Physical AI raised $47.4 billion in the first half of this year, more than the three years from 2022 to 2024 combined. Waymo's $16 billion round in February is nearly a third of it on its own. The money is no longer the constraint, but the shape of it could be.

Your Second Customer Is a Lender
Equity buys the inflection. Debt carries the scale.

We published this in full on Monday. If it already reached you, skip ahead. Everything below it is new. Hardtech can’t be financed the way software was. You cannot copy a robot for free, and paying for the hundredth identical unit with the most expensive money on the cap table quietly destroys the returns the company is pitching. The milestone that unlocks cheaper capital arrives when a unit starts repeating, and it almost never lands neatly on a priced round.

The screening list is the useful half for anyone on the equity side: everything a lender underwrites is worth diligencing years before the capex ask, because the choices that make a company financeable are cheap at pre-seed and expensive at Series B.

In the field

Four companies, and the kind of capital each one is built on.

Kubos Semiconductors — Cubic GaN microLEDs for high-speed optical interconnects
Kubos raised $2.1 million to accelerate its cubic GaN microLED work: a Growth Catalyst grant from Innovate UK, matched by the Development Bank of Wales, the Low Carbon Innovation Fund 3 and S4C Digital Media, plus follow-on from existing shareholders. Total funding now passes $8 million.
Why it matters: Public R&D money and private capital doing different jobs in the same round. Innovate UK paid for the science, the Development Bank of Wales and two funds matched it, and existing shareholders followed. Materials work this capital-intensive rarely gets built any other way.

Distance — Visual Intelligence for vehicles and cockpits
Kia’s Vision Meta Turismo won a 2026 Red Dot Design Award in the Design Concept category, with Distance’s 3D Lightfield HUD integrated into it. Kia is the first consumer automotive manufacturer to put true lightfield technology into a concept car, rendering an interface edge to edge across the windshield with per-pixel depth. Separately, Distance’s work with Supacat on battlefield human-machine interfaces was featured in The Defence Blog. On Finnish television this month, CEO Urho Konttori made the point most founders miss about defence: there is no entity called NATO that buys things. Member states buy. Point the pipeline at governments, not an acronym. Distance is two and a half years old, sells across the Nordics, the UK, Canada, the Netherlands and France, and defence is now its main revenue line.
Why it matters: Automotive design cycles run years, defence procurement runs longer, and it is fragmented across governments rather than centralised behind one buyer. Neither is a market you can reach on an eighteen-month runway, which is the whole argument for matching the capital to the physics.

Kia x Distance : 3D Lightfield HUD | Distance Technologies

🎉 Congratulations to Kia Worldwide on the Vision Meta Turismo's 2026 Red Dot Design Award in the Design Concept category.

LinkedIn

ai-coustics — Audio intelligence for voice AI
Voice AI breaks on bad sound: a noisy room, a poor line, someone else talking nearby. ai-coustics builds the layer that cleans the audio before the AI ever hears it. This month it shipped a model that is over 70% better than the open standard at telling real speech from everything else, tested on drive-thru recordings the team made themselves by parking a car in Berlin and reading fake orders through the window for an afternoon. A second release made their call-scoring model 77% faster and 90% lighter, which is the difference between checking a sample of calls and checking every one.
Why it matters: No lender underwrites research. This is what equity is for: the stretch between an idea and a number somebody else can check. Two shipped models and a published benchmark is what that money is supposed to buy. They also published a guide to picking between the two.

Nest Talks #6: How to choose the right ai-coustics model | ai-coustics

We keep getting the same question from teams building voice AI: which model do I actually need?

LinkedIn

nunu.ai — AI agents that use software the way people do
Launched browsers for its agents, built with Browser Use, alongside cross-platform support so a single agent runs across PC, phone and web. Customers are already using it across QA, compliance and customer support.
Why it matters: Games are the cheapest place on earth to teach an agent to act. Millions of attempts, no broken hardware, no safety case. That is why we backed nunu: agents that learn to act in games are the shortest route we can see to agents that act in the physical one. Working across devices is the first evidence those skills travel.

APK Arena Benchmark for AI Agents on Phone-Use | Jan Schnyder posted on the topic | LinkedIn

Today we’re launching APK Arena: a frontier benchmark for evaluating AI agents on phone-use. 🤖 📲

LinkedIn

Radar

Beyond our portfolio, what caught our attention this month.

• NEURA bought its fifth company in 16 months, and every one came with robots already in the field. Fresh from the €1.2 billion Series C that gave European robotics its biggest half on record, it took full ownership of Adlatus, an Ulm cleaning-robot maker with hundreds of systems deployed and twenty years of operational knowledge, weeks after acquiring Bosch Rexroth’s ACTIVE Shuttle. Europe’s best-funded robotics company has decided the bottleneck is installed base, and it is spending its Series C to buy it. Which is this month’s deep dive from the other direction: a deployed fleet is valuable enough to acquire and still not valuable enough to borrow against. (PitchBook · Sifted)

• Humanoid robotics has its first listed company, and the market priced it at five times the IPO on day one. Unitree listed in Shanghai on 19 August and closed up 460 percent, near $50 billion. Every private mark now gets set against that print. (Bloomberg)

• DeepMind's new model runs robots it did not build. Gemini Robotics 2 controls a full humanoid from feet to fingertips, demonstrated on Apptronik, Franka and Dexmate hardware. Years of proprietary control stack, now a Gemini API key. (Google DeepMind)

• The biggest round construction robotics has ever seen went to autonomous machines, not humanoids. SoftBank led Gravis Robotics' $200 million Series A. Diggers on real sites, with infrastructure repair and reshoring behind the demand. (The Robot Report)

• The humanoid race stopped being a demo race and became a manufacturing race. XPENG’s physical AI unit raised more than $900 million at a $6.3 billion valuation to scale its IRON platform, with mass production targeted before the end of this year. Whatever you think of the form factor, the production date is the number that matters, and right now those dates are being printed outside Europe. (TechNode)

• More capital, fewer deals, and the squeeze is landing at the early end. European deal count fell to a six-year low in the first half while total capital rose, with AI absorbing around 60% of it. Call it a rotation. It also means the gap between a good pre-seed team and its first cheque is wider this year than last, whatever the aggregate charts say. (Tech.eu)

Hiring across the portfolio

Highlighting fresh roles.

• Distance — Senior FPGA Engineer, plus six more including two defence roles (Helsinki, Bath)

• Scenario — Full-stack ML Engineer, plus nine more (Paris, London, Brussels, NY, SF)

• Iconic AI — AI Research Engineer, Model Optimization and Inference (London)

• M-XR — Research Lead, Applied Research, Machine Learning (London)

• Flow Computing — Senior LLVM Compiler Backend Engineer (Helsinki)

• ai-coustics — Systems Software Engineer, Rust and ML inference (Berlin)

Where to find us

Nordic Tech Week, Stockholm, 8 and 9 September. Petri is there for Physical AI Day and the Investor Day that follows. Say hello if you are around.

Zurich and Paris, 8 to 10 September. Sointu is in Zurich, and David is in Paris during IPEM. Say hello if you are around.

Autonomy, Malmö, 14 September. Petri will be there. William Godfrey of Tangible Finance is on stage with a session titled "When should debt enter the capital stack?", which is the argument in this month’s deep dive, in a room full of the people who have to act on it. Find them if you are going.

Levtek opens the factory, Malmö, evening of 14 September. After Autonomy wraps, our portfolio company opens their new facility: an agentic workflow closing the loop on real hardware, the production line up close, and the Levkart to drive yourself. Capacity is limited and registration needs approval. Sign up.

Bits & Pretzels, Munich, 28 to 30 September. Dave is a Table Captain at the founders' festival, which is the easiest place to get twenty minutes with him.

Minus One, Helsinki, Monday 16 November. The night before Slush Investor Day, for the fourth year running, with Google returning as partner, joined by Wilson Sonsini and Dragon Chasers. 300+ investors and LPs, north of €60B of AUM, in the room before the week starts. Request an invite.

Minus One 2026 · Luma

The legendary Minus One drinks is back.
300+ GPs and LPs in one Helsinki room, the night before Slush Investor Day kicks off. No agenda. No pitches. Just…

luma.com/g2b7lk3s

Your turn

We are looking for the first European robotics fleet that could be financed against its own uptime. Running one? Lending against one? Hit reply, we read every response.

See what others miss.
— The FOV Team

Viewpoints is brought to you by FOV Ventures, the European fund investing in the next era of computing: robotics, physical AI, spatial.

LinkedIn | Website | Pitch us

Back in March we hosted our Vantage Points ‘Robotics Investor Summit’ in London. One of our partners for that event was Tangible Finance, whose mission is to be the capital stack co-pilot for ambitious hardtech companies — helping founders understand, then structure, non-equity financing options.

Debt has long been misunderstood by the venture community, which has focused mostly on underwriting pure software. Tangible are working to change that, and it matters more every quarter as VC capital flows into hardware.

They’ve now published Hardtech’s Debt-Raising Map. We asked them to write the companion piece for the other side of the table: what debt-readiness means for those of us buying the equity.

So what’s the FOV view?

You might wonder why we, a venture investor, are deep-diving on debt and pushing our founders toward it. Dilution is our product. We buy ownership, and we’d rather buy it cheaply than expensively. So why argue for less of it?

Because we don’t think it’s about how much equity a company sells. It’s about what that equity buys. Equity priced against technical risk is exactly what venture exists for, and everyone is well served by it. Equity spent on the thousandth identical unit is something else: the priciest capital available paying for something a lender would have funded at a fraction of the cost. That trade destroys the return for the founder and for us in the same motion.

It also crowds out what equity is actually for. Every euro spent on the hundredth identical robot is a euro not spent on the next product, the go-to-market, the team that builds both. And the arithmetic gets worse as the company scales, because funding deployment from the cap table means coming back for ever-larger rounds at valuations that eventually outrun what any venture investor can reasonably underwrite. Somewhere around the hundredth unit, that price stops existing. So the companies most reliant on equity to fund their assets aren’t simply the most diluted. They’re the most likely to die.

We’re not arguing for a smaller slice of the companies we back. We’re arguing that the slice we buy be backed by the technology and the team, and that the steel be financed by someone whose whole business is financing steel.

What follows is Tangible’s piece, written for fellow investors and best read alongside their Debt-Raising Map. Let us know what you think.

A guest post by Tangible Finance

Tangible is the capital stack co-pilot for hardtech companies.

Introduction

We’re at the start of one of the largest reallocations of private capital in a generation, and the companies that will define it are being built out of atoms. Here’s the part most investors miss: whether a company can raise asset-backed finance, more than the product itself, is becoming the thing that separates the ones that compound from the ones that stall. The best of them are already scaling on capital their competitors don’t know exists.

This piece is meant to arm you. Debt-readiness is now an equity question. Every euro of deployment a company funds with debt is a euro of your ownership that survives to the exit, and every round after it gets easier to assemble. The funds that learn to pick for this, and to push for it, will own more of their winners. What follows is how we, sitting in the diligence rooms on both sides, tell the contenders from the pretenders.

I. The supercycle, and why it changes your selection problem

Capital is moving from software to the physical economy. As AI thins the moats vertical SaaS once enjoyed, the durable returns are migrating to robotics, energy, advanced manufacturing, and the infrastructure beneath them.

Supercycles like this are never funded by equity alone. They turn when asset-backed finance arrives to make deployment viable at scale. Created in 1919, GMAC (the General Motors Acceptance Corporation) turned the car into a financeable receivable. The mortgage built the housing market. Aircraft fly on leases, not on airline balance sheets.

What you’re funding is physical, so the financing problem is structurally different from the SaaS playbook you were handed. You wouldn’t let a company design its hardware and contracts without the customer in the room. In hardtech the capital markets are the second customer. So when a founder says they’ll figure out financing later, read the flag. Later means diluted.

The capital has already arrived to meet them. Private credit has crossed two trillion dollars and is climbing toward four and a half by the end of the decade, and its fastest-growing slice is asset-backed finance — the instrument built to fund exactly what hardtech builds. The buyers exist. The only question is which of your companies is designing its cashflows, processes and collateral for them, and which is still planning to sell equity to buy steel.

II. Equity and debt must be married, and that tells you what to back

Software hit $100M of recurring revenue on a few hundred million of equity because its marginal cost was near zero. Atoms can’t. Funding each unit of deployment with the priciest capital on the cap table is value destruction, plain and simple.

Software scales inside one financing ecosystem. Hardtech has to marry two. Chart by Tangible.

The objective is the lowest blended cost of capital at each stage. Over time, cheapest cost of capital wins. Venture stays essential, but its job in hardtech is to retire technical risk and get the company to the point where debt can do the heavy lifting.

Picture the journey on two axes: technical risk falling, contracted traction rising. As the first drops and the second climbs, the optimal mix should shift in lockstep, equity-heavy to debt-heavy. Equity buys the inflection. Debt carries the scale.

What this means in diligence: back companies whose capital mix is built to shift, and whose every financing leaves them positioned to raise cheaper and more flexibly next time. A company with no credible path to debt is a company that will keep coming back to you for the priciest money it can find. And not all debt is equal, either. The wrong structure can cost more than the equity it was meant to protect. Getting it right is a craft, not a checkbox.

III. Timing: it’s about repeatability, not funding rounds

If there’s one idea here to carry into your next board meeting, it’s this one. The timing of debt is not a question of which equity round a company is on. The instinct to bolt a facility onto the last raise, the way venture debt usually works, is precisely the thinking that keeps companies small. Round-tied debt scales with your equity and therefore caps out with it. The capital that funds real scale is unlocked by something else entirely.

Sum it up in one word, and the word is repeatability.

• Debt doesn’t underwrite a story or a round. It underwrites a unit it can watch repeat. The bespoke, first-of-a-kind asset belongs to equity. The thousandth identical unit — built to a known cost, performing to a known curve, sold on a known contract — belongs to a lender, who’ll fund it gladly.

• Repeatability is what lets a lender stop underwriting the company and start underwriting the asset. That shift is the whole game. It moves a company from corporate risk, priced expensively, to asset risk, priced cheaply, and it lets a facility scale without scaling your dilution.

• So the milestone that matters isn’t the Series B. It’s the moment a company can say: we’ve done this many times, here’s the cohort data, here’s the contract template, and the next unit looks exactly like the last. That’s the event debt is waiting for, and it almost never lands neatly on a priced round.

• This reframes the founder’s job, and yours. The work isn’t to time a raise to a milestone on the equity calendar. It’s to define the repeatable unit early, standardise it, instrument it, and build the company so that when repeatability arrives it’s legible and financeable: clean contracts, per-asset data, structures that lift into an SPV without a year of legal surgery.

The full stage-by-stage map is at foundations.tangible.finance.

For you as the equity investor, repeatability is the timing question worth asking in the room.

• Ask what the company’s repeatable unit actually is, and how close they are to proving it. A team that answers crisply has understood its second customer. A team that fumbles is years from scalable debt, however large its last round was.

• Push your companies to define and standardise that unit sooner than feels necessary. Repeatability is engineered, not stumbled into, and the choices that enable it are cheap at pre-seed and expensive at Series B.

• When you find repeatability proven, or nearly so, you’re looking at a company about to unlock capital that isn’t yours to provide. That, not the next markup, is the inflection that should move your conviction.

A closing word on psychology, because it’s the real obstacle. The fear of debt, and the quiet preference to push it down the road, isn’t only a founder’s instinct. Plenty of investors share it, and would rather write another equity cheque than ask a company to do the harder work of becoming financeable. That’s a mistake, and it’s worth calling one. The bar for debt is high; we won’t pretend otherwise. But a high bar is exactly why debt has to be the lens a company sees itself through from day one — the destination it’s built toward, not a bolt-on fastened to a finished business. Bolt-ons don’t work. Neither do founders who wave a hand and promise it’ll all be fine. The companies that win here treat debt-readiness as the goal, and the investors who win here expect nothing less.

IV. What the wrong sequence costs — you, and the whole category

The fastest way to see the value of sequencing is to watch a good company hurt itself without it. We see it constantly: real technology, real customers, real demand, undone by a single avoidable decision. They bake their capex into an equity round.

Consider the pattern. And it is a pattern, not an exception.

• A Series B company needs sixty million. Forty-five of it is capex: the first plant, the first fleet, the first thousand units bound for customer sites. With no debt access built in time, it raises all sixty as equity.

• The round is now too big to fill. Fewer funds can lead a raise that size, diligence drags, and the company spends nine months selling a story it should have spent building a business. Plenty of these rounds simply never close, and the company stalls, or dies, with a working product still in the ground.

• Equity is the most expensive money on the table. Funding depreciating physical assets with it — assets a lender would have financed in the low teens or below — quietly guts the returns the company is pitching. The cap table pays for steel.

• You pay for it too. Founders and early backers get diluted against assets that should never have touched the equity line, and the valuation required to make that dilution bearable sets a bar the company may never grow into. The down round is seeded years in advance.

• Then the worst part, and it lands on all of us. When the returns come in mediocre, because so much equity went to steel, they confirm the exact bias the category is fighting: hardtech is too capital-intensive, the returns are bad, the maths doesn’t work. Every bloated, mis-sequenced round that underperforms erodes the market’s trust in the whole asset class and makes the next good hardtech founder’s raise harder. The damage is collective.

The fix isn’t more conviction. It’s the sequencing above: equity to the point of repeatability, debt beyond it. Timing is as much the lever as access.

• Equity should buy the team, the technology and the commercial traction — the things only equity can buy. A facility funds the assets. Do that and the round shrinks to what equity is actually for, becomes fillable, the dilution turns sane, and the returns on the equity slice start to look like the software-grade returns this room already knows how to underwrite.

• Debt that arrives a year too late has already let the company over-raise equity. The work of becoming financeable has to begin before the capex ask does — which is exactly why the screening that follows is your job and not only the founder’s.

• The funds that get ahead of this do two things at once. They protect their own ownership, and they help rehabilitate a category that’s spent a decade paying for its own mis-sequenced rounds.

V. What you need to understand to tell the contenders from the pretenders

You know the venture path well. To pick well here, you need enough fluency in the rest of the stack to spot who’s bluffing.

5.1 The instruments. There’s no single thing called “debt.”

• Corporate debt lends against the company. Asset-backed finance lends against the asset and its cashflows.

• The menu is wide, and each instrument has its moment: leases, floorplan, project finance, instalment structures, as-a-service facilities, order-book and offtake financing, factoring, and blended structures.

• A useful frame: debt funds the inputs, credit funds the outputs.

• Much of it sits off the operating company’s balance sheet, inside an SPV built precisely so asset risk and corporate risk can be priced apart.

5.2 The lenders. Not interchangeable, and neither are their mandates.

• Banks, credit and hedge funds, family offices and pension funds each underwrite to a different cost of capital and a different risk tolerance.

• Venture debt is the one most often misread. It’s tied to the last equity round, sized at a quarter to a third of it, and doesn’t scale with the asset base. It’s a bridge, not an engine.

• Watch for the misdiagnosis: nearly every founder is convinced they need project finance, and most don’t. A team that knows which instrument it actually needs is already ahead of the pack.

5.3 What they care about, and why it isn’t what you care about.

• You can lose seventy percent of a portfolio and still win on the one position that returns a hundred times. A lender can’t. A lender needs every deal repaid.

• To you, risk is the thing you’re paid to take. To a lender, risk is the thing that has to be engineered down.

• A first-of-a-kind plant is catnip to equity and repellent to credit. The founder who understands why is the founder who’ll eventually raise both. We’ve written about this in Capital Intensity Isn’t Bad.

5.4 The two languages. A reliable tell when you’re sitting across from a team.

• To raise equity you sell upside, a share in a future that doesn’t exist yet. To raise debt you sell a contract — a defined risk for a defined return.

• The same sentence lands as optimism to a venture investor and as cashflow-and-collateral to a lender. Debt is allergic to spin. We mapped this translation in Capital Stack Parallax.

• The founders who can already speak both are the ones who’ll clear the bar when it matters.

Increasingly, the venture round itself waits on this. The Series A doesn’t get signed until a company can prove it’s capable of raising a scalable asset-backed facility at all. Debt-readiness has become an equity gating item. That’s the shift in one sentence.

VI. The hiring signal: a CFO and a capital markets operator are not the same person

The clearest early signal of all is who a company hires, and when. The serious finance leader is no longer a Series B afterthought. Increasingly the CFO is an early, even founding, hire, brought in to design the capital stack from the first contract rather than to clean it up later. No wonder. When the second customer is the capital markets, the person fluent in their language belongs in the room while the company is being built, not after it’s diluted its way through three rounds.

But look past the title. Most startup CFOs are excellent at what startups have always needed: budgets, board decks, the next equity round, the audit. Capital markets are a different discipline, and they’re the one that unlocks scale. Very few startup CFOs have ever structured a facility, sat across from a credit fund, or carry the lender network a structured raise actually runs on. That experience and that rolodex are the scarce asset. Find a hardtech company that has it — in the CFO, a founder, or a deliberate advisor — and you’re looking at a company that can be financed. When you don’t, assume the capability has to be built or bought, and price that gap into your view of how the next three years get funded.

The exercise itself is the value. The decisions compound from the earliest days: how exit clauses are written, whether a customer contract transfers cleanly into an SPV, whether the supply chain is legible enough to underwrite. Pre-seed choices echo for years.

And there’s a signalling effect in your favour. A company that matures to the point where lenders find it interesting has, in the process, become the kind of company everyone else finds interesting too.

VII. Your screening checklist: the seven things lenders underwrite are the seven things you should diligence

Whatever the asset class, lenders converge on the same short list. We offer it as a selection lens, because a company that can answer it is a company that will preserve your ownership — and most of the list is simply ordinary hygiene, the discipline a serious company would want regardless. It has to come if the company is going to succeed at all.

1. A credible equity position and continued access to equity. Do the owners have conviction, and is the next round reachable?

2. Genuine financial sophistication on the team. Can someone in the building speak to the second customer, the capital markets?

3. Sound credit and underwriting at the asset level. Do the unit economics survive scrutiny, not just a pitch?

4. Disciplined asset-level data hygiene and automated reporting. Can the company show, and monitor, what it’s asking a lender to fund?

5. Automated, reliable collections. Does cash arrive predictably enough to be financed?

6. Demonstrable volume and pipeline. Is there enough repeatable deployment to build a facility around?

7. Well-defined corporate and asset-level operations. Is the structure clean enough to isolate and price risk?

VIII. The pattern, proven

This isn’t theory. The proof points are stacking up, and they share a shape worth learning to recognise.

• Enpal tracked every panel it installed like a financial institution long before it was treated as one. That discipline unlocked Europe’s first public residential solar ABS with M&G in late 2024, around €240 million, then in late 2025 what the partners called the world’s first public securitisation to blend residential solar and heat-pump loans, around €300 million, oversubscribed several times over. Discipline first, capital markets second.

• Crusoe paired the two instruments deliberately from its Series D onward: equity for the vision, asset-backed facilities for the build. Its CEO puts the philosophy plainly — put each major category of capex under its own asset-backed facility, and reserve equity for technology, team and growth. The result is one of the fastest-scaled infrastructure balance sheets of the cycle, each facility unlocking the next.

• Elvy refused the default. Rather than burn equity to put hardware on roofs, it secured a €500 million debt facility from the platform Scayl and a banking partner, financing the rollout of solar, heat pumps and batteries almost entirely with debt. Its founder has said he never wanted to use equity to finance asset deployment, and that full debt backing for projects like this was something he hadn’t seen done before. Equity for the company, debt for the steel.

Where this leaves you

Much of what the market still repeats about hardtech is wrong. Too capital-intensive. Poor returns. Can’t raise debt early. Tangible are in the rooms, for the funds and the founders, watching that consensus get dismantled and hundreds of millions deployed to put it to rest.

The funds that learn to screen for debt-readiness, and to push their companies toward it, will compound a structural advantage. When a portfolio company finances its assets with the right capital instead of the most expensive capital, your ownership is preserved, your winners stay financeable, and your fund gets to participate in a supercycle equity alone was never built to fund.

We’re inviting you to learn to pick them — and to tell us where you think we’ve got this wrong.

Tangible exists to help hardtech companies raise from their second customer, the debt markets. They’re working with dozens of hardtech companies to help founders navigate them. Start with Hardtech’s Debt-Raising Map, or find out more at tangible.finance.

Hardtech's Debt-Raising Map is free to read at foundations.tangible.finance.

Viewpoints is brought to you by FOV Ventures, the leading European fund investing in the next era of computing.

LinkedIn | Website | Pitch us

Last month a German robotics company raised €1.2 billion in a single round. NEURA Robotics is now worth about $7 billion, backed by Nvidia, Amazon, Qualcomm and the European Investment Bank. Capital stopped being the bottleneck in physical AI. The right kind of capital became it.

Deep Dive: Why physical AI needs a different kind of capital

Why physical AI can’t be funded like software

Robotics and physical AI startups raised $27.6 billion in 2025, more than double the year before. In Silicon Valley, AI took 93% of every venture dollar. The money has arrived. Whether it can actually build robots is a separate question.

Software AI rewards speed. Ship, measure, ship again. Robotics runs on different physics. A humanoid that climbs a flight of stairs in a demo is a long way from a fleet that runs a warehouse for a year with nobody stepping in.

The autonomous vehicle sector already ran this experiment. More than $100 billion went in. A decade later, a handful of US cities have driverless commercial service. Capital alone did not collapse the timeline, and it rarely does when atoms are involved.

When generalist money meets hardware, three things tend to go wrong. Valuations drift away from technical milestones. Founders skip the unglamorous engineering to chase a moonshot demo. And early cap tables fill with story investors who go quiet the moment an actuator redesign eats six months. You can see the incentive in the pitch language: too many early-stage teams now introduce themselves as “the OpenAI of robotics” instead of explaining what they have actually built.

Specialist capital behaves differently because it has to. It understands the physics, the supply chain and the competitive map. It can sit still through a twelve-month detour into a new actuator design, because it has lived through that detour before. As Lancaster and Paseman put it, niche investors have an edge precisely because they are not trying to be everything to everyone.

That is the thesis we have been running since 2022, across 33 companies at the convergence of AI, robotics and spatial computing. None of them is a magazine cover. All of them have to work on day one. Four are below.

Read David Ripert’s full essay, Why Physical AI Needs Specialist Capital 

In the Field

Four companies from the essay above, each building physical AI that has to work in the real world, not just in a demo.

Makiina — The general factory for physical intelligence
Raw material in, different form factors of physical agents out. Makiina designs a shared manufacturing and software back end from day one, with common actuation, electronics, perception and policy layers spanning multiple robotic form factors. This month it showed its arms operating in real time and fully autonomously.

Why it matters: One back end across many form factors compresses the cost of building physical agents faster than any single-robot roadmap in the press.

🎥 Watch the arms demo below where robots are assembling their own actuators.

Twitter tweet

Levtek — Robot working companions for physical labour
Levtek’s Levkart, a ride-on autonomous robot, is in live logistics trials with PostNord, driven in part by vision-language-action models. An operator can redirect it by voice in a space it has never seen. No facility retrofit required.

Why it matters: Levtek fits the world as it is. No QR floors, no rebuilt aisles, which means a sales cycle that closes in months, not years.

Levtek is already shipping several robots to one and the same customer.

And these are just a few of many robots going out to customers these weeks

LinkedIn

Distance Technologies — Deviceless mixed reality for vehicles and cockpits
Distance signed Galvion, integrating its AR computational optics and Acuity OS into Galvion’s head systems, with market-ready products targeted for 2027. It also joined Project Orbion alongside Niantic Spatial, ICEYE and BlackSky, deployed for the first time at the US Coast Guard Aviation Training Center.

Why it matters: Defence and aerospace buyers move on multi-year evaluation cycles. That is the patient, expert-capital story in practice. Spatial computing reaches scale by upgrading the surfaces people already use, not by selling them new headsets.

🎥 Watch the Eurosatory demo below.

Distance at Eurosatory 2026

Demonstration of their Field Operator HUD integrated with the General Dynamics Mission Systems–Canada MESHnet® system.

LinkedIn

Spogen — Voice and vision AI assistants for heavy machinery
Spogen’s assistants are now running in tractors and seed drills through the EIT Food Test Farms programme, with Valtra, Väderstad, Lännen Tractors and NHK Group. Construction, forestry, mining and ports are next.

Why it matters: Heavy machinery is fragmented, slow to sell into, and gets no hype. It is exactly the kind of physical-AI problem generalist capital walks past and specialist capital underwrites.

🎥 Watch Spogen's Talking Tractor (AGCO Corporation) demo below.
📖 Read → Natalia Leinonen on machine knowledge beyond the manual.

Radar

Beyond our portfolio, here’s what caught our attention this month.

• NEURA Robotics raised €1.2 billion. The number to watch now is the deployment count, not the valuation. Europe’s largest robotics round of the year, backed by Nvidia, Amazon, Qualcomm and the EIB at roughly $7B. The Deep Dive above in one headline: the money is here, the milestones have to catch up. (CNBC)

• Europe’s industrial robots are pulling growth money now, not just seed. THEKER raised a €73M Series A out of Barcelona, led by CRV with Samsung and LVMH, to put AI robots on tier-one factory floors. The deployable-robot thesis, funded at scale. (EU-Startups)

• Jeff Bezos is now a physical-AI founder. His startup Prometheus raised $12B at a $41B valuation to build an “artificial general engineer” — AI that designs and manufactures physical systems, from jet engines to drugs. When the second-richest person alive bets his next act on atoms, this stops being a contrarian thesis.

• The world-model money is here. The benchmark below says the tech isn't yet. Odyssey, built by autonomy veterans Oliver Cameron (ex-Cruise) and Jeff Hawke (ex-Wayve), raised a $310M Series B at a $1.45B valuation, led by Natural Capital with Amazon, AMD and GV, to build world models that learn physics from the real world — for game worlds and robotics. Capital is racing into the exact frontier the benchmark below just called brittle. The edge goes to whoever lands these models on real hardware first. (TechCrunch)

• When DeepMind opens its first robotics accelerator, it opens it in London. Google DeepMind’s Accelerator: Robotics kicked off with a 15-startup cohort, several of them European. The platforms and the talent are converging on this continent. (PPC Land)

• The week the US cut Europe off from the best AI models was the strongest case for tech sovereignty yet. In June, Washington forced Anthropic to block every non-US user from its most capable models, overnight. Europe runs on AI it does not control, and a dependency that can be switched off in a day is not a foundation. The continent needs its own stack. (Al Jazeera)

• The billion-dollar world-models bet just met its first hard benchmark, and the models came back brittle. A new evaluation of Yann LeCun’s AMI research direction found today’s world models still fragile. We think that’s healthy. The category needs benchmarks more than it needs bigger seed rounds. (TechTimes)

One Thing to Watch

The teams turning world models into something a robot can actually use are the ones to watch.

Yann LeCun’s $1B raise put world models back on the front page, but the benchmark that followed (see Radar) showed today’s models are still brittle in the real world. The work we find most interesting in Europe is narrower: not a model that reasons about everything, but one compressed enough to give a specific robot a reliable guess at what happens next, on real hardware, in one environment.

We’ve taken a handful of meetings, and the signal is getting louder. If you’re shrinking a world model down to something that runs on a machine that has to work on Monday, we want to hear from you.

Hiring across the portfolio

Our companies are building, which means they are hiring. A sample of open roles this month:

• M-XR — Applied Research & Machine Learning (London)

• ai|coustics — Systems Software Engineer, Rust/ML (Berlin)

• Flow Computing — Compiler & CPU-modeling engineers (Oulu)

• Levtek — Mechanical Engineer (Malmö)

• Iconic AI — AI Research Engineer (London)

• Scenario — Full-stack ML Engineer (Paris)

• Starstuff — AI/ML Engineer (Helsinki)

Each company name links straight to its careers page — more roles listed there.

Save the date: Minus One, the day before Slush Investor and Founder Days

Google returns as a partner for the fourth year running.

On Monday 16 November, the day before Slush opens its doors, we host Minus One in Helsinki. Three editions in, it has become one of the sharpest rooms of the week: 300+ investors and operators, north of €60B in AUM, gathered before the noise starts.

It is invite-only, and the first invites go out this week. If you are on the list, watch your inbox. And book your flights accordingly — for those travelling into the city, you’ll need the earlier flight on Monday.

Your turn

We are mapping the European founders building physical AI in markets nobody calls exciting: heavy machinery, inspection, agriculture, logistics. Building one? Know someone who is? Hit reply, we read every response.

— The FOV Team

Viewpoints is brought to you by FOV Ventures, the leading European fund investing in the next era of computing.

LinkedIn | Website | Pitch us

The money is already moving, and faster than most people realise. Robotics and physical AI startups raised on the order of $27.6 billion in 2025 across more than 1,000 deals (PitchBook), more than double 2024, and 2026 is running hotter still: by mid-year the sector had already crossed $23 billion, on track to pass all of last year, with humanoid rounds setting the pace and Tesla committing $20B in Optimus capex for 2026. Figure, a single humanoid company, closed more than $1 billion last September at a $39 billion valuation (Figure), roughly fifteen times what it was worth eighteen months earlier. In Europe, Neura Robotics just raised $1.4 billion at a valuation of about $7 billion, making it one of Europe's best-funded AI startups. In Silicon Valley, AI took 93 percent of every venture dollar in 2025, $103.5 billion out of $111 billion (Mind the Bridge "Silicon Valley Bets Big on Physical AI”). Robotics, humanoids and embodied intelligence have become the new front of the AI gold rush.

The pattern is familiar. Generalist mega-funds are doing in robotics what they already did in software AI: crowding into the most obvious deals, at the highest prices, on the shortest timelines. Simon Lancaster makes the point in Unlocking Alpha: The Rise of Niche VC, where he argues the industry has “traded away focus for breadth and conviction for consensus. The result is a model that struggles to generate true alpha.”

Robotics is where that model breaks.

Robotics runs on different physics

Software-AI investing rewards speed. Back the team with the best model, scale the GPU bill, race to product-market fit. Robotics does not work that way. A humanoid that climbs a flight of stairs in a demo is a long way from a fleet that runs a warehouse for a year with nobody stepping in. Sensors, actuators, training data, certification, supply chains: every one of them needs slow, patient engineering. We have already seen where the shortcuts lead. After more than $100 billion poured into autonomous vehicles, with only a handful of US cities — San Francisco, Phoenix, LA, Austin — running fully driverless commercial service today.

The problem is not generalist capital itself. It is a mismatch of capital to the stage. When money that is priced for software speed sets the tempo for robotics, three things tend to go wrong. Valuations drift away from any real technical milestone. Founders skip the messy learning years and jump straight to deliver-a-moonshot-or-die. And the early cap table fills up with investors who are there for the story more than the substance, and who cannot help much when the hardware misbehaves.

It is not only the headline rounds. We watch early-stage robotics teams quietly rewrite their decks to ride the foundation-model wave, trading a credible story about a new gripper or perception stack for a vague line about being “the OpenAI of robotics.” That pivot might close the round. It rarely builds the company.

What specialist VC actually looks like

The fix is not less capital. It is better-matched capital. Frontier hardware needs investors who can price a deal because they actually understand the physics, the supply chain and the competitive map. Investors who can sit still through a twelve-month detour into a new actuator design, because they have lived through that detour before. Investors who know which Tier 1 OEM to call for a pilot, and which one to keep well away from.

Lancaster and Sabrina Paseman put it well: “Niche VCs have an edge because they aren’t trying to be everything to everyone. Founders can tell when an investor deeply understands their space and can plug them into a relevant network immediately. That kind of credibility wins trust, and access, faster than brand alone.”

This is how we work at FOV. We are a specialist fund backing European founders building for a new era of computing, one that is more intelligent, more spatial and more human. The convergence of AI, robotics and spatial computing has been our mandate since 2016. Across Fund I we have backed 33 companies on that thesis, and the physical AI names are where the specialist edge shows up most clearly.

To be clear, that edge is most powerful when it is paired, not isolated. We like to lead or co-lead the pre-seed, where deep domain knowledge does the most work. We also like co-investing as the specialist alongside a larger generalist fund, where we bring the technical conviction and network and they bring the firepower to scale. Some of the best cap tables in this category have both.

A few of them:

Makiina (Finland) Makiina is building the general factory for physical intelligence. Raw material in, different form factors of physical agents out.

The core idea is designing a shared manufacturing and software back end from day one with common actuation, electronics, perception, and policy layers spanning multiple robotic form factors.

Spogen AI (Finland) is bringing AI to heavy machinery, the tractors, harvesters and excavators that actually move the physical economy. Hard problem, fragmented buyers, long sales cycles. Exactly the kind of company that gets overlooked when capital is chasing the obvious narrative, and that a specialist syndicate can build patiently.

Levtek (Sweden) is developing ride-on cognitive robots that reason, learn and adapt, putting AI to work in physical labour. Embodied intelligence for real workplaces, not stage demos.

Distance Technologies (Finland) is turning windscreens and aircraft canopies into mixed-reality displays, the optics and perception layer physical AI needs to operate inside vehicles. GV led the seed round, following the conviction we had backed earlier.

None of these fits a one-line trend. They fit a long bet on the layer where atoms meet algorithms. That bet only pays off if the early investors can price it, support it and stay in the room when the work gets hard.

Why this matters for LPs

For limited partners, the physical AI cycle is the cleanest test of specialist-versus-generalist returns we have seen in years. The shape keeps repeating: a handful of mega-rounds wrapped in billion-dollar narratives, and a long tail of category-defining companies that need patient, expert money to get built at all. Underwriting the heroic version of those mega-rounds, at scale, is a lot of LP capital riding on a binary outcome.

The LPs we are talking to for Fund II keep telling us the same thing in different words. They are over-exposed to generalist funds chasing the same logos. They want specialists who see what others miss, who hold price discipline in hot markets, and who know the difference between a robotics company that should raise $5 million and one that genuinely needs $50 million. Niche is becoming a core allocation, not a hedge.

For founders the message is simpler. Money is fungible. Partner value is not. When the prototype dies at 2am, the investor who knows the field beats the one with the biggest logo.

For our peers in the generalist tier, none of this is zero-sum, and we do not treat it that way. Several of them are partners we co-invest with happily, and some are investors in our own funds. Mega-funds will and should power the growth rounds once a physical AI company is ready to scale. The point is narrower than “generalists versus specialists.” It is that the earliest rounds in physical AI reward a particular kind of expertise, and the strongest syndicates pair that expertise with the capital to scale later.

Where this leaves us

Frontier tech is too important to fund on reflex. The next decade of physical AI and robotics will be built by founders who go deep, and by investors who know how to back them at a sensible pace and a sensible price. Focus beats frenzy. That is the bet we are making with Fund II, and we think the next generation of European hardware-and-AI champions will prove it out.

Reference: Unlocking Alpha: The Rise of Niche VC by Lancaster, Paseman, Felix Staeritz, Michael Ströck (Amazon).

Viewpoints is brought to you by FOV Ventures, the leading European fund investing in the next era of computing.

LinkedIn | Website | Pitch us

The biggest physical AI bet isn't only the robot. It's the world we're rebuilding around it. Every space designed for human eyes and human hands is quietly being retuned for machines, and almost nobody is calling that what it is: the most consequential redesign of the built environment in a generation.

Reinventing Reality

How physical AI is redesigning the built world

For decades, we redesigned the digital world to be machine-legible. Web pages carry schema markup. APIs expose structured data. Phones and laptops were designed around how machines process information, not just how humans use them. That redesign was so thorough it became invisible.

Physical AI flips the same problem into three dimensions. Every street, store, warehouse, field and hospital was originally built around humans: our eyes, our hands, our walking speed, our one second of reaction time. As autonomous systems take over parts of that work, those environments stop being purely human. They become shared spaces, tuned for both people and machines. It is the physical equivalent of adding schema markup to a page.

You can see it everywhere already. Warehouses are shrinking aisles and QR-coding their floors. Retail is splitting into walk-out front-of-house and dark-store back-of-house. Streets are being rebuilt around autonomous vehicles, not the other way around. Hospitals are widening corridors for autonomous carts and reorganising nurse stations around machine-generated signal. The same arc reaches farms, ports, mines and construction sites.

Almost every physical AI bet sits somewhere on that spectrum. Both ends produce winners. The interesting companies are the ones that know exactly which end they're choosing, and why.

We see three investable layers in this shift:

1. The machines. Robots, autonomous vehicles, drones, surgical systems, agri robots.

2. The perception stack. Cameras, LiDAR, radar, tactile sensors, specialised silicon, and the foundation models that fuse them.

3. The environments. Digital twins, robot-ready infrastructure, V2X, warehouse operating systems. The software and hardware that turn a legacy space into a machine-friendly one.

Three companies, three layers

Three of our portfolio companies are already proving the thesis above.

Distance Technologies

Deviceless mixed reality for vehicles and cockpits

First consumer auto integration unveiled with Kia at Milan Design Week: a panoramic lightfield HUD live in the windshield, no headset, no glasses. The Field Operator HUD is now in NATO and allied field trials.

Why it matters: Distance is proving the perception-stack thesis. Spatial computing reaches scale by upgrading the surfaces we already use (windshields, cockpits, industrial panels), not by selling people new hardware categories.

Distance Technologies x Kia Worldwide: The World is the Interface

❝ This is not simply about adding information to the windshield. It's about redefining how digital intelligence can be integrated into the vehicle in a way that feels intuitive, refined and aligned with the future of automotive design. ❞

— Robert Oliver Samson, Senior Chief Designer, Kia European Design Center

LinkedIn

Makiina

Modular, low-cost robotic arms that learn from human demonstration

Makiina are building a full-stack, intelligent, modular robotics solution. This stack makes it cheaper and easier to address problems with robotics that simply wouldn’t have made sense before. Whilst still in stealth, the company has been catching the attention of investors who understand the huge potential here.

Why it matters: Makiina compresses the cost curve on manipulation. Cheaper arms multiplied by demonstration-based learning expands the addressable market for robotic work faster than any humanoid roadmap currently in the press.

Levtek

Robot working companions for physical labour

First production batch of the Levkart is quietly shipping to early, flagship European customers across retail, micromobility and logistics. Three different verticals out of the gate, no facility retrofit required at any of them. Levkart is in-part driven by VLA (vision-language-action) models, so an operator can say "go down the corridor, take a left, stop by the plant" in a space the robot has never seen, and it goes.

Why it matters: Levtek sits at the opposite end of the spectrum from greenfield re-engineering. The robot fits the world as it is. No QR floors, no rebuilt aisles, no fixed-cell deployment. Meaning a sales motion that is more likely to be closing in months, not years.

Radar

Beyond our portfolio, what caught our attention this month.

• "Hardware finally fits the 10-year VC model. AI compressed the cycle." Andre Retterath at Data Driven VC argues AI didn't just accelerate software — it shrunk hardware development timelines into a window that actually closes inside a fund's life. The thesis we've been underwriting for two years just became consensus.

• "Three rounds, three countries, 30 days. European warehouse robotics has its breakout cohort." Sereact closed a $110M Series B in Stuttgart, led by Headline. Smart Robotics raised €10M Series A in the Netherlands, led by Rotterdamse Havendraken with Innovation Industries. KEMARO's $20M Series B kicked off with a $5M first close in Switzerland, led by Spicehaus. The deals are happening on the continent. The dollars increasingly are too.

• "The largest seed round in European history came from someone walking out of DeepMind." Ineffable Intelligence closed a $1.1B seed at a $5.1B valuation. UK-based. Founded by DeepMind's David Silver. Europe is no longer the place where serious AI infrastructure founders leave to raise.

• "€300M of sovereign capital just changed the maths for European scale-ups." The EIC STEP Scale Up scheme's 2026 budget hit €300M, and the first eight winners were named on 27 April. Ticket sizes (€10–30M per company, targeting €50–150M total rounds) finally line up with the geography we want our founders to build in.

• "European deep tech now takes 32% of all European VC. That's not a thesis post anymore. It's the data." Dealroom's 2026 European Deep Tech Report, with Lakestar and Walden Catalyst, puts 2026 European deep tech at $20.3B — an all-time high share of continental VC. The cap-table conversation is changing in real time.

• "Spatial computing's commercial wedge isn't gaming. It's the factory worker who can see through walls." Industrial AR plus digital twins are delivering double-digit downtime reductions in early manufacturing deployments. Foundation models that interpret spatial data in real time are the unlock. The frontier is industrial, not consumer.

One Thing to Watch

The missing middle layer.

We've been spending more time recently with founders building the connective tissue between machines and environments. Digital twins that simulate a factory before a single machine moves. Warehouse operating systems that orchestrate AMR fleets in real time. V2X middleware that lets vehicles, infrastructure and edge compute talk to each other without a single OEM owning the stack.

This is the underrated category in physical AI. The machines get the headlines. The models get the funding. The infrastructure layer in between is what makes either of those investable at scale, and it's where a disproportionate share of the value will compound over the next decade. If you're building here, or you know someone who is, we want to hear from you.

The map behind the thesis

We mapped the early-stage European robotics landscape across 12 application verticals and 6 enabling stack categories. It's the ground truth behind everything in this issue.

Hiring across the portfolio

Highlighting fresh roles.

• nunu — QA Product Support Specialist, remote →

• ai|coustics — Marketing Lead, Berlin →

• Iconic AI — Tech Lead, London →

• GEEIQ — Product Manager (AI-First), London →

• M-XR — Software Engineer, London →

• Roro — UI Programmer, Remote →

• Levtek — Mechanical Engineer, Malmö →

More roles at jobs.fov.ventures →

We're actively looking at the missing middle layer: robot-ready infrastructure, digital twins, V2X, warehouse operating systems. Building one? Know someone who is? Hit reply.

See what others miss.

— The FOV Team

Viewpoints is brought to you by FOV Ventures, the leading European fund investing in the next era of computing — robotics, physical AI, spatial.

LinkedIn | Website | Pitch us

For decades, we have been quietly redesigning the digital world to be machine-legible. Web pages carry schema markup and semantic HTML so search engines can parse them. APIs expose structured data so software can talk to software. Even our laptops and phones are designed around how machines process information, not just around how humans use them.

Physical AI flips that same problem into three dimensions. Every street, store, warehouse, field and hospital was originally built around humans: our eyes, our hands, our attention spans, our walking speed. As autonomous systems take over parts of that work, those environments stop being purely human spaces. They become shared spaces, tuned for both people and machines.

That is a much bigger redesign than most of the AI conversation currently admits, and it is already underway. Warehouses look nothing like they did a decade ago. Aisles are narrower, floors are QR-coded, and lighting and ergonomics no longer cater to humans in aisles where people have stopped walking. The same logic is reaching hospitals, farms, construction sites and eventually our streets. Standardised lighting. Robot-readable signage. New architectural conventions for spaces that machines and humans both occupy. It is the physical equivalent of adding schema markup to a page.

Streets: the city rebuilt around autonomous vehicles

Swap human drivers for autonomous systems, and almost every urban design assumption becomes negotiable.

Roads, kerbs, signs, traffic lights and parking were all designed around a human driver with two eyes, two hands and roughly one second of reaction time. Swap in autonomous vehicles, delivery robots and smart sensors, and almost every one of those assumptions becomes negotiable. Lanes can be narrower. Signage can be machine-readable. Intersections can be orchestrated in real time instead of governed by a fixed light cycle. Kerbside becomes a choreography of autonomous drop-offs and robotaxi handovers. Some street furniture disappears, replaced by sensors, V2X beacons and edge compute.

The biggest shift is not the self-driving car itself. It is the city that gets rebuilt around it. This is precisely why we backed Distance Technologies: deviceless mixed reality and heads-up displays for automotive and aerospace, turning the windshield and the cockpit into a new perception layer that speaks both to drivers and to machines.

Stores: retail reshaped by physical AI

When restocking, inventory and checkout are all handled by machines, the shape of the store itself changes.

Retail was the first place most consumers met physical AI, through self-checkout and shelf analytics. The next decade looks more radical. When shelves are restocked by robots, inventory is read by cameras, and checkout is a walk-out event, the store no longer needs to look like a store. Aisles optimised for trolleys and human browsing can shrink. Dark stores and micro-fulfilment units take on the volume. The front-of-house turns into a showroom, a fitting lab, an experience. The back-of-house turns into a densely packed, highly automated fulfilment engine. The very ratio of public to private square footage flips.

Warehouses and factories: designed for robots, not forklifts

Warehouses built in the 2030s will look very different from those built in the 2010s.

This one is already in motion. Amazon, Ocado, BMW, Figure, Agility and hundreds of others are deploying AMRs, humanoids and automated picking at scale. Ceiling heights, floor markings, lighting levels, aisle widths and safety zoning all change when the dominant actor on the floor is a mobile robot rather than a forklift driver. Factories are on the same path. The cell, the line and the jig were human-scale. The next generation of physical AI blurs the boundaries between manipulation, logistics and inspection, and the architecture of production will follow.

We see this directly in our portfolio. Makiina is building lower-cost robotic arms for industrial use, expanding the addressable market for manipulation. Spogen AI layers intelligence onto heavy machinery so the same physical asset can operate more safely and more autonomously in a shared human and machine environment. Distance's perception stack is as relevant on a factory floor as it is on a motorway.

Fields, hospitals and beyond: physical AI everywhere we work

Precision agriculture is not just a software upgrade, it is a reshaping of how a field is laid out.

The same pattern repeats almost everywhere humans still do physical work. In agriculture, autonomous tractors, weeding robots and drone swarms turn row spacing, crop patterns and even field boundaries into variables rather than constants. Spogen AI is a good illustration here too: bringing modern AI to heavy machinery is as meaningful for a harvester in a field as it is for a loader on a building site.

Corridors widen for autonomous carts, theatres are designed around robotic arms, nurse stations reorganise around machine-generated signal.

In hospitals, surgical robots, delivery carts and ambient clinical intelligence will reshape floor plans, with even the nurse station reorganising around machine-generated signal rather than paper charts.

Construction sites, ports, mines, energy facilities and logistics hubs all follow the same arc. Each one is a large physical estate originally designed for human labour. Each one is becoming a robotics estate.

None of this happens overnight. But the direction is unmistakable. Once an environment is shared between humans and machines, the optimum design is no longer purely human.

Two ends of the physical AI spectrum

There is a useful question hiding inside all of this: do we want a world that has been re-engineered for robots, or robots dexterous enough to handle our world as it is? Almost every physical AI bet sits somewhere on that spectrum, and where a startup lands says a lot about its thesis.

At one end you have re-architecting the environment. New warehouse layouts, new theatres, new fields. Deep, durable, slow, and only really pays off where the operator is buying fresh greenfield capacity. At the other end you have machines that show up and just work in the messy world we already live in. Easier go-to-market, faster wedge into existing customers.

Levtek is a clean example of the second approach in our portfolio. The robot is deliberately designed to slot into existing warehouses and the way humans currently work, with autonomy added as a layer rather than as a precondition. That makes the sales motion a lot easier than competing end-to-end systems, which tend to need a brand new facility to be built around them. The wedge today is a smarter robot that fits the environment. The longer arc is increasing autonomy on top of that wedge.

Both ends of the spectrum will produce winners. The most interesting companies are usually the ones that know exactly which end they are choosing, and why.

Where the investable physical AI opportunity sits

At FOV we see three layers of opportunity in this shift, and we are actively backing founders across all of them.

• The machines themselves: robots, autonomous vehicles, drones, surgical systems, agri robots. The physical half of physical AI.

• The perception stack: cameras, LiDAR, radar, tactile sensors, specialised silicon, and the foundation models that fuse them. This is exactly where our spatial computing thesis meets our robotics thesis.

• The environments themselves, which we think are underrated. Digital twins, robot-ready infrastructure, V2X, warehouse operating systems, the software and hardware that turn a legacy space into a machine-friendly one.

You cannot deploy a million robots into a world that was never designed for them, and the companies building that missing middle layer will capture an enormous share of the value.

We recently mapped the early-stage European robotics landscape across those first two layers, the machines and the perception stack, into twelve application verticals and six enabling stack categories. It is the ground truth sitting behind this thesis, and a useful starting point for anyone looking at where European physical AI is actually being built. Explore the European robotics landscape map.

Reality, in other words, is up for reinvention

We started with the digital world. The web, the API, the laptop and the phone were all quietly redesigned over the last few decades to make machines first-class users. The same redesign is happening to the physical world. Streets, stores, factories, fields and hospitals are being rebuilt to accommodate intelligent machines, and a new generation of robots is being designed to fit the world we already have.

That is a generational redesign of the built environment, driven by a generational shift in who and what operates inside it. It is one of the most consequential places to be investing right now.

If you are building in this space, at any of the three layers above, come talk to us.

Welcome to Viewpoints, news and investment activity covering the next era of computing - brought to you by FOV Ventures.

This Month:

• European robotics scales: Sereact's $110M raise and TUM's RoboGym

• Apple at 50: John Ternus on spatial computing's inevitability

• Portfolio wins: Distance x Kia, Scenario MCP, nunu's William launch

Dave Haynes, Petri Rajahalme, Sointu Karjalainen & David Ripert (FOV Ventures)

The last 18 months have seen European robotics move from research-heavy to infrastructure-heavy. The capital is flowing into the enabling layer - AI models, training facilities, and sovereign data platforms that let robots learn faster and deploy more reliably.

April brought two significant announcements that illustrate this shift. Sereact raised $110M in Series B to fund Cortex 2, an AI model that trains robots on different physical behaviours and picks the one most likely to succeed. The focus is narrower, they don't build robots or sell services, just the AI model that runs across any hardware - single arm, dual arm, humanoid, fixed cell. Meanwhile, Neura Robotics and TUM launched the TUM RoboGym, Europe's largest Physical AI training center, with €17M committed and 2,300 m² of space where humanoid robots will train under real-world conditions from mid-2026. The data feeds Neuraverse, Neura's hardware-agnostic training platform, creating a sovereign data flywheel for European robotics.

We spoke with Dr. Máté Péntek from TUM Venture Labs, who's building this infrastructure, about Europe's advantages in physical AI and where the commercial opportunities are. Read the full conversation here

These are part of a broader pattern where the constraint in robotics is shifting from "can we build capable hardware" to "can we generate enough high-quality training data to make robots work reliably in unstructured environments," this month has also seen NVIDIA continuing to cement its position as the platform layer beneath robotics with new open models, and Japan committing $6.3B to physical AI as a matter of national survival.

We wrote last month about how the infrastructure layer is being built in real-time across geographies, with Europe's advantage is in manufacturing expertise, deep tech and operational environments that generate training data US labs can't easily replicate.

That infrastructure push extends beyond robotics into how we interact with intelligent systems in the physical world more broadly. Apple's 50th birthday this week brought that into sharper focus. John Ternus is stepping in as CEO with a hardware background/ And in his first interview as incoming CEO, he said Apple is still in the "early innings of spatial computing" and that combining the digital and physical world is an "inevitability."

He didn't give a timeline for smart glasses but reiterated Vision Pro as "a peek of the future." With Ternus' hardware background, the question now is which form factor wins and when.

FOV NEWS

Pathways: The Female LP Report 2026

Pathways 2026: the first benchmark on female LP participation in European venture capital.

There's a piece missing from the diversity conversation in venture: the LP layer. The people writing cheques to fund managers. They've never been measured.

We wanted to do our part. So we ran the first survey on female LP participation in European venture capital. Six months across 21 countries. Starting with data.

We expected a confidence gap. We found a design gap. 8 in 10 of the women we surveyed say they're ready to invest in venture capital today. Almost none have access to a structure that fits.

Launched live at the Pathways Summit in Helsinki with Tesi and 15 partners.

Europe's Physical AI Playbook: Q&A with Dr. Máté Péntek, TUM Venture Labs

We spoke with Dr. Máté Péntek from TUM Venture Labs about Europe's position in physical AI and robotics. Máté is at the center of the infrastructure buildout happening across European universities - including the TUM RoboGym we mentioned above.

The conversation covers why the built environment is the right anchor for physical AI, where Europe has credible advantages (precision engineering, industrial relationships, infrastructure expertise), how sovereignty and dual-use change the strategic picture, and why moving from research to revenue is the priority now.

FOV Portfolio News

• Distance Technologies announced their collaboration with Kia to integrate Distance's panoramic Lightfield HUD inside the Kia Vision Meta Turismo concept car. Distance's true 3D lightfield places images, graphics, and driving intelligence naturally within the real world at correct depth and distance. The windshield becomes the interface with the visual intelligence layer moving from smartphones and headsets into environments we already inhabit. more here

• Scenario opened their MCP server to public beta, connecting AI agents directly to Scenario's full generation stack - images, video, 3D, audio, hundreds of models, custom trained models, assets, and workflows. Works with Claude Desktop, Claude Code, Cursor, VS Code, and any MCP-compatible environment. more here

• nunu.ai launched William - the first AI agent with a phone. No SDKs, no integrations, no code. William sees the screen, taps, scrolls, types, and completes tasks end-to-end on real devices. Major gaming and enterprise companies are using it in production for QA, compliance, and market research workflows. The public beta is live now. more here

• M-XR launched their image-to-material foundational model. Input a clean 3D mesh and get six PBR maps back. The foundation is M-XR Marso measured data from years of physically capturing real-world materials. That measured PBR data is the moat, separating it from synthetic, scraped, or non-standardized alternatives. more here

• Ray Browser opened unlimited credits. Ray Unlimited lets you describe any app or workflow and builds it with no code. Gmail, Drive, Slack connected in a single prompt. The browser as AI interface layer is underappreciated, and this is the fastest way to understand why it matters. more here

• Flow Computing has been meeting with customers and investors across Silicon Valley. Strong signal: growing interest from both customers and investors, highly engaged technical conversations, and multiple discussions moving forward. Particularly encouraging traction with US-based investors focused on semiconductor and computing technologies. more here

• Spogen evolved from AI-powered machinery assistance into a broader operational knowledge platform for industrial equipment. Through customer deployments, they've seen how resource-intensive it is for industrial teams to build and maintain the structured operational knowledge that modern support and AI systems depend on. The platform now covers the full lifecycle from creation to delivery and continuous improvement. New website reflects this evolution. more here

FOV Job Board

• Roro - UI Programmer

• ai|coustics - Talent Pool: Audio Data Engineering & Machine Learning (Berlin, Germany, Mid-Senior Level)

• Waveline has multiple positions: Growth Intern (United States, Internship) and QA Product Support Specialist (Remote, Entry Level)

• GEEIQ - Client Manager (London, UK, Mid-Senior Level)

• Iconic AI has multiple positions in London, UK: AI Research Scientist/Engineer for Expressive Embodiment, AI Research Scientist for Open-Endedness & Reinforcement Learning, and AI Research Engineer for Model Optimization

• Levtek - Mechanical Engineer (Malmö, Sweden, Senior)

• M-XR - MTS Software Engineer (London, UK, Mid-Senior Level)

View all open roles here

Following the FOV Vantage Point Robotics summit, we sat down with one of our fireside speakers to unpack the European Physical AI thesis. Below is the conversation, lightly edited for clarity.

Dr.-Ing. Máté Péntek is the Director of the Built Environment focus area at the TUM Venture Labs, where he eagerly supports research and technology transfer from academia to the market through the entrepreneurial endeavour of startups. His background encompasses research on applied supercomputing to structural and civil engineering, consulting, and startup involvement.

TUM Venture Labs have established themselves as a premier European deep-tech hub, supporting over 1,100 start-up teams in 2024 alone and contributing to a record 103 spin-offs. Ranked #1 in Europe for three consecutive years (2024-2026), they foster major successes, including the unicorn Isar Aerospace (reaching unicorn status in 2025) and numerous AI and health-tech ventures.

Q1. Why is the built environment the right reality anchor for Physical AI?

Physical AI becomes most meaningful when it leaves the lab and enters complex, constrained, high-value environments. Construction sites, existing building stock, infrastructure networks, logistics hubs, tunnels, industrial assets, and cities are exactly that.

From this perspective, Europe's robotics opportunity may be less about general-purpose humanoids and more about task-specific systems: inspection robots, maintenance drones, retrofitted machinery, teleoperation, assisted surveying, concrete spraying, painting, demolition, and infrastructure monitoring.

The built environment forces the question: can Physical AI work under real-world constraints?

Q2. Where does Europe actually have a credible advantage, and what does it look like in practice?

This is where Europe has a credible edge. We have world-class technical universities, deep industrial relationships, precision engineering capabilities, dense infrastructure, demanding customers, and a strong need to retrofit, reuse, and upgrade existing assets.

Munich and TUM are one example of this advantage in action: robotics and AI research, venture-building infrastructure through TUM Venture Labs and UnternehmerTUM, strong industrial proximity, and a startup ecosystem increasingly focused on deeptech translation.

Alongside the frontier technology narrative, this matters because Europe's Physical AI thesis is strongest when it is tied to real industrial demand.

Q3. How do sovereignty and dual-use change the strategic picture?

The third layer is strategic. Robotics, autonomy, sensing, AI infrastructure, cyber-physical systems, and dual-use technologies are becoming central to sovereignty, resilience, and infrastructure security. What used to be niche research is now part of Europe's strategic capability stack.

That creates a moat: not just technical IP, but ecosystems that can connect research, founders, industrial customers, public mandates, and capital.

Q4. So what is your verdict on moving Europe from research to revenue, and why do convenings like Vantage Point matter?

Europe does not lack technical depth. It lacks enough speed, connectivity, and scale in turning that depth into category-defining companies. This has been acknowledged by many, repeated by even more. So the priority now is clear: move faster from research to revenue, from pilots to production, and from ecosystem potential to venture-scale outcomes.

That requires early-stage support for technical founders, more ambitious industrial pilots, capital that understands deeptech timelines, and stronger European collaboration across universities, startups, corporates, investors, and public actors.

My takeaway: Europe has the ingredients to shape Physical AI. The question is whether we can connect and scale them quickly enough.

That is why convenings like Vantage Point matter. They aim to turn a fragmented European discourse into a shared ecosystem and investment thesis.

You can find a recap of FOV Ventures's Vantage Point: Robotics Edition 2026 here.

In 2025, we started asking a simple question: why aren't more women investing in venture capital funds?

Not as founders. That conversation is well-documented. As Limited Partners. The people who write the cheques that fund the funds.

Six months later, with respondents from 21 countries, I think we got a more useful answer than the one we expected. We expected to find a confidence gap. The data shows something else.

8 in 10 say they're ready to invest in venture capital today. Almost none have a way in. Almost none have been told where to start. The barrier isn't confidence. It's a design gap. The LP layer between angel cheques and institutional fund commitments hasn't been built.

IT'S NOT CONFIDENCE. IT'S DESIGN.

Before we ran the survey, the working assumption I'd built up across a lot of conversations was that the answer would be some mix of confidence, knowledge, or risk appetite. Get women more comfortable with the asset class, and the participation gap closes. The data didn't back that up.

76% of respondents come from finance, VC, or angel investing backgrounds. The confidence-scale gap between active LPs and explorers is 0.38 points on a five-point scale. Basically noise. And the famous cliff (94% would invest at €10K, 38% at €100K) doesn't show a willingness gap. It shows a regulated ticket size that almost no fund actually offers. And the bigger frame: 81% of respondents want in. 3.6% have a way. A 77-point access gap that confidence can't explain.

So if it isn't confidence, what is it? It's a design gap. Specifically: ticket sizes set above where most of this capital actually sits, regulatory tiers where a €10K position in a diversified VC fund triggers higher friction than a €50K position in a single startup, FoF concentration preferences that filter out smaller cheques before discovery even starts, and a network-gated discovery layer where the people who hear about funds are mostly the people who already know other LPs.Not just women. Not just VC.

Not just women. Not just VC.

This isn't a women-in-VC story. It's a missing-layer-of-European-VC story. Sub-€100K LP capital is where most emerging LPs want to start. It's also where a lot of operators, angels, and family-office juniors sit.

Feeder funds, syndicates, and pooled vehicles do exist. But they're scattered, network-gated, and rarely surface outside an existing LP circle. And VC fund LP investing isn't angel investing. 10-year lockups, real loss risk, slow capital return. The bar is real. The demand is too. 80% of respondents are interested in a pooled structure. 72% prefer pooled over direct. The supply hasn't been built for the entry tier.

This is a snapshot, not an answer. A start, not the last word. The point isn't that Pathways has solved anything. The point is that we're now measuring it, together with 15 partners. We hope everyone working on this takes part in what comes next.

THREE ASKS

There's real capital looking for a way in. The market exists. The product hasn't been built.

Here are three things I'm asking the people who design this layer to do next. The full report goes deeper, with audience-specific recommendations for GPs, LPs, FoFs, ecosystem, and policy.

1. Measure what's missing.

   No European body currently tracks LP diversity. Preqin, Invest Europe, the national VC associations all measure GP-side composition extensively. The LP side has been a data void. That void protects no one. It also costs FoFs the ability to read GP LP-base concentration as a portfolio risk signal, which matters more, not less, as the cycle keeps testing how stable funds' capital sources actually are. The fix is procedural, not philosophical: make LP-base composition a routine diligence question, the way LP geography or vintage already is. What gets measured gets managed. What isn't measured can't be priced.
   

   "This data makes clear that the issue isn't a lack of diverse capital, but a lack of access."

   Meghan Stevenson-Krausz, CEO, Diversity VC
   

2. Design for the entry tier.

   60% of respondents cite minimum ticket size as the #1 barrier. Twice the next factor. Most have the wealth to invest more. They want a smaller way in first. That's a fund-design choice, not a market constraint. Lower-ticket structures already exist: feeder funds, pooled vehicles, syndicate layers. They need broader distribution, not just network-gated access. A €25K LP today is a €100K allocator in five years and a €500K commitment in a decade. This is a graduation path, not a destination. GPs who treat it that way build a more diversified, more durable LP base than GPs who don't.
   

   "Experienced women who want to invest in funds but are locked out by EUR 100K minimums and network-gated access. This research has measured what the industry has preferred not to see."

   Matt Penneycard, Co-founding Partner, Ada Ventures
   

3. Make the pathway visible.

   82% of respondents have never received guidance from a bank or wealth manager on VC investing. Fund discovery runs almost entirely on personal networks. If you're not already connected, you don't hear about opportunities. The fix has two parts. First, surface the diverse LPs who already exist, so aspiring LPs can see the pathway is real. Second, build the structured education and community on-ramps that personal networks have so far monopolised. Pathways aren't aspirational when they're invisible. They're aspirational when they're named.
   

   "Diversity in the LP industry is worse than in VC itself. The first step for change is always data — and we need recognition of the diverse LPs that already exist to show that there is a pathway available to aspiring LPs."

   Anisah Osman Britton, Unlock VC

One program already exists. The LP Institute runs a two-month cohort for accredited investors exploring fund-LP investing. Small, peer-led, invitation-based. They've kindly extended their application deadline for the Pathways audience. Mention FOV on the form. The answer to "what does a pathway look like?" should be a real program, not a report or a blog post.

WHAT HAPPENS NEXT

This is only the beginning. The data is published. The asks are concrete. What happens next depends on whether the people who design the LP layer treat its current shape as a market opportunity or a market failure.

I think it's an opportunity. The capital is there. The interest is there. Nobody has built the structures around it yet.

Read the full report below.

Welcome to Viewpoints, news and investment activity covering the next era of computing - brought to you by FOV Ventures.

This Month:

• Vantage Point: Robotics Edition recap and key takeaways

• FOV Robotics Interactive Market Map 2026 now live

• Doublepoint acquired by Oura - our first Finnish exit

Dave Haynes, Petri Rajahalme, Sointu Karjalainen & David Ripert (FOV Ventures)

We just wrapped Vantage Point: Robotics Edition in London - 300 investors, founders, and operators, 25 speakers, 6 panels, and 14 startup showcases.

The conversations showed how much has shifted in 18 months. Early-stage robotics companies now have customers, deployment pipelines, and paths to unit economics. The focus has moved to scaling - capital structure for hardware-intensive growth, building manufacturing partnerships, and navigating enterprise procurement cycles.

We published a full event recap covering the key sessions: NVIDIA and DeepMind on infrastructure readiness, the research-to-revenue translation problem, where capital is actually flowing, dual-use as default strategy, and what real deployment looks like across logistics, construction, and healthcare.

Recap - VANTAGE POINT: Robotics Edition 2026

300+ robotics investors, founders and operators.

Viewpoints

FOV NEWS

Portfolio Exit: Doublepoint → Oura

Doublepoint has been acquired by Oura for an undisclosed amount. One of our earliest bets from Fund I.

We backed Ohto, Jamin, Lauri and Eemil at seed when they were building gesture recognition tech that many couldn't yet see a use case for. Today they're joining Oura, an $11B company to build the next generation of human-device interaction.

Ohto was an obvious first call (and our first Finnish founder investment) when we launched FOV Ventures in 2020. We completely bought into Doublepoint's vision that new human-computer interfaces would be fundamental building blocks for the next era of computing.

Huge validation for the Finnish deep tech ecosystem, and you can read more about the decision to join Oura on Ohto’s personal blog.

Pathways Summit: The Female LP Report Launch May 6, Helsinki

We're launching the first data-driven report on female LP participation in European venture capital. Over six months, FOV Ventures, Tesi, and partners across 16 countries surveyed women about their relationship with venture capital as an asset class.

Pathways Summit: The Female LP Report Launch · Luma

This is an early invite to the launch of the first data-driven report on female LP participation in European venture capital — launched live in Helsinki.

luma.com/yc41kyjk

The event includes:

• Report launch with key findings and data.

• Workshops and working sessions, including a live DD simulation.

• Networking and facilitated introductions.

This is for women exploring investing in VC funds, GPs building a broader LP base, institutional investors, and anyone working to widen the European VC funding base.

FOV Robotics Market Map 2026 Now Live

We published our European Robotics Market Map - mapping where the next generation of physical AI companies are being built across 12 application verticals and 6 enabling technology layers.

The map covers industrial robotics, logistics, healthcare, agriculture, construction, defence, and the full enabling stack from hardware components to AI models. It's a collaborative resource - contribute companies, corrections, or updates directly.

This is a collaborative map that will continue to evolve and expand - contribute companies, corrections, or updates here.

ICYMI: Europe's Deep Tech Frontier

Last month of Vantage Point, David Ripert published an in-depth analysis on why Europe is positioned to lead in physical AI and robotics. The piece covers Europe's research-to-real-world advantage, structural advantages in spatial computing, and the case for European technological sovereignty.

FOV Portfolio News

• Scenario opened their MCP server to Public Beta, connecting AI agents directly to Scenario's full generation stack - images, video, 3D, audio, hundreds of models, custom trained models, assets, and workflows. More here

• Edailabs (Emma) was named to the VivaTech Top 100 2026 - the only edtech company selected from 100 startups across 12 countries and 28 sectors. Emma now has the 2nd highest retention rate globally among language learning apps, ahead of Babbel and SpeakAI. More here

• ai|coustics launched Quail Voice Focus 2.0, their most powerful speaker isolation model yet. It delivers 10-43% reduction in Word Error Rate under real-world conditions by stripping competing speakers, background noise, echo and codec artifacts before audio reaches ASR. More here

• Flow Computing was selected for the Deep Tech Momentum Watchlist from 5,000+ startups and will be at DTM26 in Berlin meeting with investors, partners, and industry leaders across the deep tech ecosystem. More here

• Starstuff hit 35M views on TikTok (+3M on Instagram) while completing a major infrastructure overhaul. They're preparing to launch their creator program and UGC marketplace, with significant user growth coming from Japan and South Korea. More here

• Move launches Genesis on April 8th - their next-generation inertial motion capture system. Real-time data and post solve, exponentially higher quality than previous systems. Markerless motion capture is here. More here

• Graswald released AI 2.0, rebuilding their platform to become a core operating system for fashion imagery production. The new version organises everything into projects, treating each product as a trackable task. More here

FOV Job Board

• Nunu ai has multiple positions: Growth Intern (United States) and QA Product Support Specialist (Remote)

• ai|coustics has openings in Berlin, Germany: Talent Pool for Audio Data Engineering & Machine Learning, and Content & Growth Marketer

• GEEIQ - Client Manager (London, UK)

• Atopia Space - SDR / Outbound Sales Freelancer (Remote or Munich, Germany)

• Iconic AI has multiple AI Research positions in London, UK across Expressive Embodiment, Reinforcement Learning, and Model Optimization

• Levtek - Mechanical Engineer (Malmö, Sweden)

View all open roles here

We brought 300 investors, founders, and operators together in London on March 25th. We had 25 speakers, 6 panels, 2 keynotes, 14 startup showcases, and even several robots showed up…!

FOV Ventures VANTAGE POINT: Robotics Edition 2026

LinkedIn

The last 18 months has seen a big shift in Robotics, many of the early stage companies raising capital now look very different, with customers, deployment pipelines and clear paths to unit economics.

This is is because as the cost of parts has been dropping, technical capabilities have caught up - robots can navigate warehouses, perform precision tasks, and operate in semi-structured environments reliably enough that companies are buying them.

Our latest Vantage Point event focused on this dawn for Europe in robotics and it was great to host many conversations where founders are working through capital structure for hardware-intensive scaling, figuring out manufacturing partnerships, and navigating procurement cycles. Likewise, there are more investors looking at the sector are trying to separate real traction from research projects dressed up as startups.

At FOV, we've been investing in spatial computing technologies for years - computational optics, 3D sensing, real-time spatial mapping, depth perception. That infrastructure is now essential for robots to navigate physical environments, positioning us at the convergence of AI, spatial computing, and robotics. It was good to bring together founders, investors, and LPs working across this stack.

The Infrastructure Is Ready

Mozhgan Kabiri Chimeh (NVIDIA) opened with a clear read on where we are post-GTC '26: the tooling, compute infrastructure, and simulation environments that startups need to train and deploy physical AI systems are finally production-ready. NVIDIA's partnership with cloud providers like Nebius is building the sim-to-real stack that lets robotics companies tackle data bottlenecks and edge cases without burning through capital on physical infrastructure. Daniel Woods from Nebius walked through how their platform handles the full workflow from synthetic data generation to deployment, cutting iteration cycles from weeks to days for companies building robot foundation models.

Svetlana Grant (Google DeepMind) built on this in her fireside with Dave Haynes, laying out the path to general-purpose robots. The message was, breakthrough robotics companies will be built with frontier AI labs, not isolated from them. When the compute and tooling are available, the bottleneck shifts from "can we build this" to "can we turn this into a real business."

That's where the conversation naturally moved - from technical readiness to commercial reality. Europe produces world-class robotics research, but converting that into companies that can raise, scale, and compete globally remains the challenge. The "From Research to Revenue" panel featuring Kirsty Lloyd-Jukes (Stateful Robotics), Alex Li (UCL/Cortica), Svetlana Grant, and Muji Ahmedi (ARIA) dug into exactly that problem.

The Translation Problem

Talent gravitates to the US because the pay gaps are real and the funding rounds are larger. Early-stage European rounds rarely exceed £3.5M for robotics startups, while US equivalents routinely hit $10M+. The deep tech investor base in Europe remains thin compared to the capital flowing into fintech or enterprise SaaS. These aren't new observations, but what's changing is the response.

ARIA is funding high-risk ideas that traditional VCs and grants won't touch, deliberately backing projects that might fail but could unlock entirely new categories if they work. Universities like Oxford and UCL are making it easier to spin out companies, dismantling bureaucratic barriers that historically made academic founders wait months just to incorporate. Recent successes like Wayve raising $1.5B and Oxa demonstrating autonomous deployments prove the model works when the capital arrives at the right stage.

Kirsty's take was pragmatic - stop pretending Europe can ignore the US. Use the relationship strategically through dual listings, talent arbitrage, and market access. The UK-US connection becomes an advantage if you build for it rather than around it, which set up the capital conversation naturally, because once you accept that European companies will interact with US markets and investors, you need to understand how capital flows into this sector.

Following the Money

The investor panel with Troy Estes (NVIDIA), William Godfrey (Tangible Finance), and Dominic Keen (Britbots) tackled funding reality directly. European rounds remain smaller than US equivalents, but the capital structure conversation is evolving beyond pure equity. William explained how Tangible Finance is opening up debt instruments for hardware-intensive robotics companies, letting founders scale deployments without diluting equity every time they need to manufacture more units. Robotics companies that prove unit economics can now access asset-backed financing and equipment loans to fund fleet expansion - capital tools that didn't exist for this sector five years ago.

Troy's perspective from NVIDIA was - Europe's edge lives in specialist systems solving real industrial problems with domain expertise that's hard to replicate. Manufacturing automation, precision agriculture, infrastructure inspection - sectors where Europe's operational knowledge creates defensible moats.

Dual-Use as Default

The defence panel with Rosalyn Moran (Stanhope AI), Samuel Burrell (Expeditions Fund), and Gui Wainwright (Occam Industries) made the case that dual-use isn't optional anymore - but how you build viable robotics companies in Europe. The continent's sovereign imperative to rebuild defence capabilities is creating sustained procurement demand for autonomous systems, and companies building for both defence and commercial markets aren't hedging their bets.

This focus on real-world deployment conditions carried through to the logistics and built environment panels, where the conversation shifted from what's possible to what's actually being deployed and why.

Where Robots are Deployed Today

The logistics and warehousing panel brought in operators and system integrators who deploy robots daily. Most deployments still happen in structured environments - warehouses, distribution centres, controlled production lines - because these settings offer predictable conditions and economics that justify the investment.

Construction, healthcare, and outdoor environments present harder challenges. The panel featuring Dr. Máté Péntek (TUM Venture Labs) and Ali Haddad (XRLabs) explored what it takes to deploy physical AI where you can't engineer the environment around the robot. Healthcare robots operating in hospital wards need to navigate unpredictable human behaviour, equipment left in hallways, and strict safety protocols. Construction sites change daily, with materials moved, terrain altered, and weather conditions shifting.

The through-line across all these conversations is economic viability. The technology has reached the point where capability isn't the primary constraint, which brings us to the startup showcases and what we're actually seeing in the market.

What We're Seeing

Fourteen companies demoed live during the showcase sessions, and the quality reflected how far the European robotics ecosystem has matured.

These companies have customers, revenue, and deployment pipelines actively delivering value in real operational environments. That shift is significant because it we've moved past the "is this even possible" phase into genuine commercial execution.

Europe has the talent, research institutions, industrial demand, and increasingly the capital infrastructure to build global robotics leaders. The Vantage Point conversations showed that the ecosystem has matured past theoretical discussions about whether this is possible. The real work now is execution - helping founders navigate the capital stack, connecting technical breakthroughs to commercial applications, and building the networks that let everyone learn faster from each other.

That's what Vantage Point was built for, and we'll be back!

Explore the European Robotics Landscape: We've published an interactive market map covering 12 application verticals and 6 enabling technology layers. It's a living resource - contribute companies and corrections directly. HERE.

Read the Full Analysis: Ahead of the event, we published an in-depth piece on why Europe is positioned to lead in physical AI and robotics. HERE.

We built this European robotics landscape to map where the next generation of physical AI companies are being built - and why Europe has genuine advantages in making it happen.

Europe leads in industrial robotics, surgical systems, and precision automation. German companies dominate factory floor automation. Swiss precision robotics sets global benchmarks. The UK produces surgical robotics competing directly with US incumbents. Built on decades of manufacturing heritage, engineering standards, and university research now accelerating into commercial deployment.

At FOV Ventures, we see robotics as part of a broader "beyond screens" paradigm where spatial computing, AI, computer vision, and edge computing converge to enable machines that perceive and interact with the physical world. Technologies originally built for mapping human environments (XR, vision, foundation models) are now the foundation for robots operating in unstructured, real-world settings.

The companies on this map are building at that intersection.

This is a collaborative map - contribute companies, corrections, or updates here.

LINK TO FULL MARKET MAP HERE

The landscape is organised in two layers.

At the top are twelve application verticals - where robots are actually being deployed, the industries being transformed, and the problems being solved.

Underneath sits the enabling stack: hardware components, software infrastructure, and intelligence systems that make those applications possible. Innovation flows upward through this stack. Better sensors enable better perception; better perception enables more capable autonomy; more capable autonomy unlocks deployment environments that were previously out of reach.

Understanding both layers - and how they connect - is essential for understanding where the real opportunity lies.

Scope & Approach: This report maps out the European robotics startup landscape and proposes a taxonomy by application domains and enabling technology layers. We focus on startups (privately-held companies, typically early to growth-stage) and exclude large corporates or public companies, except for a few we consider “European champions.” It is not a comprehensive list and the decision to include which companies is a subjective one. Feel free to reach out to us if you’d like to make sure you’re included.

Market Map Contents

• Layer A: Applications

  • A1. Industrial & Manufacturing Robotics

  • A2. Logistics, Warehouse & Last-Mile Robotics

  • A3. Service Robotics

  • A4. Healthcare & Medical Robotics

  • A5. Agriculture, Forestry & Food Systems

  • A6. Construction, Infrastructure & Built Environme …

  • A7. Drones & Autonomous Vehicles

  • A8. Humanoid & General-Purpose Robots

  • A9. Defence, Security & Dual Use

  • A10. Mobility & Autonomous Transportation Robotics

  • A11. Science, Research & Laboratory Robotics

  • A12. Domestic & Consumer Robotics

• Layer B: The Enabling Stack

  • B1. Robotics AI, Models & Agents

  • B3. Autonomy, Navigation & Localisation

  • B4. Human-Robot Interaction & Interfaces

  • B5. Robotics Middleware, OS & Infrastructure

Building in any of these areas? We'd love to hear from you - and if you're in London on 25 March, join us at Vantage Point: Robotics Edition, our investor summit and startup showcase for Europe's frontier robotics founders. Register here.

Layer A: Applications

A1. Industrial & Manufacturing Robotics

INTERACT WITH THIS MARKET MAP HERE

Europe's manufacturing base is one of its most significant competitive advantages in robotics - and one of its most urgent deployment environments. Germany alone has 429 industrial robots per 10,000 workers, one of the highest densities in the world, and the continent's SME-heavy manufacturing sector creates constant demand for more flexible, more affordable automation. The frontier has moved beyond rigid, single-purpose industrial arms toward collaborative robots that work alongside humans, AI-powered systems that handle variation across SKUs, and no-code programming platforms that let operators retrain robots without external expertise. For the majority of European manufacturers running high-mix, low-volume operations, this flexibility is the unlock that makes automation viable for the first time.

A2. Logistics, Warehouse & Last-Mile Robotics

Logistics was one of the earliest commercial beachheads for robotics in Europe and remains one of the most active. The continent produced some of the world's first warehouse robotics unicorns and continues to attract significant investment across autonomous mobile robots, robotic picking, yard automation, and last-mile delivery. The frontier has shifted to harder problems - unstructured environments, mixed SKU picking, and the genuine complexity of last-mile delivery in European urban settings. As e-commerce volumes grow and warehouse labor tightens, the ROI case only strengthens. Europe's strength here is in systems integration: combining hardware, software, and operational know-how into deployments that actually scale.

A3. Service Robotics

INTERACT WITH THIS MARKET MAP HERE

Service robotics covers a wide range of commercial and consumer deployments - cleaning, security, hospitality, retail - unified by the challenge of operating safely in unpredictable, human-populated environments. Traction has been strongest in cleaning and security, where the operational domain is more bounded and the labor economics are clearest. Hospitality and retail deployments are growing, particularly in high-wage markets where staff turnover is a persistent problem. The form factor that wins in this category is typically one that solves a clear, repeatable operational friction - not one that showcases technical sophistication for its own sake.

A4. Healthcare & Medical Robotics

INTERACT WITH THIS MARKET MAP HERE

Europe's ageing demographics make healthcare robotics less an opportunity and more a necessity. The continent has produced genuine global challengers in surgical robotics, alongside a strong ecosystem in rehabilitation exoskeletons, hospital logistics, and assistive devices. The regulatory environment is demanding - CE marking, clinical validation, long procurement cycles - which acts as both a barrier and a moat. Companies that earn clinical trust early build durable competitive positions. The most accessible near-term opportunity is in hospital operations: disinfection, supply logistics, room turnaround, where ROI is demonstrable and the regulatory burden is lower than in surgical settings.

A5. Agriculture, Forestry & Food Systems

INTERACT WITH THIS MARKET MAP HERE

European agriculture is under compound pressure: labor shortages, climate volatility, sustainability mandates, and thin margins. Robotics is being applied across the full value chain - autonomous weeding and spraying, precision harvesting, crop monitoring, and indoor farming. Europe's diversity of agricultural contexts - large arable operations in France and the Netherlands, specialty horticulture in Spain and Italy, smaller mixed farms across Scandinavia - provides an unusually rich testing ground. The sustainability angle is particularly strong: precision robotics that reduce chemical inputs and cut fuel consumption align directly with EU agricultural policy, creating regulatory tailwind alongside commercial pull.

A6. Construction, Infrastructure & Built Environment

INTERACT WITH THIS MARKET MAP HERE

Construction is among the least automated industries on earth, and Europe's startup ecosystem is beginning to address that seriously. The sector loses enormous value to delays, rework, and labor gaps - all areas where robotics can help. Early traction has come from well-defined sub-tasks: autonomous site monitoring, scaffolding material handling, precision drilling, infrastructure inspection. Europe's large infrastructure pipeline - offshore energy, housing, transport - creates sustained demand. The constraint is deployment complexity: construction sites are among the most unstructured environments any robot navigates, which is why the most traction starts in bounded, repeatable tasks.

A7. Drones & Autonomous Vehicles

INTERACT WITH THIS MARKET MAP HERE

This category spans three distinct but related domains: aerial robotics (drones and UAVs), maritime and underwater systems, and autonomous ground vehicles outside warehouse settings. Europe has particular strength in the first two. Defense and industrial drone capability is well-established, with companies in Germany, Portugal, and the Nordics building serious dual-use platforms. Maritime and underwater robotics is a category where European infrastructure need is acute - North Sea energy assets, rapidly expanding offshore wind, and major port operations all create sustained demand for autonomous inspection and monitoring. Regulatory frameworks for drone operations are also maturing faster in Europe than in many other markets, which is beginning to unlock commercial scale.

A8. Humanoid & General-Purpose Robots

INTERACT WITH THIS MARKET MAP HERE

Humanoids are attracting the most attention in robotics, and warrant the most scrutiny on timelines. The commercial logic is real: a robot that operates in human-designed environments without costly retrofitting has a vast addressable market. Europe's contribution here is grounded in deep research heritage - from Italy's iCub to ETH Zurich's locomotion work - increasingly making its way into commercial products. The near-term opportunity is most likely in structured industrial environments where the form factor advantage is genuine. Full general-purpose capability in unstructured settings remains a longer-horizon problem, and the companies most likely to get there are those with the strongest foundation in manipulation and locomotion research.

A9. Defence, Security & Dual Use

INTERACT WITH THIS MARKET MAP HERE

Defence is one of the verticals where Europe has moved fastest, driven by genuine urgency. Spending has accelerated across the continent, and with it the demand for autonomous surveillance, ISR drones, unmanned ground vehicles, and dual-use platforms that serve both military and civilian markets. The dual-use nature of the underlying technology - perception, edge compute, secure communications - means that defence robotics and civilian robotics share a stack, and founders who understand both markets have a structural advantage. Europe's defence robotics ecosystem is strongest in the Nordics and Baltics, Germany, and the UK, and is attracting increasing institutional capital alongside government procurement.

A10. Mobility & Autonomous Transportation Robotics

INTERACT WITH THIS MARKET MAP HERE

Europe has meaningful technical depth in autonomous vehicle software, and growing deployment in specific mobility contexts: autonomous yard trucks, campus shuttles, port logistics, and sidewalk delivery robots. Full open-road autonomy remains a long-horizon problem, but constrained environments with clear commercial use cases are seeing real traction. EU cross-border regulatory frameworks are beginning to align with deployment ambitions - a meaningful enabler for companies operating across multiple national markets. The most defensible near-term positions are in defined operational domains where the environment is structured enough to make the economics work today.

A11. Science, Research & Laboratory Robotics

INTERACT WITH THIS MARKET MAP HERE

Laboratory automation is a less visible but commercially significant category, with strong European presence. Life science research, drug discovery, and materials science all involve highly repetitive, precision-dependent workflows that are well-suited to robotic automation - and where automation genuinely accelerates outcomes rather than simply reducing cost. The AI-driven lab is an emerging concept: robots that don't just execute predefined protocols but can adapt experimental workflows based on real-time results. Europe's strength in biotech and chemistry research creates natural demand, and the category benefits from shorter sales cycles and clearer ROI than many other application verticals.

A12. Domestic & Consumer Robotics

INTERACT WITH THIS MARKET MAP HERE

Consumer robotics has historically been a difficult commercial category, but the underlying technology is improving faster than ever. Home robots, companion devices, and educational platforms are being revisited with more capable AI, better manipulation, and more realistic price points. Europe's contribution has been strongest in educational robotics and social assistive devices - products designed around genuine human need rather than technical spectacle. The form factor that wins in consumer markets is one that solves a clear daily friction at a price consumers will actually pay, which remains a harder problem than it looks.

Layer B: The Enabling Stack

Every application above depends on a set of underlying capabilities that determine what robots can actually do, how reliably, and at what cost. This is where the foundational innovation happens - and where European technical heritage is deepest. The stack runs from physical hardware at the base to intelligence at the top, with each layer enabling the one above it.

B1. Robotics AI, Models & Agents

INTERACT WITH THIS MARKET MAP HERE

At the top of the enabling stack sits intelligence: foundation models for robotics, spatial AI, perception systems, and the planning and control algorithms that let robots handle real-world complexity. This is where the influence of large language models and vision-language models is being felt most directly in robotics - enabling robots to generalise across tasks rather than being explicitly programmed for each one. It is also where European sovereignty considerations are sharpest. Foundation models trained on industrial data from European factories, with European safety and regulatory standards built in, are a meaningfully different product from those developed elsewhere. The research institutions producing this work - ETH Zurich, TU Munich, INRIA, Oxford - sit at the frontier, and the rate at which that research is converting into commercial products is accelerating.

B2. Simulation, Digital Twins & Dev Tools

INTERACT WITH THIS MARKET MAP HERE

Simulation is increasingly central to how robots are developed and trained. Rather than running every experiment in the real world - expensive, slow, and often dangerous - developers use simulators to test robot behaviour across thousands of scenarios, generate synthetic training data, and validate deployments before they happen. Digital twins of specific environments allow companies to replicate real-world conditions at scale. Europe's gaming and real-time rendering heritage feeds directly into this layer: the same capabilities that built world-class 3D engines are now being repurposed as the development infrastructure for robotics.

B3. Autonomy, Navigation & Localisation

INTERACT WITH THIS MARKET MAP HERE

This layer covers the core capabilities that let robots move intelligently through the world: SLAM (simultaneous localisation and mapping), sensor fusion, motion planning, and fleet autonomy stacks. It's where a significant portion of European robotics IP is concentrated, particularly in vision-based approaches that reduce dependence on expensive infrastructure like GPS or pre-mapped environments. The companies building here often have deep roots in academic research - and the ones that have solved hardware portability and real-time calibration challenges are able to deploy across a much wider range of robot configurations and environments.

B4. Human-Robot Interaction & Interfaces

INTERACT WITH THIS MARKET MAP HERE

As robots move into environments shared with humans, the interfaces between people and machines become increasingly important. This category covers teleoperation systems, XR interfaces (augmented and mixed reality), gesture and voice interaction, and haptic feedback. Europe has particular strength here, with a heritage in display technology, optics, and human-computer interaction research that translates directly into robot interface design. The quality of human-robot interaction is often the determining factor in whether a deployment actually gets used - and used safely.

B5. Robotics Middleware, OS & Infrastructure

INTERACT WITH THIS MARKET MAP HERE

Above the hardware sits the software infrastructure that makes robots operable at scale: robot operating systems, fleet management platforms, cloud and edge robotics infrastructure. This layer matters particularly for European deployments, which frequently span multiple countries, regulatory environments, and languages. Fleet management platforms that coordinate robots across a French warehouse and a German distribution centre, and OS layers that abstract hardware complexity, are the picks and shovels of the robotics buildout - less glamorous than the applications above, but essential to scaling them.

B6. Hardware Components & Subsystems

INTERACT WITH THIS MARKET MAP HERE

At the base of the stack sits hardware: sensors, actuators, end-effectors, power systems, and custom silicon. This is where European optics, photonics, and precision engineering heritage is most directly applicable. LiDAR, vision systems, force-torque sensors, and compliant actuators are all becoming cheaper and more capable - and the companies building robot-native hardware components are often the least visible but most structurally important players in the ecosystem. Cost reduction here is what makes entire application categories commercially viable.

In conclusion, Europe’s robotics startup landscape in 2026 is dynamic and promising, with activity across all layers of the value chain. It offers a fertile hunting ground for investors who share FOV’s vision of the next computing platform being rooted in the physical world, a world where intelligent robots, enabled by spatial computing and AI, augment human productivity in every domain.

Europe has the talent, market demand, and increasingly the capital to build global robotics leaders. With a strong thesis guiding investment (the convergence of spatial computing and robotics) and by leveraging Europe’s collaborative framework, we are confident that the coming years will see many of these startups thrive – delivering returns for investors and strategic value for partners, and quite literally transforming how we live and work through robotics.

Building in any of these areas? We'd love to hear from you - and if you're in London on 25 March, join us at Vantage Point: Robotics Edition, our investor summit and startup showcase for Europe's frontier robotics founders. Register here.

With so much going on in AI each week, you may have missed that NVIDIA's new DreamZero model just crushed robotics benchmarks by doing something novel - teaching robots to predict what will happen before they act.

The model jointly generates both future video and robot actions, essentially letting robots simulate outcomes before moving. This spatial prediction capability has been the missing piece keeping robots confined to controlled environments.

Snapshots from the DreamZero rollout gallery.

That same shift - from map-dependent systems to AI that understands space dynamically - is what just attracted $1.5B to Wayve at an $8.6B valuation. The Cambridge-founded company pioneered what they call "Embodied AI" for autonomous vehicles, using camera-first vision and end-to-end learning instead of expensive sensor stacks and pre-mapped routes. Their cars have been navigating unmapped London streets since 2020. Now, backed by SoftBank, Microsoft, NVIDIA and Uber, they're launching commercial robotaxi trials in 2026.

These are strong signs that the infrastructure layer for physical AI is finally mature.

At FOV, we've been focused on spatial computing technologies for years. Many saw our thesis through the lens of AR/VR consumer applications, but the tech stack we've been building - computational optics, 3D sensing, real-time spatial mapping, depth perception - is now fundamental to physical AI.

Europe has real advantages here, something that makes us very bullish on our thesis over the next decade. Ahead of our Vantage Point event where we're bringing together robotics founders and investor, we wrote about why Europe is positioned to win this wave.

Read more from David Ripert here

FOV NEWS

We're hosting VANTAGE POINT: Robotics Edition on March 25th in London - a curated summit bringing together leading founders, investors, and strategic partners at the frontier of robotics and embodied AI.

First speakers confirmed:

• Troy Estes and Mozghan Kabiri Chimeh, NVIDIA

• Svetlana Grant, Google DeepMind

• William Godfrey, Tangible Finance

• Rory Daniels, techUK

• Robert MacKenzie, Ellipsis

• Kirsty Lloyd-Jukes, Stateful

• Andrei Danescu, Dexory

The half-day program features keynotes, fireside chats, curated startup pitches, and matchmaking sessions designed for signal over noise. We're prioritising thoughtful dialogue, high-quality deal flow, and long-term relationships.

This early registration phase is limited to VCs, LPs, angels, and partners only. Attendance is by approval to maintain a highly curated guest list.

Powered by Venture Café London
For partnership enquiries: sointu@fov.ventures

ICYMI…

We published 10 Early-Stage European Robotics Startups to Watch for 2026 = mostly pre-seed, all deeply technical, building full physical-intelligence stacks from perception to deployment. These teams are moving beyond demos toward systems designed to survive factories, logistics sites, and contested real-world environments.

https://viewpoints.fov.ventures/p/10-early-stage-startups-to-watch-in-2026

FOV Portfolio News

• Edailabs (Emma) achieved the 2nd highest retention rate globally among language learning apps, ahead of apps that have raised tens of millions. The metric validates that their teaching approach is engaging and keeps users coming back consistently. (more here)

• Distance Technologies was featured in Tech.eu discussing how they're building the always-on visual intelligence layer between humans and machines. CEO Urho Konttori describes it as "the operating system for human-AI interaction, enabling intuitive decision-making and making autonomous AI systems safer." (more here)

• Scenario launched Nodes, a new visual workflow engine that lets users build and run processes on one canvas. The node-based system connects models, tools, and assets into pipelines that execute from start to finish. Users can also build custom workflows and publish them as simple apps. Everything runs entirely in the browser - no installs required. (more here)

• GEEIQ was featured in The Business of Fashion for their latest State of Brands in Virtual Worlds data. The analysis shows fashion brands are recalibrating investment in gaming platforms like Roblox and Fortnite. While overall spend pulled back in 2025, integrations rose and brands are focusing on formats that drive repeatable performance. (more here)

• Graswald released AI 2.0, rebuilding their platform to become a core operating system for fashion imagery production.

FOV Job Board

• Levtek - Mechanical Engineer (Malmö, Sweden).

• GEEIQ has multiple positions open in London, UK: Senior QA Engineer, Sales Development Representative, and Senior DevOps Engineer.

• ai|coustics has multiple openings in Berlin, Germany including ML Research Engineer, Backend Developer, Audio Signal Processing Engineer, and Developer Relations roles.

• Atopia Space - SDR / Outbound Sales Freelancer (Remote or Munich, Germany).

• Iconic AI has multiple AI Research positions in London, UK across Expressive Embodiment, Reinforcement Learning, and Model Optimization.

At FOV Ventures, we back the founders building the next era of computing. Europe has a once-in-a-generation opportunity to lead this transformation. We’re especially excited right now about frontier tech that touches the real world, where Europe has structural advantages Silicon Valley cannot easily replicate.

This means anything where atoms merge with bits — putting digital intelligence into physical environments, through robotics, mixed reality, computer vision, and autonomous systems. Europe is uniquely positioned to build in this arena.

Europe's industrial DNA now aligns with geopolitical necessity for the first time. Manufacturing heritage meets AI maturity meets sovereign urgency.

Here's why this moment matters.

Sections

• Europe’s Research-to-Real-World Advantage

• Real-World Applications, Real-World Urgency:

  • Industrial settings

  • Agriculture

  • Construction

  • Healthcare

  • Precision and Safety Applications

• Spatial Computing: Europe's Hidden Superpower

• The Case for European Sovereignty

On 25 March 2026, FOV Ventures is bringing together Europe's leading robotics founders, investors, and operators in London for Vantage Point: Robotics Edition — a half-day showcase and summit for the teams building embodied intelligence into the real world. Everything below reflects how we think about where the opportunity is. If you're building in this space, we'd love to see you there.

VANTAGE POINT: LONDON - The Robotics Summit

Robotics is entering its inflection point — join the platform bringing together top founders, investors, partners and labs shaping embodied intelligence in Europe.

vantage.fov.ventures

Europe’s Research-to-Real-World Advantage

In the global race for frontier-tech dominance, Europe has found its unfair advantage: a $400 billion startup funnel rooted in its historic universities.

While traditional VC funding has cooled globally, European academic spin-offs are defying the trend - reaching a near-record $9 billion in funding last year and producing 76 ‘centaurs’ and unicorns in 2025 alone.

This is the maturation of a cross-border ecosystem where world-class IP is finally meeting world-class commercial execution.

As a result, research institutions are producing commercial companies at an accelerating pace. Switzerland's ETH Zurich alone produced 46 new spin-off companies in 2025. Germany's TUM set a record with 103 tech start-ups in 2024, and its affiliated incubator UnternehmerTUM was named Europe's top startup hub for the second year running.

These startups are attracting serious capital and corporate buyers too.

ANYbotics AG raised approximately $60M to commercialize its quadruped robots. Sevensense Robotics, an ETH vision-guided robotics spin-off, was acquired by ABB in 2024. While Oxbridge’s research gave birth to WaveOptics (acq by Snap for $500M) and Latent Logic (acq by Waymo in 2019).

University spin-offs now command roughly one-fifth of Europe's deep-tech VC dollars, thanks to strong technical founding teams and protected IP.

This expertise clusters geographically in ways that create compounding advantages. France boasts world-class AI labs (INRIA, Paris-Saclay) spawning leaders like Prophesee and Mistral AI. Switzerland's ETH Zurich produces a steady stream of robotics and computer vision spin-offs, including our portfolio company nunu ai, and has R&D offices from Google, Apple, and NVIDIA.

We've seen it up close in our own portfolio. Flow Computing emerged from VTT Technical Research Center of Finland, founded by researchers alongside a commercial veteran. Flow is building a revolutionary chip architecture to accelerate parallel computing, essential for powering real-time AI in physical systems. 

The ecosystem is also maturing beyond first-time university spin-offs. Experienced founders are launching new ventures with deeper technical expertise and stronger industry networks. Urho Konttori founded Distance Technologies after leading Varjo's world-class VR development. Harri Valpola launched System 2 AI after selling Curious AI and spending three years at Apple. This pattern creates companies with better technical foundations and clearer commercial strategies.

Now, European startups are building across the deep tech stack for virtually every sector, solving real-world problems in factories, warehouses, hospitals, farms, and households.

Real-World Applications, Real-World Urgency:

The true test of a technological revolution isn't how it performs in a clean room, but how it handles a construction site, a hospital ward, or a battlefield. Europe’s competitive edge in robotics stems from its status as the world’s 'living laboratory' - a diverse landscape of asset-heavy industries hungry for automation. By grounding AI in the messy reality of the physical world, European startups are moving beyond the novelty of humanoid research to solve the urgent demographic and labor crises of the next decade.

Industrial settings

Startups develop robots and automation aiming to make manufacturing more flexible or to assist human workers on the factory floor. Europe's strong manufacturing base and research in industrial automation have spawned many ventures in collaborative robots (cobots) and robotic cells for SMEs.

Take our portfolio company Makiina, pioneering modular, low-cost, AI-driven robotic systems built from scratch in Finland. Designed to learn tasks through human demonstration, they make adaptable automation viable far beyond factory floors.

Makiina

Agriculture

Agriculture faces similar pressure from farm labor shortages and demands for sustainable practices. European startups build robots for planting, crop care, harvesting, and farm management. Europe's mix of large industrial farms and smaller specialty farms provides a testing ground for field robotics such as weeding machines, harvesters for fruits and vegetables, and drones or rovers for monitoring crops.

Many agri-robotics startups emphasize environmental benefits like precision weeding to reduce pesticide use. Our portfolio company Spogen is building a voice and vision-guided co-pilot that helps operators interact with complex machinery like tractors using natural language. Think Alexa running on diesel with steel tracks.

Construction

Construction has been relatively late to automation, but European startups are emerging to tackle heavy labor tasks and productivity issues. These companies build robots for construction sites, including robotic systems for on-site material handling, automated construction processes, and surveying and monitoring. Europe's construction robotics scene is nascent but growing, often leveraging partnerships with construction firms.

Healthcare

Healthcare faces demographic realities including aging populations and staff shortages that make robotic solutions urgent. Europe is home to several leading medical robotics startups, particularly in surgical robotics and rehabilitation.

The continent has produced notable challengers to Intuitive Surgical's dominant da Vinci system, including Switzerland's LEM Surgical and UK-based CMR Surgical with their Versius system. Focus areas include minimally invasive surgical assistants, robotic endoscopy, prosthetics and exoskeletons, and micro-robots for targeted therapies.

Precision and Safety Applications

Defence represents another area where Europe's deep tech edge matters for both national security and dual-use innovation that spills into civilian industry. Portfolio company Distance Technologies shows how this works. Founded by two Nokia veterans who previously built Varjo into a leader in enterprise and defence-focused VR headsets, they recognised the hardware barrier preventing mass adoption. They're now combining hardware, software, and AI to create light fields and images that appear naturally within your field of view at correct depth and distance in the real world. Their technology serves both automotive and defence markets, appearing on car windshields or in fighter jets. Europe needs more dual-use startups that thread the needle between national resilience and commercial scale.

Distance Technologies

Beyond these vertical applications, heavy investment flows toward general-purpose humanoid robots with human-like form factors (bipedal legs, arms, torso, head). Europe has a legacy of humanoid robotics research, including Italy's iCub open-source humanoid and NASA and ESA collaborations. A few startups are now pushing toward commercial humanoids, including 1X originally from Norway.

These applications across robotics share a fundamental requirement. Machines need to understand space, depth, and physical context. That brings us to Europe's second structural advantage in spatial computing.

Spatial Computing: Europe's Hidden Superpower

We are moving past the era of looking at data on screens to an era where humans and machines inhabit the same space.

The "engine room" of this revolution is firmly rooted in Europe. The continent has spent decades perfecting the core disciplines - optics, photonics, and real-time rendering - long before "spatial computing" was a buzzword.

This technical foundation wasn't built for headsets, but it trickled down from Helsinki’s gaming studios, Swiss optics labs, and the telecommunications labs of Nokia and Ericsson. Today, that DNA powers the entire spatial stack. Real-time 3D games engines like Unity and Unreal, have been repurposed as the operating systems for reality. They now run the simulations for our military, the training grounds for our robots, and the "digital twins" of our autonomous vehicles.

These engines have transitioned from entertainment tools to the essential infrastructure for understanding physical space.

As these simulated environments become more complex, the industry bottleneck has shifted from simply building 3D worlds to capturing the existing world at scale. European innovators are breaking this barrier through breakthroughs like Gaussian splatting and image diffusion, with companies like M-XR and Graswald making high-fidelity 3D content generation near-instant and hyper-realistic.

However, capturing reality only matters if the hardware is practical enough for everyday use.

This is where European startups are solving the "physics problems" of miniaturisation, with Kubos Semiconductor reinventing microLEDs for efficiency and IXI perfecting electronically tuneable lenses that mimic the human eye’s natural focus.

The strategic value of this talent has triggered a global acquisition wave; in the lead-up to the 2024 Vision Pro launch, nearly half of Apple’s targeted acquisitions were European-based, while Snap’s $500M purchase of WaveOptics underscored the region's dominance in waveguide technology.

We are now witnessing a sophisticated "third wave" of innovation as veterans from Nokia and Varjo spin out new ventures like defence tech, Distance Technologies, and $11B valued Oura, with Europe’s heritage in optics and computer vision the indispensable substrate for the next interface layer of computing.

The Case for European Sovereignty

This concentration of high-value expertise is shifting the conversation around European technological sovereignty from a philosophical debate into a practical strategy for owning a digital destiny.

Reducing the long-standing reliance on foreign cloud providers and hardware manufacturers allows Europe to build homegrown capabilities that reflect its own industrial standards. This transition ensures that the critical technologies shaping the future stay rooted in local values while maintaining control over essential data and infrastructure. There is a massive opportunity to lead in the field of physical AI by defining the core principles of the industry alongside the products themselves.

Take drone motors and actuators. Chinese firms control 70-90% of the global commercial drone market, and European defence forces are acutely aware of the dependency (up to 70% of components in Ukrainian battlefield drones come from Chinese suppliers). European-made alternatives exist but typically cost 10x more, making reshoring economically unviable. Our Portfolio company Makiina is building drone motors and actuators entirely in-house in Finland, producing dual-arm robotic systems for under $1,200 with motors that early benchmarking suggests match Chinese alternatives on both torque and cost. If those economics hold at scale, more than just sovereign supply chains - it's a platform for truly European robots, built with European parts, designed for European deployment.

The current policy landscape offers a unique advantage for this transition. European AI regulation provides a stable framework that actually favours industrial innovation. Because physical AI relies on sensor data from factories and warehouses rather than sensitive consumer information, it naturally avoids many common privacy hurdles. This allows developers to focus on safety and interoperability without the friction typically associated with personal data. A company can innovate within a logistics hub or a manufacturing plant while staying perfectly aligned with existing standards.

Public funding and institutional support are now accelerating these efforts through high-impact initiatives. New cross-border testbeds for autonomous drones and digital hubs for robotics are helping to de-risk early deployments for startups. These efforts are further supported by the introduction of EU Inc, a transformative legal framework that allows startups to scale across the entire bloc under a single company structure. This system removes the administrative complexity that historically fragmented the European market and slowed down the growth of deep tech firms.

The practical impact of this structural shift is significant for any company dealing with physical hardware. A robotics startup can now test machines in German factories and deploy software in French warehouses while hiring research talent in Poland through one legal entity. This streamlined approach simplifies investment structures and speeds up cross-border growth by harmonising operations across the single market. These changes are creating a clear path for a new generation of sovereign tech giants to rise from the lab and compete on the global stage.

We believe this is the most exciting time to back European deep tech founders in decades, because the convergence is here (areas once considered futuristic like AI, robotics and spatial computing are reaching maturity together), because the urgency is real, and because the infrastructure is catching up. Europe has a clear edge through engineering depth, industrial heritage and cross-border collaboration, and this is finally aligned with what the world needs most to reap real-world benefits from AI: intelligence in the physical world.

At FOV Ventures, that’s the bet we’re making. Not just on technology, but on people. Founders who want to build hard things, with long-term impact, in the places they know best.

If you’re one of them, let’s talk.

If you're building in any of these areas, we'd love to see you in London on 25 March. Vantage Point: Robotics Edition is our investor summit and startup showcase for Europe's frontier robotics founders.

VANTAGE POINT: LONDON - The Robotics Summit

Robotics is entering its inflection point — join the platform bringing together top founders, investors, partners and labs shaping embodied intelligence in Europe.

vantage.fov.ventures

Welcome to Viewpoints, news and investment activity covering the next era of computing - brought to you by FOV Ventures.

This Month:

• Six Shifts Shaping 2026: Our investment themes

• 10 European robotics startups to watch at the pre-seed stage

• Vantage Point: Robotics Edition coming to London in March

Dave Haynes, Petri Rajahalme, Sointu Karjalainen & David Ripert (FOV Ventures)

At FOV, we've been persistently focused on a set of truly transformative technologies that seemed like science fiction just a few years ago - AI, robotics, and spatial computing. Now they're simultaneously reaching maturity and creating massive new opportunities.

Physical AI and robotics have shifted from theoretical possibilities to practical realities. We've moved from the "wow" phase of seeing a robot walk to the "how" phase of making it industrial-grade. The frontier tech of yesterday is becoming the infrastructure of today.

We just published our investment focus for 2026 across these key themes. While some areas race ahead with major leaps arriving weekly, others are recalibrating and moving away from the hype machine toward real deployments.

https://viewpoints.fov.ventures/p/six-shifts-shaping-2026

Robotics and physical AI is one of the natural evolutions of our long-held thesis - the shrinking gap between the digital and physical realms, and the application of AI beyond our 2D screens (and chat windows).

As we outlined in the piece above, robotics in 2025 saw VC investment rebound sharply - but the majority of capital went into the same three ideas: robot brains, robot training, and humanoids.

Now, as foundation models commoditise, alpha has moved to the fringe — incl. vertical-focused solutions solving specific problems exceptionally well.

Europe's robotics renaissance in these areas is being built in small teams, early labs, and first production pilots across the continent. While most attention flows to later-stage humanoids, the next breakout companies are forming much earlier - at pre-seed, where new architectures and software stacks are still fluid.

We're deep in this space, so we wanted to spotlight what we're seeing on the ground. We published 10 Early-Stage European Robotics Startups to Watch for 2026 = mostly pre-seed, all deeply technical, building full physical-intelligence stacks from perception to deployment. These teams are moving beyond demos toward systems designed to survive factories, logistics sites, and contested real-world environments.

https://viewpoints.fov.ventures/p/10-early-stage-startups-to-watch-in-2026

FOV NEWS

VANTAGE POINT: Robotics Edition 2026

We're hosting VANTAGE POINT: Robotics Edition on March 25th in London - a curated summit bringing together leading founders, investors, and strategic partners at the frontier of robotics and embodied AI.

The half-day program features keynotes, fireside chats, curated startup pitches, and matchmaking sessions designed for signal over noise. We're prioritising thoughtful dialogue, high-quality deal flow, and long-term relationships.

This early registration phase is limited to VCs, LPs, angels, and partners only. Attendance is by approval to maintain a highly curated guest list.

Powered by Venture Café London.

For partnership enquiries: sointu@fov.ventures

FOV PORTFOLIO NEWS

• Doublepoint announced two major products at CES 2026: a $249 developer kit with partners STMicroelectronics, ams OSRAM, and Ambiq (available for pre-order now), and WowPlay, a consumer-ready WearOS app for everyday music and media control. (more)

• M-XR introduced Marso AI, a game-ready asset platform that generates production-grade PBR textures from any 3D model, powered by their foundational model trained with PBR data from Marso Measure - built with technical artists to deliver a full game-ready asset workflow. (more)

• Graswald continued their strong 2025 momentum, becoming a critical workflow tool in fashion as they work toward a true operating system for fashion that supports brands from early creative ideation to final publish-ready assets. (more)

• GEEIQ launches The State of Brands in Virtual Worlds 2026, a report mapping how brands actually operate across gaming and virtual platforms based on observed behavior, not opinion—register to receive it first when it launches next week. (more)

• Distance Technologies unveiled the Field Operator HUD—a lightweight, glasses-free XR system that integrates with existing soldier helmets to give defense vehicle operators the same tactical advantage as advanced fighter-pilot helmets. (more)

• nunu.ai launched PR Testing Agents, enabling fully autonomous end-to-end tests directly inside GitHub Pull Requests that interact with products like real users—reaching 85% issue detection compared to 30% with traditional PR bots, solving the QA bottleneck as AI coding tools accelerate development. (more)

• IXI signed a letter of intent with Atol and Vizio, two leading optical groups representing 3,000+ stores in France, to explore commercial pilots for their autofocus eyewear and jointly develop the optical retail experience around how opticians can best introduce autofocus technology. (more)

FOV JOB BOARD

Open Roles at Portfolio Companies:

• AI Coustics is hiring across their Berlin team with 10 roles spanning Systems Software Engineer (Rust), Developer Relations, Audio ML Data Engineer, Backend Developer, Digital Audio Signal Processing Engineer, Audio Data Infrastructure Engineer, Forward Deployed Engineer, ML PhD Internship, Junior IT Operations Engineer, and ML Research Engineer. (apply)

• Spogen.ai has a QA & Data Specialist position (Entry Level, €2,500-4k/month) open across Oregon, Atlanta, and Bucaramanga, Colombia. (apply)

• Pixieray is seeking a Digital Channels & Growth Manager (Mid-Senior Level) across Espoo (Finland), Washington DC, and Lehi (Utah). (apply)

• GEEIQ is hiring an Account Executive (Mid-Senior Level) in London, UK. (apply)

• Levtek has a Senior Mechanical Engineer position open in Malmö, Sweden. (apply)

• Atopia Space is seeking an SDR / Outbound Sales Freelancer (Entry Level, Remote or Munich). (apply)

We’ve said it before and we’ll say it again… but with fresh eyes for 2026.

As Amara's Law reminds us, human nature tends to overestimate the impact of technology in the short term but underestimate it in the long term. The constant buzz of fleeting tech trends will come and go. But at FOV Ventures we’ve been persistently focused on a set of truly transformative technologies that have been unfolding gradually and are now converging to take us into a new era of computing.

Areas like AI, robotics, and spatial computing - that to many seemed like science fiction just a few years ago - are simultaneously reaching maturity and creating huge new opportunities. This is something we went even deeper on last year, publishing our first physical book on the subject (grab your free digital copy here)

While some of our focus areas have been steady companions for years, evolving in expected ways, others have accelerated dramatically. Physical AI and robotics, for example, have shifted from theoretical possibilities to practical realities. And other building blocks - from AI and computer vision to mixed reality and edge computing - are finally clicking into place.

The rapidly shrinking gap between digital and physical realms has long been our investment thesis and in 2026, the signals are even stronger.

We've moved from the "wow" phase of seeing a robot walk, to the "how" phase of making it an industrial-grade reality. The "frontier" tech of yesterday is becoming the infrastructure of today, and whilst some areas race ahead with major leaps seemingly arriving weekly rather than yearly - other areas are recalibrating and moving further away from the headline hunting hype machine of today’s tech industry.

So let’s revisit some of the key themes from last year and look at how they’ve evolved, the nuance we’re seeing on the ground, and how we might be allocating our capital this year.

If you are working on a startup, or investing yourself across any of these themes we would love to hear from you! Reach out on LinkedIn and submit details here 

1. Smart Glasses: the most human interface for AI

Last year we predicted: Smart glasses would emerge as "the most human interface for AI."

What happened: We were right about the trajectory, but the winning insight was simpler than expected - smart glasses aren't just "phones for your face." They're "eyes for your AI." The value today is the AI's ability to see your life and provide proactive utility.

We made new investments in this category in 2025 and will continue proactively looking for opportunities. It remains early in the cycle, so we're interested in B2B opportunities while asking when and how main players will open to 3rd party developers - and what enablers still need building.

As Meta's CTO Andrew Bosworth noted, the form-factor's AI capabilities proved key in driving consumer adoption of this new wearable category. Ray-Bans validated the core assumption - people will buy tech-enabled sunglasses - de-risking the path to Meta's north star: Project Orion (full AR glasses).

Orion AI glasses: The future of AR glasses technology | Meta

A decade in the making, recent groundbreaking innovations from Meta with Orion redefine what is possible for AR, AI and beyond.

www.meta.com/en-gb/emerging-tech/orion/?srsltid=AfmBOoqS2_QvQ44zOZ4wzGVuzhq3gj7I421_cB79QnHNNGxoSc4yQmXL

They took a key step in 2025 with Meta Rayban Displays, which adds wrist-powered control and a subtle heads-up layer to everyday vision - surfacing photos, messages, and AI prompts without breaking eye contact with the real world.

Google - having announced AndroidXR and partnerships with Xreal, Warby Parker, Gentle Monster, and Samsung, plus impressing with Gemini's multimodal and memory capabilities - will be a key player in 2026. While CES witnessed a "war of hundreds of smart glasses," the other key players we're watching are Snap (next-gen Spectacles) and Apple (most likely 2027).

But as the category matures, what can startups build that Meta, Samsung, and Apple can't…

Innovation is still needed at the infrastructure level from startups that walk the fine line between true technical innovation, large customer demands, and commercial potential (e.g., WaveOptics sold to Snap for $500M).

A growing list of startups have clearly defined wedges - something in either hardware or OS that makes them competitive as AR/AI wearable emerges for mass consumers. Metra is building an open-source operating system for glasses. Raven focuses on eye-tracking HUD.

At the application level, the opportunity remains nascent and undefined, especially in consumer. We're watching closely while looking at opportunities in environments where users are paid to wear glasses and specialised AI assistants provide workers with productivity, safety, or decision-making boosts.

Our own investment, Spogen, leverages smart glasses + augmented intelligence for operators of heavy machinery. Startups like Syn2Core develop AI co-pilots for frontline automotive workers to minimise human errors. Many opportunities will emerge.

2. Voice AI: more interactions beyond the chat window

Last year we predicted: Voice would help AI break free from the 2D chat window, with LLMs entering the physical world through ambient interfaces.

What happened: We were right about voice emerging as a key interaction layer for both on-device and wearable AI, confirming the chat window isn't the be-all and end-all. For investors, it became a huge category in its own right - startups tackling everything from new wearables and vertical applications to the more critical infrastructure and tooling (like our portfolio company AIcoustics).

But reliability and context-awareness limited how autonomous these systems could be in practice. Vertical-specific solutions consistently outperformed general-purpose assistants. Consumers showed limited tolerance for poorly contextualised or overly proactive AI behaviour.

The "life-log" hype of 2025 faded. The truth many realised is that most of life isn't worth recording, so many always-on wearables will struggle with PMF, likely replaced by vertical-specific solutions.

The next interface likely isn’t "AI recording everything" but ambient AI operating in constrained domains.

3. Physical AI: from buzzword to infrastructure

Last year we predicted: "Physical AI" would become the new buzzword as world simulation became AI's next grand challenge.

What happened: We were right about the buzzword - Google announced a new team to work on AI models that simulate the physical world, NVIDIA unveiled Cosmos trained on 20 million hours of video, and the funding frenzy followed. General Intuition secured one of the largest seed rounds ever at $133.7M in November, Field AI hit a $2B valuation, renowned AI scientist Yann LeCun launched Advanced Machine Intelligence (AMI) in December, reportedly seeking over $5 billion and just last week, Skild AI raised a $1.4 billion Series C led by SoftBank.

But in 2026, the first-mover advantage for general-purpose "robot brains" evaporated. With foundation models from NVIDIA, Physical Intelligence, and OpenAI becoming ubiquitous, a robot's ability to "see and move" might just be a commodity you buy via an API.

Twitter tweet

Because the "brain" is now a commodity, alpha has moved to whoever can deploy, maintain, and scale hardware most efficiently.

The companies best positioned might not just be the ones raising $1B+ for world models - they'll include industrial incumbents with existing manufacturing footprints. For example, FANUC does $6B annually in robotic arms and ABB's robotics division did $2.4B in 2024, both these companies already have the supply chains, customers, and distribution. When they can license a “better” model from Nvidia or Google, their existing advantages compound.

4. Robotics: now riding the next wave

Last year we predicted: A new wave of robotics startups would lead to a surge in investor interest.

What happened: We were right - VC investment rebounded sharply, with notable rounds including Figure AI ($1B+ Series C), Apptronik ($403M Series A), and OpenAI-backed 1X entered talks in September 2025 to raise up to $1 billion dollars.

However, in 2025, the majority of capital deployed went into the same three ideas: robot brain, robot training, and humanoids. But as the foundational models discussed above commoditise, alpha moved to the fringe.

At FOV, we made our first forays into robotics with two, so-far unannounced, rounds.

We see that many of the most successful deployments are still vertical-focused. Soft-robotic grippers for agriculture; Quadruped inspectors for oil rigs like Anybotics; Modular appendage-only AI for surgical bays. The robots currently making money solve highly specific tasks in environments where reliability beats generality enough times.

Twitter tweet

We're also seeing more deployments than ever before, particularly in China where manufacturers are moving aggressively into industrial robots, humanoids, and autonomous systems at scale. This has triggered increased Western focus on manufacturing sovereignty and supply chain resilience, driving more capital into domestic robotics - particularly for defence applications.

5. VR & Mixed Reality: a year to recalibrate

Last year we predicted: VR's long march to mainstream adoption would continue through mixed reality, shaped by platform competition between Apple and Meta.

What happened: The march continued, but several setbacks signal a wintry year for VR - best categorised as a year of transition rather than category death. The mainstream narrative has cooled.

We didn't make any new VR investments in 2025 and will look at new opportunities very selectively.

As predicted, Vision Pro only got minor upgrades, with a next-gen (or cheaper) model unlikely until 2027. Google re-entered with AndroidXR powering a new Samsung headset, but it wasn't the game-changer hoped for. Meta just announced layoffs in Reality Labs content teams - a necessary reset after overpriming the pump.

The press narrative that "Meta is abandoning VR for AI" is overstated. Content spend is slowing, not usage. Anecdotally, VR engagement rose again over the holiday period and we predict Meta will expand the Quest ecosystem with a lighter headset later this year for broader lean-back audiences. Valve's 'Steam Frame' headset set for 2026 release should also do for VR what Steam Deck did for handhelds, solidifying the core gaming audience.

But there's two massive undercurrents:

First, top-decile indie studios are printing money. Several, off the back of Gorilla Tag’s success, are nearing $100M in revenue and 10M players with their playbook of low-budget viral game mechanics that attract social media native audiences. Any new startups in this space need to prove they can develop products and iterate quickly with lean teams, leveraging social media to build community and user acquisition without over-relying on platform support.

Second, VR teleoperation is becoming a real business. Across hospitals, factories, and logistics hubs, humans increasingly "possess" robots via headsets to resolve edge cases. One operator supervises dozens of machines. Surgeons work remotely. Hazardous maintenance avoids human entry. Autonomous vehicles hand off complex maneuvers to remote pilots.

Platforms like Formant (Series B) and Viam (Series C) provide the fleet backbone. While startups like Extend and Sunday Robotics push low-latency control and learning at scale. This is becoming a high-ROI way to extend human skill across distance, risk, and labor constraints - while gathering valuable data for generalized AI models that take us closer to full automation.

6. Gaming Tech: underpinning the simulation stack

Last year we predicted: Gaming tech would expand beyond entertainment into core infrastructure.

What happened: We were right, but the shift was even more fundamental than expected. In 2025, we saw a "Simulation-First" approach solidify across every major sector. Anduril leverages game engines for high-stakes defence simulations (their first acquisition was a gaming company). Tesla creates virtual worlds to stress-test autonomous systems before they hit the road. Beyond heavy industry, BMW normalised AR in heads-up displays, and Matterport turned 3D walkthroughs from a luxury into a real estate standard.

The bottleneck shifted. It's no longer the engine itself - but the accessibility of content creation and data capture.

AI-powered creation tools like FOV portfolio companies Scenario and Move.ai are already being adopted by game studios as first movers to push creative boundaries. Meanwhile, 3D capture pipelines improved rapidly through techniques like gaussian splatting and image diffusion models, making it exponentially easier to scan assets and generate environments. FOV portfolio company M-XR exemplifies this shift, while companies like Graswald and Aino apply these techniques to unlock e-commerce and other verticals.

The final piece of gaming's move into industry is AI NPCs using autonomous agents and LLMs, creating sophisticated applications from soft skills training to healthcare to collaborative learning.

If you're building in any of these areas, we'd love to hear from you. While this list captures our current thinking, we're always excited to be proven wrong by founders seeing around corners we haven't yet turned. And no doubt we’ll continue to revisit this piece through the year.

Get in touch at www.fov.ventures 

Europe’s robotics renaissance is in full swing, being built in small teams, early labs, and first production pilots across the continent. While most attention still flows to later-stage humanoids, foundation models, and defence primes, the next breakout robotics companies are forming much earlier, at pre-seed, where new architectures, software stacks, and manufacturing approaches are still fluid.

This list highlights 10 early-stage European robotics startups to watch for 2026 - mostly pre-seed, all deeply technical, and all ultimately enabling us to tackle hard problems in the real world. These teams are building robots, full physical-intelligence stacks spanning perception, control, simulation, learning, and deployment, tightly coupled to real operational use cases.

Across mobility, manipulation, autonomy, and embodied AI, they are raising capital to move beyond pitch decks and demos, toward systems designed to survive factories, facilities, logistics sites, and even contested real-world environments. 

Some are still emerging from stealth, others already in pilots, but each represents a credible attempt to make robots genuinely useful at scale - and to redefine what European robotics can become over the next decade.

Note: this is far from a definitive list, and for this post we’ve focused only on some of the really early stage companies that crossed our path in 2025. We’ll be doing a follow-up post soon with those startups that have raised larger seed or Series A rounds. If you would like to be included, or nominate a company, reach out here.

And In March 2026, we're bringing this ecosystem together at Vantage Point: Robotics Edition in London - a curated investor summit + demo day that will connect Europe's most promising robotics founders with investors and other industry leaders.

Investors and startups request your spot below…

Vantage Point: Robotics 2026

Robotics is entering its inflection point — join the platform bringing together top founders, investors, partners and labs shaping embodied intelligence in Europe.

vantage.fov.ventures

1. MAKIINA — Finland

Building the hardware and software stack for physical intelligence: modular, low-cost, scalable, and made in Finland. From drone motors and actuators to power electronics, sensor units, and a breakthrough AI robotics foundation model, each piece is built like LEGO blocks that fit together into different form factors of physical AI agents, affordably and at scale.

MAKIINA developed methods to produce dual robotic arm systems for under $1,200 all in-house, while iterating on hardware at software speed. Their robots learn tasks through human demonstration rather than rigid programming, handling "fuzzy" processes that traditional robots can't.

Stage: Seed | Founded: 2024 | Amount Raised: Not disclosed

2. Levtek — Sweden

Democratising automation through "software-defined" ultralight electric vehicles and collaborative robots. Their flagship product, the Levkart, is an autonomous electric utility vehicle designed to work alongside humans in shared workspaces like warehouses and industrial sites.

Levtek focuses on cognitive robotics - building "ride-on" robots that can reason, learn, and adapt to their environments. They've partnered with PostNord (Swedish national postal service) for logistics trials and collaborate with Malmö and Lund Universities on machine learning and human-robot interaction research.

Stage: Seed | Founded: 2023 | Amount Raised: €2.25m

3. Mimic — Switzerland

Deploying frontier physical AI across industries with AI-driven dexterous robotic hands paired with proven, off-the-shelf robot arms. "Humanoids are exciting, but there aren't many industrial scenarios where the full-body form factor truly adds value," says Co-founder Stephan-Daniel Gravert. "Our approach delivers the same capabilities in a way that is much simpler, more reliable and rapidly deployable."

Founded in 2024 as an ETH Zurich spin-off by Stefan Weirich, Stephan-Daniel Gravert, Elvis Nava, Benedek Forrai, and Robert Katzschmann, Mimic brings together a multidisciplinary team of 25 engineers developing an AI foundation model and humanoid robotic hands that make human-level dexterity deployable across industries.

Stage: Seed | Founded: 2024 | Amount Raised: €17m(€13.8m in latest round led by Elaia, alongside Speedinvest)

4. Paddington Robotics — UK

Founded in 2024 by Zehan Wang, who previously founded Magic Pony before exiting to Twitter, Paddington Robotics are building robots that help people do more. Starting in supermarkets and grocery stores, Paddington's robots are designed to lighten the load for staff working long, labor-intensive hours - boosting productivity and enabling them to do more with less effort.

Twitter tweet

"It may not be as flashy as kung-fu fighting or backflipping humanoids, but we are building for real-world impact, not just building tech for technology's sake." - Zeha Wang.

Stage: Pre-Seed | Founded: 2024

5. Loki Robotics — Switzerland

Building the autonomous backbone of facility operations. Loki is engineered to master one of the most complex physical tasks in robotics: cleaning like a person. The first robot designed to handle high-contact, high-variation environments - swapping tools, applying cleaning agents with precision, and physically interacting with fixtures and objects.

Twitter tweet

From deep overnight restroom cleaning to continuous day porter routines, Loki adapts to facility demands. "We spend over a year of our lives doing chores - time that could be spent raising kids, calling grandparents, making music, building things," says the team. At the core, they combine end-to-end learning with teleoperation to drastically shorten deployment time and unlock real-world capability fast.

Stage: Pre-Seed | Founded: 2024

6. Qualia — Denmark

Building the infrastructure layer for robotics AI. Qualia provides a platform for training and deploying VLA (Vision-Language-Action) models - foundation models that allow robots to reason about their surroundings using visual and text input, then execute physical tasks.

Twitter tweet

Their platform simplifies the complex workflow of training embodied AI with tools for data processing, model fine-tuning, and rapid iteration. Supporting leading VLA models like Pi 0.5, SmolVLA, and Gr00t 1.5, Qualia targets robotics engineers who need to iterate quickly, hardware companies adding AI reasoning via API, and foundation model labs distributing to a wider ecosystem.

Stage: Pre-Seed | Founded: 2025

7. Saturn Dynamics — UK

Saturn is developing the first world model as a virtual simulator, powered by a physics-aware generative engine. Prompt Saturn to generate and control simulations, reducing friction with your tech stack and lowering the time-to-market. Saturn takes in multimodal sensor data, state data, and text, to produce video annotated with world-state data.

Saturn aims to become the foundational simulator behind the autonomous systems that will shape our world in the coming years - powering every step from early-stage training to real-time decision-making when deployed in the real world.

Stage: Pre-Seed | Founded: 2025

8. Voyage Robotics — London

Building the world's first neural safety system for Physical AI-powered robots and vehicles, alongside humanlike AI controls for robots to automate any task. Committed to open-source robotics and partnering with other robotics and embodied AI builders to develop common, modular standards and platforms. The team is comfortable working with both software and hardware, aiming to build their own robot workers while helping others build theirs using Voyage's technology.

Stage: Pre-Seed | Founded: 2024

9. Kova Labs — Finland

Developing embodied intelligence for the next generation of unmanned systems in defence and national security.

Kova Labs builds the "brain" of autonomous hardware, giving machines a real-time understanding of the physical world to perceive, navigate, and act autonomously in complex and contested environments (where GPS might be jammed or communication restricted). Their technology processes complex spatial data like point clouds for 3D mapping, helping robots understand difficult terrains including forest environments. Backed by Lifeline Ventures and angels, they're modernising how unmanned systems operate in high-stakes defence applications.

Stage: Pre-Seed | Founded: 2025

10. Sunrise Robotics — Slovenia

Sunrise Robotics revolutionises manufacturing with intelligent, autonomous robotic cells that enhance flexibility, scalability, and efficiency.

Stage: Seed | Founded: 2023 | Amount Raised: $8.5M

From Helsinki to Zurich, London to Copenhagen, technical founders are building the physical intelligence stack that will define the next decade of automation - and they're doing it earlier, leaner, and faster than ever before.

These ten companies represent just the beginning.

If you're building in this space, or backing those who are, we want to meet you.

Vantage Point: Robotics March 2026

Robotics is entering its inflection point — join the platform bringing together top founders, investors, partners and labs shaping embodied intelligence in Europe.

vantage.fov.ventures

A curated investor summit and demo day connecting Europe's most promising robotics founders with the investors and industry leaders shaping embodied intelligence.

Request your spot: vantage.fov.ventures

Welcome to Viewpoints, news and investment activity covering the next generation of computing - brought to you by FOV Ventures.

This Month:

• Our Slush Recap

• What is Gaussian Splatting and why you should pay attention now

• Portfolio wins: Roro's €2.7M seed, IXI's 22g prototype, and more!

Dave Haynes, Petri Rajahalme, Sointu Karjalainen & David Ripert (FOV Ventures)

This was hands down our biggest Slush yet. We led with our annual pre-slush Minus One event with Google and KPMG - bringing our portfolio together for dinner before bringing in over 250 more investors and founders into our wonderful venue at Allas Sea Pool.

We also organised and hosted an excellent Female LP side event (more on that below) and collaborated with OpenAI’s Wulfie Bain and Alice Guo for an intimate session with a curated group of technical founders. Between that and our invitation to the exclusive Slush launch of Gemini 3 we were blown away by the accelerating pace of new AI capabilities and the fact that Helsinki was at the centre of that for European startups, during the conference.

Another big topic this year, even dominating on Slush’s main stage, was defence tech. European defence startups raised $4.2 billion in the first nine months of 2025, putting the year on track for record investment.

For us, the defence opportunity sits squarely at the intersection of our core thesis: spatial computing technologies that enable machines to see, understand, and act in the physical world have obvious dual-use applications. More on that here.

Case in point: our portfolio company Distance Technologies brought their XR Night Vision system to Slush - military-grade optics that fuse lidar, infrared, and thermal data into 3D displays. The demo drew serious attention from both contractors and commercial players, along with serval other interesting robotics displays.

Both defence and robotics applications share the same fundamental requirement: high-fidelity 3D representations of the physical world. Whether it's a military vehicle operator navigating at night or a robot manipulating objects in a warehouse, these systems need to understand spatial relationships with precision and speed that traditional 3D modelling simply can't deliver.

Which brings us to two announcements that Meta made last week -

• First, SAM 3D - AI that takes any 2D image and reconstructs it in 3D with accurate geometry, texture, and pose. They're already shipping it in Facebook Marketplace so you can preview furniture in your room before buying.

• Second, Meta Hyperscape, which lets you create photorealistic 3D replicas of real spaces using just a Quest 3, then share and explore them with up to 8 people simultaneously.

Meta’s Hyperscape Capture

3D capture has been a core part of our investment thesis for years. We see it as essential infrastructure for spatial computing - whether that's training AI systems to understand physical environments, enabling defence applications like Distance's system, or powering the next generation of robotics.

One of the key breakthroughs making scalable 3D capture possible is Gaussian Splatting, a family of AI techniques that can reconstruct photorealistic 3D from ordinary photos or video, delivering the speed and fidelity required to finally bridge the digital-physical divide.

We caught up with Michael Rubloff of RadianceFields.com to understand why this matters now, where the opportunities are, and how this unlocks the infrastructure layer for spatial computing and physical AI.

Full piece with Michael Rubloff here.

FOV NEWS

A new robotics movement is forming…

Something significant is coming. We’re not announcing full details yet — but those driving the next era of robotics in Europe will want a front-row seat. If robotics is part of your future…

Making the Invisible Visible: The Female LP Pathways Survey

Mid-week at Slush, our Female LP Side Event drew 150+ attendees with Anisah Osman Britton MBE moderating a powerful panel on pathways into LP investing. The conversations proved there's a real appetite and economic imperative for bringing more women into venture.

Our part in this is launching The Female LP Pathways Survey - the first Europe-wide study mapping how women invest in venture capital, the barriers they face, and what could unlock the next wave of female LPs.

This 6-10 minute anonymous survey invites all women in Europe interested in venture — from seasoned investors to those simply curious — to share their story.

Every response feeds into a transparent, open-access dataset publishing in 2026, available to funds, LPs, networks, and policymakers.

In collaboration with Tesi, unlock VC, Ada Ventures, Kvanted, and many more partners, we're building a shared resource the whole ecosystem can use.

Change starts with visibility. When we make the invisible visible, we make it possible for thousands more to join.

Read the full piece: 

Making the Invisible Visible: The Women Behind Europe’s Venture Capital Flows

The Female LP Pathways Survey launches to uncover the journeys, barriers, and opportunities for women in venture.

viewpoints.fov.ventures/p/making-the-invisible-visible-the-women-behind-europe-s-venture-capital-flows

FOV PORTFOLIO NEWS

• Roro raised €2.7M seed led by Behold Ventures, node.vc, and 1Up Ventures to build Hello Home, the first true 3D digital dollhouse launching next year, bringing gaming as the new social to an expanding demographic looking to create and connect through self-expression—the team is hiring to build little homes with big dreams. (more)

• Spogen ai was selected as Startup Winner of the Tuottava Idea 2025 competition, earning recognition on the Nasdaq Times Square billboard and the honor of ringing the Nasdaq bell in Helsinki. (more)

• Scenario partnered with Alan, a leading HealthTech company with over one million members, enabling them to produce 10× more on-brand content across image, video, and 3D—including 27,000 personalized postcards in 48 hours—turning speed and creative control into a competitive advantage. (more)

• M-XR launched the Marso Measure 3D Capture Programme starting January 2026, offering 3D scan artists free access to Marso Measure with prizes including up to £3,000, Sony A7 Mark IV cameras, annual licenses, and Certified Scan Artist status for the best scanners. (more)

• Distance Technologies showcased their latest capabilities at a private Slush event, demonstrating how they're seamlessly fusing visible, infrared, and thermal signals with pixel-level depth precision to power physical AI with superhuman visual acuity for defence and automotive applications. (more)

• IXI revealed their latest prototype weighing just 22 grams—the lightest smart glasses frame in the world—featuring true-fit frame architecture that allows precise, personalized adjustments at temples, nose piece, and lens angle for all-day comfort, with the waitlist now open. (more)

• Atopia partnered with The Metropolitan Museum of Art to transform two iconic collections into immersive worlds explorable for free in web or VR: Oceania featuring Pacific-inspired landscapes from the Rockefeller Wing, and Dendur Decoded exploring the Temple of Dendur's story through interactive chapters. (more)

FOV JOBS

VIEW ALL OPEN ROLES HERE

• Brainspark Games is hiring a PR Events & Marketing Assistant Manager in London, UK to help amplify their AI-driven educational games following their Dragons' Den investment. (apply)

• AI Coustics has an opening for an Experienced Backend Developer (Mid-Senior Level) in Berlin, Germany to build infrastructure for their speech enhancement technology. (apply)

• Roro is hiring two Unreal programmers in Stockholm: Unreal Gameplay Systems Programmer and Unreal Generalist Programmer (both Mid-Senior Level) to build Hello Home, the first 3D digital dollhouse. (apply)

• Atopia Space is seeking an SDR / Outbound Sales Freelancer (Remote or Munich) to help scale their metaverse platform for arts & culture. (apply)

• IXI has an Entry Level Development Engineer position open across multiple locations for their optical technology team. (apply)

• Doublepoint is hiring a Senior ML Engineer in Helsinki, Finland to advance their gesture control technology for AR experiences. (apply)

Michael is the founder of Radiance Fields, a consulting firm specialising in 3D reconstruction technology. He built his expertise and online presence in radiance fields before the technology hit mainstream adoption, positioning him at the forefront of this emerging space. Through his consulting work with corporations exploring 3D applications and his industry-recognised content, Michael offers both technical depth and market perspective on how radiance fields and Gaussian splatting are reshaping industries from e-commerce to construction.

Every digital product today pretends reality is flat.

For 40 years, we've forced the complexity of the 3D world onto rectangular screens. This worked well enough for film and web content, but it created two structural gaps that are now barriers to progress:

1. Physics Gap: Traditional 2D tools and manual 3D modelling cannot accurately capture the physical behaviour of light and materials in the real world.

2. Scale Gap: Traditional 3D pipelines are too slow and labour-intensive to digitise the physical world at the speed and fidelity required.

If computing is going truly spatial - if AI, robots, and AR are to operate convincingly in physical environments - digital objects must behave like real ones.

The industry's historic solution - creation (modelling scenes by hand in tools like Blender and Unreal) - simply does not scale.

The revolution lies in switching focus from manual creation to scalable capture.

The technical breakthrough is a family of AI techniques called Radiance Fields, with Gaussian Splatting (GS) emerging as the most practical implementation. GS can reconstruct photorealistic 3D from ordinary photos or video.

Gaussian splatting excels at capturing large, expansive spaces without sacrificing smaller details.

Gaussian Splatting is the first 3D technology to simultaneously deliver the necessary speed and fidelity to bridge the digital and physical divide. It transforms the capture of real-world physics into a scalable, accessible process.

At FOV Ventures, this capability is the foundation of our investment thesis: spatial computing requires solving content creation at scale first. We believe this technology creates the essential, missing infrastructure layer.

We spoke with Michael Rubloff of RadianceFields.com to get a view on this rapidly accelerating space and what is now suddenly possible.

Why This Matters Now: The Shift to Real-Time Reality

The entire value proposition of Radiance Fields lies in their technical depth - they bypass traditional 3D construction entirely.

Instead of creating geometry - the manual, low-fidelity process of modelling with polygons - these AI methods learn how light travels through space from multiple input images. This captures not just shape, but the physical behaviour of light, ensuring material fidelity that manual methods could never match.

Polygons vs Splats: two ways of building a 3D world

While this concept emerged years ago, the initial implementations, like Neural Radiance Fields (NeRFs), were computationally intensive. They delivered photorealism but required vast compute - often minutes to render a single frame - making them impractical for any real-time application. However, in 2022, that changed with the publishing of NVIDIA’s Instant NGP, dropping reconstruction time and VRAM requirements. 

The critical leap occurred with Gaussian Splatting (GS), which won best paper at SIGGRAPH in 2023. The paper was published in April that year and for months, nobody reacted. Then it won at SIGGRAPH and the response was immediate - companies like Niantic's Scanniverse pivoted overnight, redirecting all resources toward Gaussian splatting.

While NeRFs still produce better visual quality in many cases, they can't be distributed easily across devices or rendered in real-time on standard hardware (for now). Gaussian splatting achieves the necessary compromise by representing scenes as millions of explicit, optimised 3D points rather than an implicit neural network. This allows photorealistic data to be efficiently rasterised on standard GPUs at real-time speeds (100+ FPS), transforming a research curiosity into something commercially viable.

A small, but growing number of engineers Michael speaks with are questioning the long-term viability of triangular meshes as the fundamental representation for 3D. NeRFs have the capabilities of producing photorealistic 3D in files under 1MB - smaller than traditional meshes (although more work here is needed). The only problem is they need powerful GPUs to render. But as GPUs get cheaper and faster, that constraint disappears. Some engineers think neural representations like NeRFs will eventually replace meshes entirely - smaller file sizes, better quality, and they capture how light actually behaves rather than approximating it with geometry.

Gaussian’s practical advantage came from three things converging at the same time:

1. Improved Camera Accessibility: Now any camera can produce the data for radiance fields, since their input is normal 2D images.

2. Accessible Compute: Real-time rendering runs on standard GPUs, not expensive datacenters.

3. Sophisticated AI Models: Can handle messy, real-world conditions (shake, imperfect lighting).

As a result, the technology is rapidly moving from research to production.

Meta's recently launched Hyperscape uses radiance field capture on Quest headsets to create photorealistic virtual replicas of physical spaces, now rolling out multiplayer features so multiple people can explore scanned rooms together.

Hyperscape Capture remains one of the highest profile deployments of Gaussian Splatting for consumer VR, translating a scan made on a standalone headset into a view dependent, radiance field representation that users can revisit later.

Meta have also released SAM 3D Objects for single-image 3D reconstruction, which powers Facebook Marketplace's "View in Room" feature for furniture visualisation. The research is rapidly moving beyond demos into real products being shipped to serve users. When Big Tech invests at this scale across complementary 3D technologies, it validates that the infrastructure layer has crossed from experimental to essential.

Hardware too is evolving to match the software capabilities.

The capture barrier is rapidly falling. Today, smartphone cameras - particularly the iPhone with its depth of field, sharpness, and computational photography - are good enough for quality reconstructions with no specialised equipment meaning people can build photorealistic 3D models with the phone in their pocket.

And elsewhere, consumer hardware is being repurposed in unexpected ways. Action cameras like Insta360, originally designed for extreme sports footage, are now being rigged up in arrays of 30+ units for volumetric capture. Companies are using tools in ways manufacturers never intended, a clear signal that demand is outpacing purpose-built solutions.

XGRIDS arrived as the first cameras designed specifically for Gaussian splatting capture, with built-in SLAM (Simultaneous Localisation and Mapping) for spatial tracking. At $5,000, the Portal sits at prosumer pricing - expensive for consumers but within range of what professionals already spend on high-end mirrorless cameras and lenses.

Michael expects this tier to expand rapidly over the next two years, with truly consumer-grade devices arriving within five years. Major manufacturers like Canon and Sony will likely be forced to release lines that shoot natively in 3D.

The trajectory mirrors photography's evolution:

• → professional equipment shrinks to consumer form factors

• →→ costs drop by orders of magnitude

• →→→ suddenly everyone creates content that once required specialists.

When capture becomes that accessible, the constraint shifts from "can we create enough 3D content" to "what do we build with all this 3D data".

This shift unlocks immediate, high-value opportunities across multiple verticals.

Where The Opportunities Are

Michael has been tracking this space long enough to separate near-term opportunities from distant possibilities.

His assessment is that startups need to focus on industries with clear, funded problems where better 3D visualisation creates immediate, measurable value rather than chasing consumer applications that may take years to materialise.

1. Hard services industries offer the clearest path.

Geospatial companies, architecture firms, and construction operations already pay significant money for 3D data through LiDAR scans, photogrammetry, and drone surveys.

Current methods work but they're expensive and time-consuming, which means radiance fields don't have to convince these buyers that 3D data is valuable - they just have to prove they can deliver it faster and cheaper. "Even the visualisation itself is a huge step forward from what they have traditionally worked with," Michael notes. The value proposition is immediate: better spatial data, lower cost per capture, faster turnaround from field work to usable models.

Horizon-GS, tackles the unified reconstruction and rendering for aerial and street views with a new training strategy, overcoming viewpoint discrepancies to generate high-fidelity scenes.

Architecture, engineering, and construction workflows need accurate spatial documentation for planning, change detection, and as-built verification. These industries understand ROI in concrete terms, which makes them attractive beachheads for startups that can navigate longer sales cycles and complex procurement processes.

2. Education and training applications create repeatable value.

Creating realistic training environments typically requires either expensive physical mockups or months of 3D modelling work by specialised artists.

Radiance fields let you capture a real space once and train people in that exact environment repeatedly, which matters enormously for medical training, industrial safety protocols, and military simulations where the physical environment is critical and mistakes carry real consequences.

The content creation bottleneck has kept many training applications stuck in simplified 3D environments or 2D video instructional content. Radiance fields remove that constraint entirely, enabling organisations to build libraries of real-world scenarios that would be too dangerous, expensive, or logistically complex to recreate physically for every training session.

This capability is already being demonstrated in the medical field, as 3D Gaussian Splatting makes cinematic, photorealistic, interactive 3D visualisations of human anatomy accessible on consumer devices, potentially moving complex anatomical study from restricted lecture halls to mobile, personal learning environments.

Application of 3D Gaussian Splatting for Cinematic Anatomy on Consumer Devices

3. The middleware layer remains mostly unbuilt.

Reconstruction tools exist, viewers exist, but the connective tissue between capture and application is sparse. Major platforms are already adding native support - Houdini announced Gaussian splatting integration, as did Octane Render, and plugins exist for Unity and Unreal. Michael expects consolidation here, with existing software platforms absorbing basic functionality rather than startups building standalone tools.

The question for companies building in this layer is whether they're creating features that eventually get absorbed or solving problems those platforms fundamentally can't address because they're too general-purpose. The opportunities exist in workflow-specific tools: automated quality checking for captures, format conversion pipelines, optimisation tools that reduce file sizes while preserving visual fidelity, and integration bridges between capture hardware and industry-standard software.

4. Vertical-specific solutions will win over horizontal platforms.

Michael emphasises there won't be one winner serving all industries. Each vertical has different technical requirements and workflow needs. Geospatial applications need outdoor capture at scale with georeferencing and integration with GIS systems. E-commerce needs controlled lighting, precise colour accuracy, and fast iteration for product catalog consistency. Architecture requires integration with BIM workflows, construction documentation standards, and compliance with building codes. Training applications need real-time manipulation, scenario editing capabilities, and assessment tracking.

Companies building horizontal platforms will lose to specialists who understand specific industry requirements deeply enough to build the right features, integrations, and go-to-market strategies that resonate with buyers who already have established workflows and purchasing processes.

5. Consumer and prosumer tools are emerging.

The fundamental accessibility of Gaussian Splatting - running on commodity hardware - guarantees a vibrant market for consumer and prosumer tools, even if the immediate ROI is less defined than in enterprise applications. The technical capability is already shifting the creator economy around 3D content.

Beyond creation, this accessibility enables the most personal use case: Memory Capture. Families can now scan rooms, objects, and moments, creating the spatial equivalent of photo albums. Apple's adoption of Gaussian Splatting for its Vision Pro platform validates this future, using the technology to power its highly immersive "Spatial Scenes" and the lifelike "Personas" used in virtual communication. The technical capability exists today to capture grandparents, childhood homes, and meaningful objects in full 3D, creating 3D versions of important moments that people will wish they had 50 years from now.

Apple’s “Spatial Scenes”

Building on this, Prosumer creators are finding workflows. Tools like PlayCanvas's Super Splat make basic editing accessible, and platforms like Arrival Space enable sharing captures with spatial audio and multi-user exploration - imagine touring a friend's apartment renovation or a wedding venue before booking.

The creator economy hasn't fully materialised yet, but the tools are readying the opportunity for platforms to distribute at scale.

The Ultimate Prize: Gaussian Splatting as the Foundation for Physical AI

While the commercial opportunities in AEC and training are significant, the ultimate, high-value application of radiance fields is much more than better graphics - but creating the foundational data layer required for Physical AI.

AI models need massive amounts of high-quality 3D data to truly understand spatial relationships and physical behaviour. You can't train robots to navigate spaces or manipulate objects without photorealistic 3D representations of how the world actually works.

NVIDIA's integration of Gaussian Splatting into Omniverse and Isaac Sim through their NuRec neural rendering API makes this concrete. They are using it to create physically accurate digital twins of factory floors and real-world spaces where robots are trained before deployment.

Integrating 4D Gaussian Splats into Omniverse and Isaac Sim

Progress, Challenges, and What Comes Next

insights.evercoast.ai/p/integrating-4d-gaussian-splats-into

Their Physically Embodied Gaussians (real geometry, surfaces, and properties) approach combines particles for physics simulation with Gaussians for visual appearance, creating a continuously synchronised world model. This speed is key: Gaussian splatting enables real-time SLAM, allowing robots to build live internal models of environments as they navigate.

Research groups are already building on this foundation. SplaTAM and Gaussian-SLAM create dense, photorealistic maps of unknown environments in real-time as robots explore. GaussianGrasper allows robots to understand and interact with objects using language prompts and view-robust policies and Volvo uses Gaussian splatting for safety simulations and virtual world creation for autonomous vehicle development.

The pattern = wherever robots need to understand physical spaces, Gaussian splatting provides the training data.

GaussianGrasper lets a robot look at just a few pictures of a scene, quickly build a simple 3D version of it using “Gaussian splats,” understand which object you’re talking about (from your words), and then figure out the safest place to grab it.

Gaussian Splatting to enhance, train and validate Volvo’s driver safety systems.

This connects directly to FOV's thesis on physical AI and robotics.

The convergence of spatial computing, computer vision, and embodied AI all depend on the same underlying requirement: high-fidelity representations of the physical world. Radiance fields provide that foundation in ways traditional 3D modelling never could, because they capture not just geometry but the physical behaviour of light, materials, and spatial relationships that AI systems need to learn from.

Companies like Gauzilla Pro (web editor for 3D Gaussian Splatting) are already thinking beyond static visualisation toward how lifelike 3D becomes the foundation for AI systems that need to understand and interact with physical spaces - and their tools make those high-fidelity 3D maps simple to create.

This convergence is not linear; but a powerful feedback loop.

Better capture methods create superior training data, which produces more capable AI models, which enable smarter robotic systems, all demanding even higher fidelity spatial data. Radiance fields finally provide the bedrock for this loop.

This is the missing infrastructure that will unlock physical-world computing.

At FOV Ventures, we're investing in companies solving different pieces of this stack - from material capture to object digitisation to spatial AI applications. The companies that win will be the ones that understand both the technical capabilities and the specific industry problems where those capabilities create measurable value.

If you're building at the intersection of 3D capture, spatial computing, or physical AI, we'd like to hear from you. Reach out.

Thanks to Michael Rubloff of RadianceFields.com for sharing his insights on this rapidly evolving space. For anyone working with or researching radiance fields, his site remains the definitive resource for tracking developments in NeRFs, Gaussian splatting, and related technologies.

Twenty years ago, a few women quietly changed the course of my life - opening doors, sharing playbooks, and showing me that venture capital could be a path long before I even knew it was one.

Now, in a rare full-circle moment, one of those early mentors has joined as an LP in a new syndicate we’re building - a vehicle designed to make smaller ticket entries possible for women (and men) alike.

It’s proof of something I’ve seen again and again: representation compounds. When one woman invests, she often opens the path for many more.
But for most women, those bridges don’t yet exist - or if they do, they’re hidden.

We’ve all seen the charts showing how few female founders and GPs there are in Europe. But there’s another layer shaping everything that happens above it - the people deciding where the money flows.

The limited partners.

While women now make up slightly more of the LP base than of GP seats, their presence remains largely invisible. We don’t have real data on their journeys - what drew them in, what held them back, or what could make it easier for others to follow.

Because who funds the funds decides who gets funded.

That’s why we’re launching the Female LP Pathways Survey - the first Europe-wide study mapping how women invest in venture, the obstacles they face, and the opportunities that could unlock the next wave of female funders.

There’s no visible roadmap for women who want to start investing in funds - we’re each building our own bridge as we walk it. That’s why this survey matters - it can make the path clearer for those who come next.

The message is clear: progress begins with visibility. Understanding who gets access — and who doesn’t — is the first step toward lasting change.

Building a Shared Map

In collaboration with FOV Ventures, Tesi, unlock VC, Ada Ventures, Kvanted, and many more, this 6–10-minute anonymous survey invites all women in Europe interested in venture — from seasoned investors to those simply curious — to share their story.

Every response will feed into a transparent, open-access dataset published in 2026 — available to funds, LPs, networks, and policymakers alike. Because this isn’t about another private report. It’s about building a shared resource the whole ecosystem can use.

We’ve seen growing data around female founders and GPs — but the LP side remains a blind spot. I’ve seen what happens when representation compounds: when one woman invests, she makes it possible for many more to follow. This survey is about making those journeys visible — turning curiosity into clarity, and clarity into change.

Why Now

Governments and institutions are starting to take notice - from the British Business Bank’s £500 million commitment to diverse fund managers to EIF’s inclusion targets.

But we’re still missing the upstream piece: who’s investing behind the funds, and what’s needed to make LP participation more inclusive.

If we want capital to flow more equitably, we need to start at the source - with the people who decide where it goes. 

Your Invitation

If you’ve ever thought about investing into a fund but weren’t sure where to start - this survey is for you.
If you’re already investing and want to see more women beside you - this survey is for you.
If you work at a fund and want to understand how to welcome more first-time LPs - this survey is for you.

Take the 6-10-minute Female LP Pathways Survey
Help us map the pathways, barriers, and opportunities for women investing in venture. 

A Shared Future

Change starts with visibility.
When we can see the paths that already exist, we can build better ones.
When we make the invisible visible, we make it possible for thousands more to join.

This isn’t just about who invests - it’s about who gets to shape the future of venture capital.

Let’s build that future together.

Help us spread the word

Change starts with visibility.

When we can see the paths that already exist, we can build better ones.
When we make the invisible visible, we make it possible for thousands more to join.
The more voices we reach, the stronger the data becomes.

If you or your organisation want to support this movement, you can find ready-to-share assets and templates here. 

A Note of Thanks

This initiative wouldn’t exist without the support of all the amazing partners behind this initiative. Thank you. And my partners and GPs at FOV Ventures - who never once questioned the idea, only asked, “How can we help?”

And to every woman - and ally - who’s shared ideas, experiences, or encouragement along the way: thank you. Your openness is what makes change possible - and unstoppable.