22% in 2025. For context, Anthropic grew over 1,000% in the same period. OpenAI grew roughly 300%. But yea, we knew that the only reason anyone used Grok was for NSFW reasons or because they wanted to be on the IPO.

The segment is burning approximately $14 billion annually in cash. The math is almost exact: every dollar Starlink earns in operating profit gets consumed by the AI segment.

The prospectus also discloses that SpaceX faces litigation from the absorption of xAI and X that will likely cost $530 million. Good thing xAI generates enough profit to absorb that.

The AI bull case is not Grok subscriptions. It is orbital compute. SpaceX is already shifting some launch capacity from Starlink toward transporting AI servers into orbit.

The thesis is that reusable rockets make space-based AI infrastructure economically viable for the first time. The S-1 claims a total addressable market of $28.5 trillion, of which $26.5 trillion is AI, including $22.7 trillion in enterprise applications. I mean sure, and the TAM for my newsletter is $30 trillion. Don’t ask how I got there.

Part VI: Return of the Valuation

SpaceX is targeting $1.75 trillion on a raise of $70–75 billion, the largest IPO in history by a very wide margin. Saudi Aramco raised $35.4 billion in 2019. SpaceX is trying to raise double that.

At $1.75 trillion against $18.67 billion in 2025 revenue, you’re at roughly 94x trailing sales. Against $6.58 billion in Adjusted EBITDA, you’re at roughly 266x. I simply report the news and cannot give investment advice but just to be clear those numbers are…high.

The honest valuation framework looks something like this. Starlink alone, using satellite broadband comps with a growth premium, is defensibly worth somewhere around $500–700 billion. The launch business adds a meaningful chunk on top. PitchBook puts the total fair-value range at $1.1–1.7 trillion. Morningstar called $1.5 trillion “expensive and risky, but not irrational.”

This means somewhere between $500 billion and $1 trillion of the $1.75 trillion target rests on Starship reaching commercial scale, xAI generating real AI revenue, and orbital compute working as described. These are things that do not yet exist as revenue lines in the filing.

The share price range hasn’t been set yet. It appears in the amended prospectus before June 4. What we do know: SpaceX did a 5-for-1 stock split on May 4, and is reserving up to 30% of shares for retail investors, about 3x the industry standard. That’s a very smart way to build a million loyal shareholders who are going to tell all their friends about it. Diamond hands coming back?

Part VII: Elon Awakens

Naturally, Elon isn’t just going to hand his pride and joy over to a bunch of Reddit apes…sorry, investors…without retaining some control.

Class A shares with one vote per share are being offered as part of the IPO. The Class B shares - which Musk holds - have 10 votes per share and the Class B shareholders elect a majority of the board. So Musk owns 42% of the equity but has 79% of the votes.

The S-1 on this: “Mr. Musk will control the voting power over the selection of our board. As a result, Mr. Musk will have the power to control the outcome of matters requiring shareholder approval, including election of all our directors.”

SpaceX intends to rely on “controlled company” exemptions from Nasdaq’s governance requirements; no requirement for a majority-independent board, no requirement for a fully independent comp or nominating committee.

You’re buying an economic stake in a company where the governance outcome of every shareholder vote has already been decided. This is disclosed, in plain English, on page one of the prospectus. I would say read it before your compliance officer does, but given the entire street worked on the IPO, it’s probably on every restricted list anyways.

The Bottom Line
I didn’t get through all nine Star Wars movies - my bad - but also the last 3 should hardly count, what a way to kill a franchise… kind of like merging a social media and AI company into a profitable infrastructure business.

Stripping out xAI, you have one of the best infrastructure businesses built in the last two decades.

Starlink is genuinely extraordinary, and the launch business is operationally unmatched. Together they generated roughly $3.8 billion in combined operating income in 2025 before the AI segment consumed all of it.

Add xAI back in and you get: a $14 billion cash burn, a social media platform, an AI model that is growing at 22% in a market where competitors are growing at 300–1,000%, and a vision for orbital computing that is compelling and not yet real.

At $1.75 trillion, you are paying full price for what exists and a significant premium for what Musk says comes next. The historical record on that: Falcon 9 took longer and cost more than the original plan. Starlink took longer and cost more than the original plan. Tesla nearly went bankrupt. They all eventually worked.

Whether that pattern holds at 94x revenue is a question for your IC.

_{STRATEGIC DEAL OF THE MONTH}

Half Cash, Half Stock

I mean, do I really need to cover this one? I feel like Ryan Cohen kind of nailed it on CNBC. It’s half cash, half stock. I will take no further questions on the subject.

For those of you who somehow didn’t see the viral clip of GME CEO Ryan Cohen going on CNBC, I will attempt to explain a better understanding of the deal presented than the CEO of the acquiring company.

GameStop - you know the famous meme stock that ripped over COVID - made an offer to acquire eBay for $125 per share, or $55.5 billion, a lofty 46% premium to eBay’s unaffected share price.

Before we even get into whether this deal makes any sense or if Ryan Cohen is just trying to hit the EBITDA and market cap targets in his comp package, we have to take a look at the consideration.

As Cohen reminded us (and Andrew Sorkin) several times, it’s simple - the proposal is half cash, half stock.

Once you exit the C-Suite and come down to reality though, you have to check if that math maths. GME has lined up a $20 billion financing via a highly confident (but not binding) letter from TD and has $9.4 billion of cash on its balance sheet. The remaining ~$25.5 billion would have to come from an equity issuance, which for a company with a $10 billion market cap is…tough.

Taking the obvious financing hurdles aside, the deal is expected to be accretive in year 1 (says every banker ever), driven by $2 billion of cost synergies. Cohen wasn’t shy about the LinkedIn updates for eBay employees to generate these synergies, basically stating he could run the company from his garage.

GME actually already built a 5% stake in eBay as a beachhead and recently filed its 13D and HSR documentation to disclose the stake publicly, so the transaction isn’t totally out of nowhere even if it may seem to be at first glance.

Cohen may seem insane, but what is interesting is that he receives no cash comp at all. His entire compensation package with GME is in shares (and a lot of them) but they are tied to ambitious EBITDA and market cap metrics…which could come as a result of this transaction.

Could you make the argument that this comp package incentivizes empire building and doing bad deals to hit your targets? Sure, it absolutely does. That being said, Cohen does have a pretty solid track record at GME so far (even though that still doesn’t quite justify his run for eBay).

GME’s share price fell ~10% on announcement, and eBay’s share price popped to $109, which may be a far cry from the $125 being offered but was a welcome bump to eBay shareholders nonetheless.

This deal really puts into perspective the hard work you do as a banker or a private equity professional. You can spend hours over engineering a model to find the price that you can afford to pay to hit the required return threshold…and lose to a guy who can’t even explain his deal on TV but just has the guts to shoot his shot and figure it out later.

There is merit in both approaches and it’ll be interesting to see if cooler heads prevail for eBay or if Cohen’s crazy plan actually works. If it means anything for your internal analysis, Michael Burry has sold his GME shares as a result of the deal. And before you ask, yes - the guy who bet on the US economy blowing up in 2008 was really long GME until now.

_{PRESENTED BY F2}

Product Spotlight: Institutional Knowledge

The analysis behind the deals in this newsletter comes from nearly a decade in Private Equity and Investment Banking. Looking through hundreds, maybe thousands, of deals is what brings you the insights (and humor) you love for the most interesting deals of the last several weeks. Each prior success, risk, and unique deal nuance is baked into every Buysiders article.

F2 has built that, but on an institutional level, with their newest launch: Institutional Knowledge.

Their platform turns prior deal work, including memos, models, diligence files, IC materials, CRM notes, and portfolio data, into a queryable intelligence layer that teams can use on every future deal.

While most firms are experimenting with simple AI wrappers, firms using F2 are building firm knowledge into a proprietary asset that compounds with each win and loss.

Test out F2 today.

_{DEAL OF THE MONTH}

Shell Says Let’s Go Back to Canada, eh?

Shell CEO Wael Sawan has spent the better part of three years telling investors that Shell is a “capital discipline” company.

Value over volume. Organic growth. No big acquisitions. I think we found the limits of capital discipline the other week.

On April 27, Shell announced it would acquire ARC Resources, Canada’s largest pure-play Montney producer, in a $22 billion deal. That’s $32.80 per ARC share, split 75% Shell stock and 25% cash, representing a 27% premium to ARC’s closing price on April 24.

So why Canada?

ARC produces roughly 370,000 barrels of oil equivalent per day from the Montney shale formation in British Columbia and Alberta, think like Canadian Texas, one of the most economic shale basins in North America.

About 40% of ARC’s production is liquids, which drove 70% of its revenues last year. Somewhere someone at Deloitte is gearing up to charge Shell a few mill to tell them to do more of that production and reduce costs.

But the real story is LNG.

Shell is already the largest owner of LNG Canada. The country’s first LNG export terminal, which opened in Kitimat, BC last summer and ships liquefied gas across the Pacific to Asia.

ARC’s Sunrise production area has a direct pipeline connection to that terminal via Coastal GasLink. So what Shell is really buying is upstream gas supply to feed an LNG facility it already owns, in a country it spent the last decade divesting from, right as Qatar’s LNG exports are offline because of a war.

Sawan put it best on the investor call: “Don’t look at this as the acquisition of an E&P with a predominantly gas portfolio, look at it as a liquid-rich addition to Shell with an upside of LNG that we are uniquely positioned to be able to unlock.” That’s a very elegant way to say: we own the terminal, we need the gas, and here we are.

Shell guided ~$1.5 billion in incremental free cash flow per year from ARC once integrated, rising to ~$2 billion beyond 2030 if LNG Canada Phase 2 gets a final investment decision, which Sawan (obviously) described as increasingly likely.

Year one synergies are guided at $250 million and the deal bumps Shell’s production CAGR from 1% to 4%. I am sure those numbers looked great in the pitch deck but hopefully Shell can actually deliver them.

This deal also quietly kills the “Shell buys BP” conversation that has haunted every energy conference for two years. Hard to acquire a second major when you’re still issuing 228 million new shares to close the first deal.

ARC’s board approved it unanimously (shockingly). The deal requires Canadian regulatory approval, which Shell doesn’t see any showstoppers in but is not always straightforward. The Canadian government, and particularly British Columbia, is left leaning, which typically is not business friendly even before you add the pillaging the earth for resources bit.

All in all, Shell threw capital discipline out the window, reversed course on Canada, and did it at almost exactly the right time...maybe. If Trump decides the end the war tomorrow then Shell may be smashing the buy button at the peak. Either way, this deal is either very smart or very lucky.

_{NEWS ROUNDUP}

Top Reads

• Will GameStop acquire Ebay?

• Regis Resources and Vault Minerals agree to merger of equals

• 3 year old PE fund acquires Amex Global Business Travel

• Burry exits GameStop over eBay deal

• OnlyFans sells 16% stake to Architect Capital at $3.15 billion valuation

_{DEAL OF THE WEEK}

Amazon Looks to the Stars

I am closer to being as rich as Jeff Bezos than Jeff Bezos is to being as rich as Elon. That said, if I suddenly came into $200 billion, the odds of me writing a newsletter at 10pm on a Tuesday would be pretty low.

While that may not be the genesis for this deal, watching the guy who dethroned you as world's richest man prepare to IPO his space and satellite and everything-else company at a $1.75 trillion valuation probably makes you want to do some M&A.

Amazon decided that if you can't beat SpaceX, you should join them by buying the satellite infrastructure Apple is already paying for.

Amazon is acquiring Globalstar, the mobile satellite company powering the Emergency SOS feature on every iPhone since the 14, for ~$11.6 billion, or $90 per share in cash or stock.

Thermo Funding, which controls 57.6% of Globalstar’s votes, approved the deal by written consent, which is corporate governance shorthand for “we didn’t need to ask anyone else.” No shareholder vote required.

There’s a downward adjustment of up to $110 million if Globalstar misses certain satellite milestones before close, which is the kind of clause you negotiate when you’re buying a company that still has satellites left to build.

There’s also an adjustment if Amazon stock trades above $280/share before closing…which could happen I guess.

The premium is ~117% to Globalstar’s unaffected price from late October 2025. Amazon’s stock rose 3.8% on announcement day, adding roughly $125 billion in market cap. The market decided this $11.6 billion acquisition was worth more than ten times its price tag before Jassy finished his coffee.

What Amazon actually bought is spectrum and a customer. Globalstar operates 24 satellites with globally harmonized Band 53/n53 spectrum, the frequencies that enable direct-to-device satellite connectivity, the part of the market Starlink hasn’t cracked and Apple desperately needs.

Amazon simultaneously announced a partnership with Apple to keep powering iPhone and Apple Watch satellite features while it builds out Amazon Leo, their answer to Starlink. Apple was Globalstar’s largest customer. It now becomes Amazon’s. Elon Musk is watching from a rocket.

The plan is to have voice, data, and full messaging over satellite running by 2028, at which point Amazon Leo will be a real product rather than a $200 billion capex line item with a hopeful name. Analysts are broadly supportive of the strategy and uniformly clear that nothing meaningful happens financially until 2028 or later. Which is the polite way of saying this is a very expensive option on the future.

SpaceX will kick off its IPO roadshow in June targeting a $1.75 trillion valuation. Jassy just made sure Amazon isn’t showing up to that conversation empty-handed. Whether that was prescient or expensive is a question for 2029.

_{PRESENTED BY ACRETRADER}

Farmland: A Real Asset for the Modern Portfolio

Invest in the one thing they aren’t making any more of: Farmland. These are real assets for a rising-cost world.

The Proterra AcreTrader Farmland Fund LP combines AcreTrader’s expert sourcing and management with the institutional power of Proterra Investment Partners. The Fund provides diversification across high-quality U.S. farmland assets, offering an open-ended gateway into an essential asset class that has historically provided consistent returns.

Why the Proterra AcreTrader Fund?

• Real Assets: Direct access to high-quality U.S. farmland via a tax-efficient private REIT.

• Portfolio Stability: Access a tangible asset with the potential to serve as an inflation hedge, based on farmland's historically low correlation to stocks and bonds.

• Historical Performance: Access an asset class that has delivered a 9.4% average annual return since 1997, with low volatility.*

Learn more here: AcreTrader.com

_{STRATEGIC DEAL OF THE MONTH}

Agnico Ealge Strikes Gold in the Arctic

Gold is at $4,800 an ounce. Agnico Eagle — already the world’s third-largest gold miner — has decided that now is the time to buy the entire Central Lapland Greenstone Belt of northern Finland. Not part of it. All of it.

On April 20, Agnico announced three separate transactions totaling roughly C$3.7 billion: the acquisition of Rupert Resources for approximately C$2.9 billion, Aurion Resources for C$481 million, and B2Gold’s 70% stake in the Fingold JV for US$325 million.

One press release, three deals, one very cold region of Europe. I guess the Corp Dev team only wanted to do one trip. Have to respect the efficiency.

The Rupert deal is the centerpiece.

Each Rupert share gets 0.0401 Agnico shares — worth approximately C$12.00 based on Agnico’s five-day VWAP on the TSX as of April 17 — plus a contingent value right worth up to C$3.00 in cash paid in three tranches over 10 years, tied to reserve and production milestones.

That’s a 67% premium to where Rupert was trading the day before the announcement. Agnico already owned 13.5% of Rupert, so this is mopping up the remaining 86.5%. Guess it helps when you know what you’re buying.

The reason anyone cares about a Finnish greenstone belt is Ikkari, Rupert’s flagship gold deposit, which sits roughly 50 kilometres from Agnico’s Kittilä mine, its largest European operation.

In mining, 50 km from an existing mine means shared infrastructure, shared processing facilities, shared labor pool, and a dramatically shorter path to production than building in the wilderness from scratch.

Jefferies estimates the three transactions are modestly accretive, adding roughly 2% to NAV per share with up to C$500 million in potential synergies.

The Rupert deal values the project at roughly 0.54x NAV, which is either a bargain or a fair reflection of the fact that Ikkari is still a development asset in a country where winter lasts eight months, like actually eight months not how it feels in NYC when you don’t see the sun from October to February.

The CVR structure is the honest part of this deal. Nobody knows when Ikkari will reach 5 million ounces of reserves, or when it will hit commercial production, or what gold is worth when it does. The CVR is Agnico’s way of saying: we’ll pay you more if it works out, and we’ve got ten years to find out.

At $4,800 gold and with a $110 billion market cap, Agnico has the currency and the conviction. Whether its shareholders have the patience is a different question.

_{INT’L DEAL OF THE MONTH}

Viva USA Rare Earths

Quick quiz: what do an EV motor, a wind turbine, a missile guidance system, and the robot that is definitely coming for your job have in common? They all require neodymium, praseodymium, dysprosium, terbium and probably 5-10 other iums to function.

China produces the vast majority of all four, has export restrictions on all four, and has made it very clear it understands exactly what that means. This deal is America’s answer.

On April 20, USA Rare Earth (Nasdaq: USAR) announced a definitive agreement to acquire Serra Verde Group, owner of the Pela Ema rare earth mine in Goiás, Brazil, valuing Serra Verde at approximately $2.8 billion. The consideration is $300 million cash plus 126.849 million newly issued USAR shares at a reference price of $19.95.

Pela Ema is the only scaled producer of all four magnetic rare earth elements outside Asia. It’s an ionic clay deposit, the same geology that dominates Chinese production, and processes at a fraction of the cost of the hard rock deposits that most Western rare earth projects are stuck with.

Serra Verde started commercial production in 2024, has already sunk over $700 million into construction and permits, and is targeting Phase 1 nameplate capacity of 6,400 metric tons of total rare earth oxide per year.

Projected run-rate EBITDA by end of 2027: $550–$650 million. Combined company EBITDA target by 2030: ~$1.8 billion.

If you’re wondering why a company trading at $19.95 a share can credibly project $1.8 billion in EBITDA four years out, the answer is the U.S. government. Yea, the same people who famously can’t balance a budget, are in crippling debt and misplace billions of dollars per year.

Serra Verde comes pre-attached to a $565 million DFC financing package that fully funds Phase 1 optimization and expansion through to positive cash flow.

It also has a 15-year, 100% offtake agreement with a special purpose vehicle capitalized by U.S. government agencies and private capital, covering all four magnetic rare earths with guaranteed floor prices plus 70% upside participation above those floors. Great use of taxpayer dollars.

Pro-forma liquidity for the combined company is approximately $3.2 billion. The U.S. government has committed approximately $1.8 billion in total capital to this supply chain. Washington isn’t a passive observer here; it’s a cornerstone investor with a very strong opinion about where rare earths come from.

The management transition adds credibility. Sir Mick Davis, former CEO of Xstrata, who turned a $500 million business in 2001 into a $65 billion mining major before selling it to Glencore in 2012, joins the USAR board.

Thras Moraitis, Serra Verde’s CEO and a former senior Xstrata executive, becomes President of the combined company. The Xstrata alumni have arrived in Oklahoma (probably not for long), and they brought the Brazilian mine with them.

This deal is industrial policy with an equity wrapper. The floor prices, the DFC facility, the government-backed offtake SPV, none of this is how commercial markets normally work.

It’s how you build a supply chain when the commercial market has already been cornered by someone who doesn’t like you.

USAR shareholders are getting a leveraged call option on whether the U.S. government keeps paying to solve its own rare earth problem. Given that both parties have very strong incentives to make it work, the floor is more credible than most.

_{NEWS ROUNDUP}

Top Reads

• SpaceX targets June roadshow aimed at $1.75 trillion valuation

• Wall Street deal fees surge in Q1 despite global turmoil

• GS posts record Q1 revenues driven by equities

• Livenation found guilty of illegal monopoly

• Meta’s Manus deal blocked by Beijing

• Possible US government investments 

_{DEAL OF THE MONTH}

Unilever Does Two Deals in Two Weeks

McCormick (the company responsible for your spice rack) announced it would merge with Unilever’s entire Foods business in a deal valued at $44.8 billion. The company behind Old Bay is absorbing Knorr, Hellmann’s, Marmite, and Colman’s.

The structure is a Reverse Morris Trust, which means Unilever spins its food division into a new entity, which immediately merges with McCormick. Since it’s just past April 15th, I have to remind you that RMTs are tax-free. So for the first time ever Unilever can say their lawyers earned every penny.

McCormick pays $15.7 billion cash and ~$29 billion in stock; post-close, former Unilever Foods shareholders own roughly two-thirds of the combined company. Citi, Goldman, and Morgan Stanley are all taking a bite out of the bridge financing.

The combined business does $20 billion in revenue with ~$600 million in promised synergies by Year 3.

Margins guided to expand from 21% to 23–25%. One of the major sources of synergy is something every banker cheers for, back office optimization. This means reducing duplicative headcount and a lot of LinkedIn updates coming soon.

McCormick stock dropped 6% on announcement day, which is the market’s way of saying “we believe you, but we’re nervous.” MKC was already down 26% year-to-date. Unilever also dropped 4%. I guess no one wins on announcement day anymore.

This deal marks the end of Unilever as a food company. After spinning out Magnum Ice Cream last year and now shedding the entire foods division, what remains is a pureplay beauty and personal care business… and a vitamins business as of a few days ago…

Nine days after announcing the $44.8 billion McCormick deal, Unilever agreed to acquire Grüns, a 32-month-old greens gummy startup, for a reported $1.2 billion. Man corp dev has been busy.

The reflexive take: “Unilever just sold food and immediately bought food,” is technically wrong, even if you do eat the gummies.

Grüns lives in Beauty & Wellbeing’s Wellness Collective alongside Liquid I.V., Nutrafol, OLLY and Onnit. Supplements aren’t classified as food at Unilever, which is the kind of distinction that only matters when a journalist calls your IR department at 8am. 

The McCormick perimeter was drawn around culinary flavor. Grüns is a daily nutrition gummy with 60 ingredients including organic spirulina, prebiotic fiber, and whatever adaptogens are in vogue this week.

What makes this interesting is the multiple arbitrage. Unilever sold slow-growth, private-label-exposed culinary brands at ~14x EBITDA. It bought a startup growing 600% year-on-year at ~4x sales. A start-up that doesn’t even use AI mind you. No wonder their share price is getting killed.

Both trades were shaped by the same force: GLP-1 drugs are terrible for packaged food and excellent for supplements. Unilever sold the victim and bought the beneficiary in the same fortnight.

Grüns was founded by Chad Janis, Stanford GSB, former Summit Partners growth investor, who launched in August 2023 after 25% of his MBA cohort tested prototypes. The company hit profitability at 14 months, crossed $300 million ARR by October 2025, and is now in 6,300+ retail doors including Walmart, Target, Costco and Ulta. 

Janis reportedly retained ~50% of the company, implying personal pre-tax proceeds north of $500 million on a business he started two and a half years ago. The Series B investors at a $500 million valuation in May 2025 are sitting on ~2.4x in under a year. That’ll play well in the fundraising materials.

So a divestment and an investment in 2 weeks? Presumably someone in Unilver Corp Dev is taking the rest of the year off.

_{PRESENTED BY F2}

Built for Speed (and Precision)

In investment memos and IC decks, the inevitable question is simple: Where did this number come from?

Tracking the origin of a figure inside a complex model or spreadsheet can take longer than the analysis itself, especially when sitting in front of everyone at IC.

F2’s Audit Mode is designed to make that process fast and error-free. Inside Excel, it can trace any value back to its exact origin, including the file, sheet, cell, formula, and source material behind it.

Instead of manually retracing calculations or rebuilding work, your team can audit numbers in seconds and move forward quickly.

According to F2’s clients, the workflow can reduce diligence and review time by up to 75%, while keeping every output fully auditable.

Take Audit Mode for a test drive today.

_{STRATEGIC DEAL OF THE MONTH}

Mining Goes to the Deep Sea

You’ve heard me talk about mining before but this time we have a new twist. Mining on land is so 2025. Now, we are going to the ocean floor for our rocks because that’s…environmentally friendly I guess? Just as long as no billionaires are planning on riding the first sub we may be in business.

American Ocean Minerals announced it would merge with Odyssey Marine Exploration (NASDAQ: OMEX) in an all-stock deal valuing the combined entity at ~$1 billion. The mandate: vacuum polymetallic nodules — clumps of nickel, cobalt, copper and manganese — off the bottom of the Pacific Ocean.

OMEX entered the day at $0.83 with a going-concern qualification and TTM revenue of $354,000. The announcement triggered a 110–150% intraday pop to $2.13 on 330% of average volume. It then gave most of it back. Analysts have not published price targets. The $1 billion figure is negotiated, not market-tested. Shocking.

The structure is a reverse takeover: AOM holders control the combined company, Odyssey provides the Nasdaq listing, and the legacy Mexican phosphate business spins into a liquidating trust for OMEX shareholders. Financing includes a $75.6 million convertible bridge, a $156 million PIPE, and up to $35 million in additional notes. 

Citi and Cantor Fitzgerald placed the PIPE, Cantor being chaired by Brandon Lutnick, son of Commerce Secretary Howard Lutnick, who oversees NOAA’s deep-sea permitting buildout. That detail appeared in zero press releases.

The management team is the actual signal. Chairman is Tom Albanese, ex-CEO of Rio Tinto and Vedanta, the most credible mining executive attached to any deep-sea venture. CEO Mark Justh is a former JPMorgan and Goldman banker who was already Odyssey’s lead independent director, which explains why OMEX became the vehicle. Mike Rowe of Dirty Jobs is a founding investor and brand ambassador, so maybe this company will get its hands dirty.

The thesis is regulatory arbitrage. The International Seabed Authority has failed to finalize a Mining Code for three years running. Meanwhile, Trump’s April 2025 executive order, NOAA’s January 2026 permitting rule, and a U.S.–Cook Islands critical minerals agreement have created a parallel licensing pathway that bypasses the ISA entirely.

AOM holds licenses across the Cook Islands EEZ and U.S.-regulated CCZ waters, covering 500,000+ km² and more than 1.4 billion tonnes of inferred nodule resource. The bet is that the domestic U.S. permitting track closes before the multilateral one opens.

The environmental opposition is loud (900+ scientists, Greenpeace, €24 trillion in institutional restrictive policies) and does not currently constrain NOAA or the Cook Islands authority. China signed a competing seabed MOU with the Cook Islands in February 2025. Congress noticed.

The trade has three legs: arb on PIPE pricing vs. pro-forma EV once the S-4 is effective; optionality on NOAA permit timelines and Cook Islands PFS milestones; and a residual claim on a $37.1 million NAFTA award sitting in the PHOSAGMEX liquidating trust. All three require a tolerance for timelines measured in administrations rather than quarters.

_{MOST QUESTIONABLE DEAL OF THE MONTH}

Altman is as Altman Does

Normally, I would do exit of the month but that’s reserved for Chad and the boy’s at Grüns, obviously. While this is an exit, it’s also a deal so ridiculous, I thought it was an April fools joke when I first read it.

OpenAI announced the acquisition of TBPN - its first-ever media purchase. Price undisclosed. The FT and WSJ both reported “low hundreds of millions.” Puck put it at ~$150 million. Chris Lehane went on CNN and declined to confirm.

TBPN is a three-hour daily livestream that has become Silicon Valley’s water cooler. Cohosts John Coogan and Jordi Hays — both longtime Altman associates — run a format best described as SportsCenter for tech.

They ring a gong when a startup raises money. Guests have included Altman, Zuckerberg, Nadella, Benioff, and Ken Burns. Daily viewership: ~70,000. Target audience: the ~200,000 operators who move IPO books and enterprise buying decisions.

The business did ~$5 million in revenue in 2025 and is pacing above $30 million in 2026. As a condition of the deal, the advertising business is being wound down entirely.

OpenAI is paying ~$150 million for a media property it intends to stop monetizing. This is not a financial investment, which is obviously why OpenAI did it. I mean losing money is practically their whole business model.

What it is: pre-IPO narrative infrastructure. OpenAI’s Chief Communications Officer left in January 2026. The company is targeting a late-2026/early-2027 IPO at a reported ~$1 trillion valuation while secondary shares trade at a 10% discount to the last round.

The press reaction was unanimous. Ben Thompson: “OpenAI’s purchase of TBPN makes no sense, which may be par for the course for OpenAI.” Om Malik titled his column “OpenAI: Masters of Agitprop 2.0.” One unnamed OpenAI investor told the FT: “Frankly, I don’t understand it — it’s a distraction and it annoys me.” 

Altman’s response on X: “I don’t expect them to go any easier on us, am sure I’ll do my part to help enable that with occasional stupid decisions.” You mean…like this one, Sam?

PS: Hey Sam - the Overheard on Wall Street audience is pretty big, averaging 500K views a post if you’re in the market...

_{NEWS ROUNDUP}

Top Reads

• SpaceX Confidentially Files for the Largest IPO Ever

• Goldman beats estimates on strong trading revenue

• JPM beats estimates despite Dimon warning of complex risks

• Citi posts highest quarterly revenue in 10yrs on strong trading revenue

• Live Nation and TicketMaster face trust suit

• Trump threatens to fire Powell and demands rate cuts (again)

_{DEAL OF THE MONTH}

Ground Control to Major Elon

Can you hear me, Major Elon?

I may have dated myself with that reference, but SpaceX has spent 24 years doing things that most people said were physically impossible.

Now, it’s doing something that most people said was financially inconceivable: filing confidentially with the SEC for an IPO targeting a valuation of $1.75 trillion, which, if it prices anywhere near that level, will be the largest public offering in the history of capital markets. 

In fact, if SpaceX were a country, it would be the 16th largest country by GDP.

Reports emerged on March 26 that SpaceX is preparing to file a confidential S-1 prospectus with the SEC, targeting a June 2026 Nasdaq listing at a valuation between $1.5 and $1.75 trillion.

The company has lined up Bank of America, Goldman Sachs, JPMorgan, and Morgan Stanley for senior underwriting roles so the four largest banks on Earth are about to spend the next sixty days arguing over tombstone placement on what could be the most lucrative IPO mandate in history. Someone is going to get the top-left box, and someone else is going to be furious about it at the closing dinner.

To understand why $1.75 trillion is at least a defensible number, not necessarily a reasonable one, but defensible none the less, you have to understand what SpaceX actually is in 2026.

It is not primarily a rocket company. It is a satellite internet provider with nine million paying subscribers, a defense services contractor with billions in government commitments, and — following the xAI merger — an AI infrastructure business building orbital data centers powered by the sun. If you told me a year ago I’d write that in a sentence, I would’ve called you insane.

Starlink alone generated over $10 billion in revenue in 2025, with margins that would embarrass most software companies. SpaceX as a whole reportedly generated $8 billion in profit on $15–16 billion in revenue last year. That is a real business, you know, with cash flow and everything.

Then there is the February merger with xAI, which closed at a combined valuation of $1.25 trillion and which, depending on who you ask, was either a stroke of visionary vertical integration or the most audacious related-party transaction in private market history.

As a refresher, SpaceX acquired xAI in an all-stock reverse triangular merger, with xAI employees given the option to redeem shares for cash. There was no fairness opinion from an investment bank. There was no special committee. There was a blog post from Musk and a valuation that was, by most accounts, set internally.

The SEC is apparently taking a close interest in how all of this gets disclosed in the S-1, which partially explains why the filing is going in confidentially and why a June listing was originally supposed to happen in April.

Elon is no stranger to rubbing the SEC the wrong way. Remember when he tweeted (yes, pre-X) about TSLA going private?

The planned dual-class share structure means that even after going public and raising up to $50 billion in primary capital, Musk, who currently owns more than 40%, will retain dominant voting control over the company.

There is also an unusually large 20–30% retail allocation reportedly being pushed by Musk, who has apparently decided that keeping his fan base liquid and invested is as important as keeping the institutional books covered. Whether that results in a more stable stock or a more volatile one is a question the secondary market will answer on day one.

One detail being underreported: SpaceX plans to take advantage of the confidential filing process, which allows it to meet SEC requirements and address financial disclosures privately before initiating a public roadshow.

This is particularly important given the complexity of integrating xAI’s liabilities, intellectual property, and revenue into SpaceX’s consolidated financials, plus the defense contract sensitivities that come with being the Pentagon’s favorite rocket supplier.

Imagine the poor SEC gremlin who has to look at a rocket launching social media company that also runs a telecom empire and is somehow also an AI company? Oh, and by the way, every time Elon tweets there will probably have to be an 8-K. 

When the S-1 goes public, likely in April or early May, it will be the most-read document on Wall Street since the Facebook prospectus. Every number will be stress-tested. Every Musk quote will be pulled for the risk factors section, especially funding secured. Every analyst will be writing a 47-page initiation report by the following Monday.

There is a version of this story where the SpaceX IPO is the defining capital markets moment of this decade, the trade that either confirms or destroys the argument that private market valuations have become untethered from reality.

The June IPO is projected to be the largest in history by cash raised, potentially surpassing the $29.4 billion proceeds of Saudi Aramco’s 2019 debut.

Whether investors ultimately pay $1.75 trillion for the privilege of owning a rocket company, a satellite internet service, an AI subsidiary with a controversial chatbot, and whatever X has become at this point — is also a question for June.

For now, the banks are engaged, Elon Musk is preparing to take the most valuable private company in history public, and, according to the Financial Times, all of it coincides with his birthday and a planetary alignment. Naturally.

_{PRESENTED BY F2}

AI That Won’t ‘Jump the Gun’

Isn’t it the most frustrating thing when you prompt GPT, Claude, Grok (insert your favorite finance AI here), and they immediately jump to working rather than listening and planning?

Thus follows the typical AI workflow in finance:

1) Run it. 2) Review it. 3) Find assumption errors. 4) Redo it.

While AI may be helpful, it has a tendency to act first and ask questions later (unless prompted).

Plan Mode from F2 lets you review the entire workflow before it runs:

• the logic

• the assumptions

• the build path

Build the plan upfront, let the agents handle the grunt work.

This equates to fewer drafts, faster outputs, and way less time retracing steps.

See how Plan Mode works.

_{PE DEAL OF THE MONTH}

Boralex Bids the Public Markets Adieu

There is a specific category of company that private equity, and infrastructure PE in particular, absolutely cannot resist: a well-run, cash-generative, publicly-listed infrastructure business whose shares are trading at half of where they were five years ago because interest rates went up and the market decided it didn’t like renewable energy anymore. Boralex is that company. Brookfield is that buyer. The only mystery is why it took this long.

Brookfield Asset Management and Caisse de Dépôt et Placement du Québec (La Caisse) announced on March 25 that they have entered into a definitive agreement to acquire Boralex Inc. for C$37.25 per share in cash, implying a total equity value of approximately C$3.8 billion and a total enterprise value of approximately C$9.0 billion ($6.5 billion U.S.), including debt.

The offer represents a 31.8% premium to Boralex’s March 20 closing price, and a 36.4% premium to the 30-day VWAP preceding the first media report that the company had begun reviewing strategic alternatives. Boralex’s board unanimously approved the transaction. The deal is expected to close in Q4 2026.

La Caisse, already Boralex’s largest shareholder with approximately 15% of the outstanding shares, will vote in favor of the transaction and will make a post-closing investment resulting in a 30% pro forma ownership stake in the newly private company.

Brookfield, through its flagship infrastructure strategy and Brookfield Renewable Partners, will hold the remaining 70%. So the Quebec pension fund is going from 15% of a publicly-traded company to 30% of a private one, which is either a vote of extraordinary confidence or a sign that they’ve been waiting for exactly this structure for a long time. Possibly both.

Why go private? Boralex’s CEO Patrick Decostre said the company’s “strong pipeline” requires financing that being private will facilitate more efficiently, which, translated, means: as a public company, every time we need equity to fund development, we have to go back to a skeptical market that doesn’t want to hear about European wind farms right now. The company’s shares peaked above C$55 in early 2021, when investors were paying premium multiples for anything adjacent to “green” and “infrastructure.”

Now, AI companies and hyperscalers are clambering for power demand, bringing power companies back into favor once again.

Boralex was trading at less than half that level as of last week, not because the assets got worse — the wind still blows in France, the hydro facilities in Ontario still generate power on schedule — but because the macro backdrop moved against rate-sensitive yield vehicles. Classic infrastructure private equity hunting ground.

Boralex adds roughly four gigawatts of operating clean power capacity to Brookfield’s existing global fleet of 46 gigawatts, plus another eight gigawatts of development pipeline in Canada, France, the United States, and the United Kingdom.

The transaction is valued at ~13x 2026 EBITDA, reasonable for a high-quality renewable portfolio with long-term contracted revenue and inflation-linked tariffs in regulated European markets.

The deal is being funded by Brookfield Renewable Partners through its flagship infrastructure strategy. No bank financing details were disclosed, but given Brookfield’s balance sheet and La Caisse’s scale, the capital structure conversation was presumably brief.

One detail worth noting: Boralex’s 2030 Strategic Plan targeted 6 GW of installed capacity by the end of this decade. As a public company, hitting that number was going to require regular trips back to equity markets in an environment where renewable energy stocks were being treated like a category error.

As a private company backed by Brookfield’s global development capabilities and La Caisse’s institutional firepower, the path is materially cleaner. Whether it actually gets to 6 GW is a question for 2030, but at least now they don’t have to explain their progress on a quarterly earnings call.  

All we can say is bon chance to Brookfield and La Caisse and bravo to the Boralex team (and their shareholders).

_{INT’L DEAL OF THE MONTH}

BP Says Auf Widershen to Germany

There is a version of corporate strategy where you sell everything that isn’t your core business, repeatedly raise your cost-savings target as you do it, and at some point the press releases start to feel less like strategy updates and more like a man furiously throwing things overboard to keep a boat afloat.

BP has been writing that press release for about fourteen months.

On March 19, BP announced it has entered into a definitive agreement to sell its Gelsenkirchen refinery and associated businesses to Klesch Group, an independent European refiner, for an undisclosed price. Which means, one party was certainly not thrilled about the price. 

The refinery complex — operating two sites in Horst and Scholven in Germany’s Ruhr region under the legal entity Ruhr Oel GmbH — processes approximately 12 million tonnes of crude oil per year, with a crude distillation capacity of 265,000 barrels per day.

It produces petrol, diesel, jet fuel, heating oil, and more than 50 other products for the European chemical industry. About 1,800 employees are expected to transfer to Klesch upon completion, along with offtake arrangements that allow BP to continue sourcing ground fuels, aviation fuel, and coke from the site.

The undisclosed price is doing a lot of work in that announcement. BP invested approximately $2 billion in modernizing the Gelsenkirchen site since 2017 — upgrading steam and power infrastructure, distributed control systems, and tank farm and pipeline refurbishments. BP’s underlying operating expenditure associated with the site runs around $1 billion per year.

Disposing of that $1 billion annual cost base is what allowed BP to raise its 2027 structural cost reduction target for the third time in fourteen months — to a range of $6.5 to $7.5 billion, up from its original $4 to $5 billion target set in February 2025.

The progression, if you’re keeping a tally at home: $4-5B → $5.5-6.5B → $6.5-$7.5B.

Each increase comes with a refinery sale. At this rate, by 2027 BP will have announced $15 billion in savings and will own exclusively a headquarters, a logo, and a very good sustainability report.

Klesch Group, the buyer, is a Geneva-based independent European refiner controlled by chairman A. Gary Klesch. The company’s existing portfolio includes the Heide Refinery in Germany and the Kalundborg Refinery in Denmark, the first acquired from Shell in 2010, the second from Equinor in 2022.

There is a pattern here: Klesch buys high-quality European refinery assets from oil majors who have decided they are no longer in the European refinery business, then runs them.

This is the third time that playbook has worked. It will almost certainly not be the last, given how many oil majors are currently in the process of deciding they are no longer in the European refinery business.

Patrick Wendeler, BP’s head of country for Germany, offered a statement that was, by corporate press release standards, almost poignant: “We have a long history of operating successful assets and brands in Germany, and we are deeply grateful for the refinery’s decades of contribution to our business.”

Translation: so long, farewell, auf wiedershen, goodbye - but, we would like to be remembered kindly.

The transaction is expected to close in the second half of 2026, subject to regulatory and governmental approvals. The deal is free-cash-flow accretive for BP based on historical performance, which, again, is doing quite a bit of euphemistic heavy lifting given the $1 billion annual cost structure that is about to leave the books.

BP has now announced or completed more than $11 billion of its planned $20 billion divestment program.

Whether the strategy ultimately works, selling everything at sufficient speed and raising your cost savings target enough times eventually fixes BP’s underlying return on capital problem, is a question above this newsletter’s pay grade.

But credit where it’s due: they are executing. Every quarter, something gets sold. Every sale comes with a press release. Every press release comes with a higher savings target. It is, in its own way, a beautiful system.

_{NEWS ROUNDUP}

Top Reads

• SpaceX targets $1.75 trillion valuation in June IPO

• BofA makes entire group to help PE firms with exit drought

• Middle East Funds are the new private equity

• Carve outs as a proactive M&A strategy

• 2026’s M&A Resurgence 

• The bank to take SpaceX public

_{DEAL OF THE MONTH}

A Well Dressed Deal

If you have ever tried to acquire a company that doesn’t want to be acquired, you know that patience is not just a virtue, it’s a budget line item.

Cintas and UniFirst are the two largest uniform and workwear services companies in the U.S. Think the companies that supply, launder, and deliver uniforms to factories, hospitals, restaurants, and auto shops across the country.

Cintas is the dominant player with a market cap of $72 billion. UniFirst is the scrappy number two, family-controlled, and apparently allergic to being bought.

Cintas Corporation and UniFirst Corporation announced on March 11 that they have entered into a definitive agreement under which Cintas will acquire UniFirst for $310.00 per share in cash and stock, representing an enterprise value of approximately $5.5 billion.

For those keeping score: in December 2025, UniFirst received an acquisition proposal from Cintas for $275 per share in cash, valuing the company at about $5.2 billion, which UniFirst’s board rejected.

Cintas came back with a higher number, structured the deal as part cash, part stock, and negotiated a $350 million reverse termination fee — payable if the deal dies on antitrust grounds — which apparently was the magic combination that finally got the Croatti family off the fence.

Under the terms, UniFirst shareholders will receive $155 in cash and 0.7720 shares of Cintas stock for each UniFirst share they own, representing a combined value of $310 per share based on Cintas’ closing price on March 9. RIP to your printer for this investor presentation…

The implied total enterprise value of $5.5 billion represents a multiple of 8.0x run-rate trailing twelve months EBITDA, including approximately $375 million of operating cost synergies.

The synergies come from route optimization, plant consolidation, and procurement scale, which is a polite way of saying: when you run half of North America’s uniform delivery routes, you can probably cut some duplicate truck stops.

Here is the thing about the uniform rental business that nobody talks about at dinner parties: it is shockingly good. Cintas already generates over $10 billion in annual revenue, and the “Uniform Rental and Facility Services” segment accounts for approximately 77% of total revenue, operating on a recurring contract basis where drivers visit client sites weekly to pick up soiled uniforms and deliver clean ones. Does someone smell an LBO target next year?

Route density is everything - the more stops per mile, the more profitable each truck. Acquiring UniFirst’s routes doesn’t just add customers; it makes Cintas’s existing routes more efficient by filling in geographic gaps.

The combined company will deliver products and services to approximately 1.5 million business customers across North America. To contextualize that number: there are only about 33 million registered businesses in the United States.

Cintas and UniFirst together will be dressing, cleaning, and supplying roughly one in twenty of them. If that’s not a moat, I don’t know what is.

The regulatory situation is, shall we say, colorful (hopefully not blue like this pitch deck).

A combined Cintas-UniFirst entity would own nearly half of the market, which may require divestitures of specific local routes or branches in concentrated geographic areas. Cintas is “highly confident” on regulatory approval and has already staged a $350 million reverse termination fee as proof of conviction.

Whether that confidence is warranted or just well-dressed optimism remains to be seen. The DOJ has been known to take a close look at markets where “Big Three to Big Two” consolidation is on the table.

The transaction was unanimously approved by both boards, and entities affiliated with the Croatti family.

Pro forma leverage at closing is expected to be approximately 1.5x debt to EBITDA, preserving Cintas’s investment grade credit profile. The firms were advised by Goldman Stanley (who cares who represented who), but this will be a good one for the league tables for both.

The deal is expected to close in the second half of 2026. And somewhere in Wilmington, Massachusetts, the UniFirst family is quietly putting their lawyers on speed dial just in case.

_{PRESENTED BY TERMGRID}

The Platform Leading PE Firms Use for Their Financing Workflows

Termgrid is how private equity sponsors, advisors, and lenders execute financings.

Their vertical-focused collaboration platform centralizes debt raises and equity co-invest processes, giving you better insights and faster execution.

With Termgrid, investment firms can:

• Boost productivity: increase deal team capacity by up to 84% (see case studies here)

• Reduce financing risk: access the largest financing ecosystem (1,600+ institutions and 30,000+ professionals)

• Make faster, data-driven decisions: analyze terms and credit negotiations in real time

• Lower transaction costs: reduce data sharing and legal inefficiencies

• Build institutional knowledge: automatically capture data across relationships, transactions, and precedent terms

The best sponsors (from mid market to mega funds) are using Termgrid for their financing infrastructure. See why today.

Request a demo.

_{SPONSOR DEAL OF THE MONTH}

Monomoy Gets Lubed Up

For once, a PE shop buying a bunch of lube might not be just because bonus season is coming up for juniors.

Shell has been quietly trying to convince itself that owning a chain of quick-lube shops makes sense for a global oil major for twenty-four years. On March 9th, it finally stopped pretending.

Monomoy Capital Partners announced it has entered into a definitive agreement to acquire Jiffy Lube International from Pennzoil Quaker State Company — a wholly owned subsidiary of Shell USA — through its Fund V for approximately $1.3 billion.

Shell acquired Jiffy Lube in 2002 for reasons that probably seemed clever at the time: vertical integration, brand synergy, guaranteed lubricants demand, banker needed fees, nepo baby corp dev guy needed a win, and has spent the two decades since discovering what every oil major eventually discovers: owning retail is a very different business from drilling holes in the ground. Turns out the customers at a Jiffy Lube don’t care whose barrel the oil came from.

What makes this deal more interesting than a CIM you would use for modeling is that practice is the structure. Monomoy is acquiring both Jiffy Lube International — including the registered trademark and the franchisor business — and Premium Velocity Auto (PVA), identified as the second-largest Jiffy Lube franchisee, which runs more than 360 locations across approximately 20 states.

So you’re not just buying a franchise system; you’re buying the franchisor and the second-largest franchisee simultaneously, which gives Monomoy an immediate owned-and-operated base to work from while also collecting royalties from the other 1,600-plus independently-owned stores. If you squint, that is a pretty elegant platform structure (and I bet it banged on the structure chart page).

For Monomoy, this is squarely in their wheelhouse. Monomoy Founder and Co-CEO Dan Collin described it as a “complex corporate carve-out transaction,” noting that the firm has executed these for global corporations for over twenty years.

The carve-out complexity here is real: Shell’s lubricants supply agreement had to be negotiated separately, Pennzoil brand relationships had to be ring-fenced, and the franchise system needed new standalone infrastructure for everything from IT to HR to legal. None of that is simple, and most PE firms can’t actually execute it. 

Monomoy’s pitch is essentially: we can. Which, given their track record, is at least credible. Don’t forget, when the spreadsheet says the deal works, you still have to make that happen in practice.

As part of the transaction, Shell’s subsidiary Pennzoil Quaker State has entered into a long-term lubricants supply agreement with Monomoy, which is Shell’s way of saying: we sold the business but we’d still like to sell you the oil that goes in the cars.

You have to respect the hustle. Shell loses the headache of running a retail franchise, keeps the lubricants volume, and pockets $1.3 billion. Monomoy gets a household brand with 19 million annual customers, a turnaround opportunity, and locked in supply to support its business.

The deal is expected to close in the second half of 2026, pending regulatory approval. In the meantime, if you walk into a Jiffy Lube and the guy under your car is in a suit, it means Monomoy underwrote some aggressive cost cutting.

_{COMPLEX DEAL OF THE MONTH}

3M and Bain Do the Most Complex Deal of the Year

There is a category of M&A transaction that defies easy description. It is not a straight acquisition. It is not a merger. It is a simultaneous acquisition-and-contribution joint venture where one party buys an outside business, folds in an existing one to create a new entity, while receiving cash proceeds on the same day they write a check. It is the M&A equivalent of a Rubik’s Cube. 3M and Bain Capital solved it on March 19th.

3M announced that it has entered into a definitive agreement to acquire Madison Fire & Rescue from Madison Industries, in partnership with Bain Capital, for $1.95 billion and that it will be forming a new joint venture focused on fire, safety, and rescue solutions.

Here is where it gets interesting: 3M and Bain Capital will establish a new joint venture in which 3M will contribute its existing Scott Safety business, receive $700 million of cash proceeds upon closing, and own 50.1% of the new company, with Bain Capital owning 49.9%.

Read that again. 3M is spending money and receiving money on the same transaction. 

It is buying a fire and rescue equipment company, contributing its own safety business into the deal, and walking away with $700 million in cash on day one while retaining majority ownership of the combined entity.

Credit where it’s due, whoever modeled this one earned their carry.

Madison Fire & Rescue’s portfolio includes Amkus and Holmatro, makers of hydraulic and battery-powered rescue tools (the ones you see cutting people out of cars on the highway); Task Force Tips, a maker of high-performance nozzles for fire hoses; and Waterax, which makes mobile water pumps for wildland firefighting. These are not glamorous products.

They are the products that exist specifically for the moments when everything else has gone wrong. Scott Safety, which 3M acquired in 2017 for approximately $2 billion, makes self-contained breathing apparatus systems, the tanks and masks that firefighters wear when they enter burning buildings.

Together, the portfolio covers essentially the entire toolkit of a first responder: the truck, the hose nozzle, the breathing apparatus, the rescue cutters. If you are a fire department purchasing director, this new JV is now your one-stop shop for almost everything except the actual firefighter.

The acquisition allows 3M to expand its safety and fire protection product offerings, which are high-margin businesses that can provide steady cash flow, while Madison Fire & Rescue’s emergency vehicle manufacturing business complements 3M’s existing personal protective equipment lines.

For Bain, the story is simpler: they get nearly half of a cash-generative, defensible industrial business with a captive government customer base (every fire department in the country), funded largely by 3M’s asset contribution and the deal’s cash-in structure.

In a week dominated by carve-outs and consolidations, the 3M-Bain deal stands out as the most structurally interesting. It is not the biggest number on the tape. But it is the one your corporate finance professor would have used as a case study, if your corporate finance professor had ever spent time thinking about fire trucks, which, statistically: they probably have not.

_{NEWS ROUNDUP}

Top Reads

• Shell sells Jiffy Lube to Monomoy Capital Partners for $1.3 billion

• 3M and Bain Capital to buy Madison Fire & Rescue for $1.95 billion

• Cintas-UniFirst faces antitrust scrutiny as combined entity nears 50% market share

• Megadeal volume up 57% year-over-year as 2026 M&A hits full stride

• KMPG 2026 M&A Insights

• Middle market M&A set to rise in 2026 as PE funds sit on $2.1 trillion of dry powder

• Possible additions to the S&P 500*

_{DEAL OF THE MONTH}

Lights, Camera, Leverage

Alright, so I called Netflix x WBD but Paramount was willing to overpay, sue me.

It started like every good Hollywood story: two suitors, one prize, and somebody about to get humiliated in public.

Warner Bros. Discovery, saddled with roughly $40 billion in debt from the disastrous AT&T merger and watching cable revenue evaporate in real time_, had been exploring a sale since October.

By December, Netflix had swooped in with an $82.7 billion offer for WBD’s studios and streaming assets, looking to buy the keys to Hogwarts, Westeros, and the DC Universe in one fell swoop. Zaslav’s board took it. David Ellison’s Paramount Skydance, newly formed just six months prior from the Skydance-Paramount merger, watched from the sidelines and apparently decided: not on my watch.

What followed was one of the ugliest, most entertaining hostile takeover attempts in recent memory - shareholder letters, competing proxy slates, HSR filings, a DOJ antitrust probe of Netflix for “attempting to create a monopoly,” and a very public game of chicken between two billionaires named David.

On February 26, 2026, Netflix blinked. After WBD’s board declared Paramount’s $31-per-share offer a “superior proposal,” Netflix declined to raise its bid and quietly walked out the door. Paramount paid a $2.8 billion termination fee to Netflix, which, if nothing else, is the most expensive “thanks for playing” in entertainment history.

The final terms: Paramount Skydance acquires 100% of Warner Bros. Discovery at $31 per share in cash, implying an equity value of $81 billion and an enterprise value of approximately $110 billion.

The deal is funded by $47 billion in equity from the Ellison family and RedBird Capital Partners, alongside $54 billion in new debt commitments from a banking syndicate led by Bank of America and Citigroup.

To sweeten the deal for WBD shareholders still skittish about the Q3 2026 close timeline, Paramount added a $0.25 per share “ticking fee” for each quarter the deal doesn’t close, a nice little incentive that also signals Ellison’s confidence in getting this done.

The combined company will control a library of more than 15,000 film titles, Harry Potter, Game of Thrones, Mission: Impossible, the DC Universe, and whatever’s left of South Park, and is committed to releasing at least 30 theatrical films annually (15 per studio, with a minimum 45-day theatrical window).

On a fully synergized basis, Paramount values WBD at 7.5x 2026 EBITDA, and expects over $6 billion in cost synergies driven by tech integration, real estate consolidation, and what every press release euphemistically calls “streamlining operational efficiencies.”

You don’t need to be a media analyst to see the logic here. Netflix, Apple, Amazon, and Disney have turned content into a scale game. Without the DC and Harry Potter IP, WBD was a slowly deflating balloon. And Paramount, fresh off the Skydance merger, needed a cannon-sized content library to compete with the streamers. Together, they at least have a fighting chance.

The less obvious part of the story? The debt. At closing, the combined entity will carry a net debt-to-EBITDA of approximately 4.3x on a synergized basis, which is manageable, but only if everything goes according to plan.

That $6 billion in synergies has to actually materialize. The studios have to stop stepping on each other’s release calendars. The streaming platforms have to be integrated without alienating subscribers. And Ellison — who, remember, has been CEO of Paramount for all of about six months — has to prove he can actually run a $110 billion media conglomerate and not just ask his daddy to buy one.

Credit agencies have already downgraded the entity to junk status. WBD shares are trading at a modest discount to the offer price, reflecting real skepticism about the Q3 close timeline or regulatory risk. And every Hollywood executive with a deal in progress is nervously wondering whether the new regime will actually greenlight their project or just cut costs until the EBITDA looks prettier.

But hey, David Ellison’s dad is Larry Ellison. The equity check is already written. Sometimes you just have to bet on the guy with the most zeroes in his savings account.

The deal is expected to close in Q3 2026, pending regulatory clearance and WBD shareholder approval (vote expected in early spring). Mark your calendars for what could be the closing dinner of the decade.

_{NEWS ROUNDUP}

Top Reads

• Megadeal volume up 57% YoY — fewer deals, bigger stakes in the 2026 M&A landscape

• Global M&A boom carries into 2026 despite tightest capital squeeze in 30 years

• Goldman CEO says 2026 could be a “top decile” year for M&A

• Victory Capital makes $8.6B bid for Janus Henderson

• Danaher acquires Masimo for $9.9B in diagnostics push

• Texas Instruments strikes $7.5B deal to buy Silicon Labs

• PwC 2026 US Deals Outlook

• The future price for Nvidia’s H100s​​​​​​

_{PRESENTED BY ORNN AI}

Wall Street’s Newest Commodity: GPU Compute

As more people use AI, demand for the infrastructure powering it grows exponentially. Training models and running large-scale inference requires massive compute capacity, turning chips like Nvidia’s H100 into one of the most important inputs in the AI economy.

Until recently, the market for that compute largely lived behind closed doors. Cloud providers, model labs, and large institutions negotiate pricing for GPU capacity privately, often with little transparency or consistency.

Ornn AI has built an index designed to commoditize AI compute, tracking the real market price of GPU capacity used to train and run modern AI systems.

Through new partnerships with Kalshi and Robinhood, retail traders can now participate as well. Markets tied to specific chips allow traders to take positions on where compute prices are headed.

If you are an institution, create an account on Ornn’s data platform to access its compute pricing index and trading infrastructure.

If you are a retail trader, explore the new prediction markets powered by Ornn on Kalshi or Robinhood by searching for "H100".

_{PE DEAL OF THE MONTH}

GIP and EQT Keep the Lights On

Last month we talked about how AI is eating the power grid alive. This month, the power grid fought back… by going private.

On March 2, a consortium led by BlackRock’s Global Infrastructure Partners and Swedish private equity giant EQT agreed to acquire AES Corporation for $15.00 per share in cash, representing a total equity value of $10.7 billion and an enterprise value of $33.4 billion including the assumption of existing debt.

Joining GIP and EQT as co-underwriters are CalPERS and the Qatar Investment Authority — so yes, your California pension and Qatar’s sovereign wealth fund are now co-owners of your Indiana power bill.

AES, for those who need the refresher, is a Fortune 500 global energy company operating utilities in Indiana and Ohio, plus power generation assets across Chile, Panama, the Dominican Republic, Vietnam, Bulgaria, and roughly a dozen other countries. It’s not the sexiest name in the sector, but it generates real cash flow from real assets that the world genuinely needs.

The headline premium is 40% to the 30-day VWAP prior to the first media report of a potential deal last July. But here’s the wrinkle: the stock had already run up on deal speculation, and shares actually fell more than 16% on announcement day because the offer came in below the inflated Friday close of $17.28. In other words, the leak helped and hurt simultaneously, a classic Wall Street own-goal.

So why sell? AES management was refreshingly honest about it. Absent this transaction, the company would have likely needed to cut or eliminate its dividend and/or issue significant new equity to fund its growth plans beyond 2027. That’s not a great look for a public utility.

Going private gives AES the capital runway it needs to keep investing in new generation capacity without the quarterly earnings pressure. Bayo Ogunlesi, CEO of GIP, put it plainly: there is a need for “significant investments in new capacity in electricity generation, transmission and distribution, especially in the United States.” Translation: AI data centers don’t run on good intentions.

This deal follows an extraordinary wave of private capital flooding into the U.S. power sector. Blackstone acquired TXNM Energy for $11.5 billion. Constellation is buying Calpine for $16.4 billion. And Bloomberg estimates that more than $280 billion in power sector M&A has been announced since the start of 2025. If you’re an energy banker right now, you are having the time of your life.

The Midwest utilities (AES Indiana and AES Ohio) will continue operating as locally managed regulated utilities, per the deal terms. That didn’t stop Indiana Democrats from immediately condemning the transaction.

“Private firms having a stake in public utilities will put profits over people,” said Congressman André Carson, which is also something that gets said every single time a utility changes hands and, yet, hasn’t actually stopped a utility from changing hands.

JPMorgan and Wells Fargo advised AES. Goldman advised GIP and CalPERS. Citi advised EQT. Everyone got paid.

The deal is expected to close in late 2026 or early 2027, subject to AES shareholder and regulatory approval.

_{STRATEGIC DEAL OF THE MONTH}

Mouthwatering Premiums

Here’s a fun M&A dynamic that doesn’t get enough attention: sometimes the best acquisition target is the company you’re already in business with. You’ve seen the technology. You’ve stress-tested the clinical data. You know exactly what you’re buying. The only question is whether you’re willing to pay full price for it.

Gilead Sciences answered that question definitively on February 23. Gilead agreed to acquire Arcellx — its partner in developing a CAR T-cell therapy for multiple myeloma — for $115 per share in cash, representing a total implied equity value of $7.8 billion. That’s a 79% premium to Arcellx’s Friday closing price and a 68% premium to its 30-day VWAP. Arcellx went public in 2022 at $15 per share. Do the math.

The center of gravity for this deal is a single drug: anitocabtagene autoleucel (anito-cel), an investigational BCMA-directed CAR T-cell therapy for patients with relapsed or refractory multiple myeloma.

The Phase 2 iMMagine-1 trial showed a 96% overall response rate and a 74% complete or stringent complete response rate at a median follow-up of nearly 16 months, numbers that would make any oncology investor audibly gasp. The FDA has already accepted the BLA with a PDUFA action date of December 23, 2026, meaning a commercial launch decision is imminent.

Gilead already owned 11.5% of Arcellx’s outstanding stock and had been co-developing anito-cel through its Kite subsidiary since 2022.

The deal eliminates all future profit-sharing, milestone payments, and royalties, which could remove up to $1.5 billion in future payouts from Gilead’s obligations.

At the same time, Gilead gains full commercial economics on a drug it already believes in. CEO Daniel O’Day called it their conviction that anito-cel could become “a foundational treatment for multiple myeloma over time.”

There’s also a contingent value right (CVR) of $5 per share, triggered if anito-cel reaches $6 billion in cumulative global net sales through the end of 2029. The CVR is Gilead’s way of saying: we think this drug is going to be really big, but we’re not willing to pay for that upside today - earn it with us.

The broader context matters here. Gilead’s cell therapy business — anchored by Yescarta and Tecartus — saw overall sales fall 7% in 2025 and is expected to decline another 10% in 2026 amid pressure from newer competitors.

The company’s COVID-19 drug Veklury is also losing steam (COVID was 6 years ago now, not 3). Gilead needed a pipeline catalyst, and it had one hiding in plain sight at a company it already partly owned.

BofA Securities and Morgan Stanley advised Gilead. Centerview Partners advised Arcellx. The deal is expected to close in Q2 2026, subject to regulatory approvals and completion of a tender offer.

For Arcellx shareholders, this is a remarkable outcome. Four years from a $15 IPO price to a $115 cash takeout, driven entirely by clinical execution and a strong partner relationship. Not bad for a biotech that most people couldn’t name six months, or even 6 days, ago.

_{BUYSIDERS MACRO MEMO}

The Consequences of Controlled Stupidity

June 30, 2028 - 9:00pm // Midtown Manhattan

The unemployment rate printed at 4.3% today, right in line with expectations. The market held steady on the news, and the Dow is sitting comfortably at 60,000.

Traders and their Claude bots predicted this months ago. In fact, Bill Ackman’s AI assistant called the exact points on the Dow three weeks ago for today’s unemployment print.

M&A and capital markets remain hot. As a shock to everyone, Altman actually pulled off spending $3 trillion of CapEx for OpenAI, and other AI companies followed suit.

Think back to two years ago… Claude single-handedly tore down the entire market - causing a sell off in every industry and a panic across the Street - when people thought analysts would be replaced.

Instead, your DCFS have grown from 5 scenarios to 500. What used to take analysts 4 hours now takes you 4 minutes, so the expectations for your work product have increased.

You were told investment banking hours would improve with AI, yet you are still at the office until 3am.

Sure, you’re not manually editing Excel files anymore - an art lost to time - but now, you have to parse through the 25 bad responses your Bank’s AI bot “Jerome" - named after a Fed Chair of a bygone era - gives you, because you didn’t phrase your prompt correctly.

Mercifully, your bank has replaced all compliance, AML, KYC and a majority of its accounting staff with AI bots. So when your MD comes to you and asks you to onboard a client for his “newest long-shot deal”, you are able to get it approved in minutes without the back and forth.

As you take the elevator to get your DoorDashed slop bowl (ordered and delivered by your AI bot based on your recent health trends), you ride past the offices of a few casualties from the AI explosion.

No one expected such casualties from this AI boom; however, SaaS companies went belly up (causing KKR to be rolled up by Apollo), Blue Owl went bust, and SoftBank became the best investor in the world (after their equity investment in OpenAI and follow-on investment in Anthropic mooned following their IPOs).

The offices of what used to be Deloitte and EY now sit empty, as relics from a pre-AI consulting world. Naturally, consulting had been the next victim after the SaaS-pocalypse. After all, increasing revenue and cutting costs to “add shareholder value” is basic math.

After picking up your food, you have a notification from “Jerome” that he’s integrated the latest Monte Carlo scenario into your model. BlackRock is happy. Then, he reminds you that he will do a price refresh tomorrow, after Treasury Secretary Dimon’s speech.

Once upstairs, you settle in to wrap up another pitch for your MD, knowing it will go nowhere. You begin crafting the perfect prompt for “Jerome” knowing that one typo or hint of bias will ruin the whole night.

In the end, the AI boom simply boosted productivity, like Excel did in the ‘80s… and the internet did in the ‘90s… It just accomplished it at a scale never seen before. MDs and clients adjusted their expectations accordingly.

_{PRESENTED BY F2}

The New Verb on the Street

“I F2’d it.”

Historically, F2 - the key - has been the shortcut to reveal the formula behind a number. That key shows you exactly what’s driving the output.

Now, F2 - the company - embodies that same ideology in the tools it builds. With its new Audit Mode, every number within a report is not only cited, but fully traceable in real time. AI-generated analysis is written into a live, Excel-native Databook complete with real formulas, linked sources, and full traceability.

Trace any value to the exact file, sheet, and cell. Inspect precedents and dependents. Update assumptions with citations preserved in context. From any report or chat output, click into a number and see precisely how it was produced.

F2 merges speed with defensibility, delivering analysis you can confidently take into IC.

When your boss asks, “Where did you get these numbers?”, the answer is simple:

“I F2’d it.”

Book a demo to see Audit Mode in action.

_{NEWS ROUNDUP}

Top Reads

• Veris Residential goes private in $3.4 billion deal

• Citi to sell 24% equity take in Banamex

• Nvidia to invest $30 billion in OpenAI’s next funding round 

• Databricks raises $5 billion in latest funding round 

• Paramount increases bid for WBD to $31 per share

• Blue Owl halts redemptions 

• Will the Citrini scenario become reality?