Patrik Liu Tran / Validio

Poor data has long been an expensive, if unglamorous, problem inside large organisations. As companies generate and store ever larger volumes of information, ensuring that data is accurate, complete and usable has become increasingly difficult. 

Now, with the rapid adoption of AI across enterprise software, the issue is becoming even more acute. AI systems depend heavily on large datasets to train models, automate decisions and power new services — but if that underlying data is flawed, the results can be unreliable or even harmful.

Research from Gartner suggests that data quality and availability remain among the biggest obstacles to implementing AI at scale. Enterprises may invest heavily in data infrastructure and machine learning systems, but those investments often stall before reaching production because the data feeding them is incomplete, inconsistent or corrupted as it moves through pipelines.

For Swedish startup Validio, this growing problem represents a major opportunity. The company has raised a $30m Series A round led by Plural, as demand grows for tools that help companies ensure their data is trustworthy.

Previous investors and angels including Lakestar, J12, MongoDB’s co-founder Kevin Ryan, Snowflake’s CMO Denise Persson and Neo4j’s co-founder Emil Eifrem also participated in the round, bringing the startup’s total funding to $47m.

CEO Patrik Liu Tran founded Validio in 2019 with the aim of solving one of the most persistent challenges in enterprise data: ensuring that information remains reliable as it moves through increasingly complex data systems.

“Data management is one of those tasks where you’re not really replacing people, you’re doing work that never was done to begin with,” Tran says in an interview with Pathfounders. “95% of enterprise data is of poor quality. That’s just how it is out there. The absolute overwhelming majority of data out there is of poor quality, and why is no one taking care of that?”

Before founding the company, Tran spent years advising large organisations on their data and AI strategies. Tran says he built around 20 internal data consistency tools for different companies before deciding the problem required a dedicated platform. That realisation ultimately led him to build Validio — the name shot for valid input and output.

Today the startup describes itself as an agentic enterprise data management platform designed to automate the monitoring and validation of enterprise data. The platform includes more than 90 automated validators that check datasets for common quality issues, such as whether data is accurate, whether it has been distorted while moving between systems, whether it is up to date, whether pipelines have been disrupted and whether datasets are complete.

Part of the challenge is scale. As businesses adopt cloud infrastructure, streaming pipelines and real-time analytics, the speed and volume of data production have grown dramatically — making manual monitoring almost impossible.

“The data velocity is so huge and the amount of data that’s being produced is growing exponentially over time,” he continues. “So there is no way that we as humans can manually go through all of this when we couldn’t even do it a decade ago. So I think the agentic capabilities makes things much more scalable, even more than before.”

Validio’s platform can be deployed in several ways depending on customer requirements, including cloud environments, fully isolated virtual private clouds or on-premise installations. This flexibility is particularly important for highly regulated industries where sensitive data cannot leave secure infrastructure.

The company’s customer base already includes organisations such as Canva, Nordea, Deutsche Glasfaser, Truecaller, Surfshark, Walden and AllianceBernstein. Many of their customers operate in sectors where reliable data is critical to operations, including financial services, telecommunications and manufacturing.

These industries are also increasingly experimenting with AI-driven automation, making the reliability of underlying data even more important.

“Data quality is no longer a nice-to-have — it’s mission-critical for AI success,” Persson said in a statement. “Validio tackles the data trust challenge head-on with an AI-powered quality and lineage platform for all types of data, AI and agentic workloads.”

While Validio’s main competitors largely remain traditional data management tools, there are a handful of newer startups building dedicated platforms focused specifically on data reliability. 

Superconductive, which created the open source data quality tool Great Expectations, raised $40m Series B funding in 2022, bringing its total raised to $61m. In the same year, another data observability startup, Monte Carlo, raised $135m Series D at a $1.6bn valuation.

With its new funding, Validio plans to expand its presence internationally, particularly in the US and UK. The company is setting up offices in New York and London while continuing to grow its engineering team in Sweden.

Its focus remains primarily on large enterprises, particularly those in the Fortune 2000.

Anirudh and Akshay Oppiliappan / Tangled founders

“Digital sovereignty” has been a hot topic in Europe for the past few years, and it has been the buzzword of 2026 so far.

From cloud infrastructure to AI regulation, policymakers and founders alike have debated how much of the continent’s technological backbone should rest in American hands. Now, a pair of brothers building between Helsinki and London are bringing the debate right down to the level of where code itself lives.

Tangled, a next-generation code collaboration platform founded by Akshay and Anirudh Oppiliappan, is positioning itself as a European, open-source alternative to GitHub — one designed around decentralisation, transparency and developer ownership.

In an interview with Pathfounders, Anirudh Oppiliappan detailed the company’s mission to “giv[e] the community a choice that prioritises transparency and developer sovereignty”.

The brothers have just secured $4.5m in seed funding to work on that mission. The funding round was led by byFounders alongside Bain Capital Crypto, Antler and a number of prominent angels including the former CEO of GitHub, Thomas Dohmke, CEO of Tailscale, Avery Pennarun and former CEO of MySQL and HackerOne, Mårten Mickos.

Dohmke left GitHub last August after four years as CEO and last week raised $60m at a $300m valuation for his new startup Entire, an open source tool to help developers manage AI agent-written code.

He was at one point rumoured to be working on an alternative to GitHub which is how Oppiliappan originally crossed paths with Dohmke, but it turned out he wasn’t and instead decided to back the startup. 

While the brothers are Indian by origin, their geographic and regulatory “European” framing is something they thought about from the outset. Anirudh moved to Helsinki for a role at Finnish cloud provider UpCloud around the same time Akshay relocated to London to join code search engine Bloop and they were conscious of building within European jurisdiction.

“We get that there’s data privacy laws that are way better here and we’re happy to be building within the jurisdiction, and having our data protected by that,” Anirudh Oppiliappan tells Pathfounders. “But we love to say that Tangled is built in Europe, but for the world.”

But they do reap the benefits of being a “European” company, and even the American users want to take advantage of that.

Of Tangled’s more than 7,000 users, around 40–50% are American, roughly 40% European and about 10% based in Asia. For some US developers, the appeal lies partly in Europe’s regulatory environment and a desire for alternatives to large American incumbents.

The need for that alternative, according to Oppiliappan, has been building since Microsoft acquired GitHub in 2018. He argues that the platform’s centre of gravity shifted away from the open-source communities and independent developers that fuelled its rise.

“The open source communities and indie devs who made GitHub what it is today, they became the underserved user group on the platform, and there were no features or product development being targeted towards [them],” he says. “And then, of course, there’s also the looming ideological angle of Microsoft and I guess Copilot and all of their hyper AI stuff, which, to be fair, we are not opposed to , but we are also a lot more thoughtful about how we want to implement AI stuff down the line and not just shove it down your gullet, like how Microsoft does it.”

He added that GitHub itself is an “aging platform that hasn’t really meaningfully improved in time”. That’s why the brothers are trying to position Tangled as “the next generation of social coding and collaboration”.

The name is a nod to that vision. Tangled refers to a decentralised space made up of “knots” — self-hostable servers where code lives — forming an interconnected network rather than a single centralised repository. Instead of relying on a traditional API model, Tangled is built on Bluesky’s AT (Authenticated Transfer) Protocol, the open social networking protocol designed to enable portability and interoperability. By removing a central API layer, the founders say developers gain greater ownership over their work, identity and connections across an open network.

That open protocol architecture is also central to how Tangled views the future of software development in an AI-driven world.

“One of our core theses is that we need to fundamentally rebuild the foundational infrastructure for agentic programming,” Oppiliappan says. “As these agents become more and more autonomous and start working more autonomously in developing and writing code and features, we need this infrastructure piece — or some sort of wiring or framework — for them to hop on and operate on.”

Because Tangled is built on an open protocol, he sees Tangled as positioned to do just that.

In other words, the company is not just trying to compete with GitHub as it exists today, but to lay groundwork for a world in which AI agents collaborate alongside humans in shared codebases. The AT Protocol, originally developed for social networking, could also underpin interactions between “agentic actors” in the future.

Tangled went live last March, initially as an invite-only product. Within two hours news of the startup landed on the front page of Hacker News and demand surged. Since then, the team has remained lean and fully remote: one founder in Helsinki, the other in London. 

They have hired one engineer from South Korea, who they discovered through their Discord community, and another who previously worked with Anirudh. With fresh capital, they plan to bring on three more generalist engineers while continuing to expand their social and community spaces.

“The bottleneck in software development has shifted from writing code to reviewing and managing it at scale,” Antler partner Jussi Kallasvuo said in a statement. “We backed Tangled because the team has the technical depth to build a system that manages this complexity for both humans and AI. This is a critical piece of infrastructure for the European tech ecosystem, providing a native alternative to American legacy platforms.”

From the Editor’s desk:

• Please take our SHORT, three-question, reader survey here.

• Any other thoughts, comments, criticism: contact@pathfounders.com

• If you’re attending Mobile World Congress or 4YFN in Barcelona this week, you might catch sight of Pathfounders Editor Mike Butcher. He’s on mike@pathfounders.com 

• Sign up to be the first to hear about the launch of our member network here.  Make sure you’re a subscriber to this newsletter and follow all our channels below. To partner with the site, newsletter, and/or events, contact our head of partnerships: Ana Williams, ana@pathfounders.com

Like your news and updates from us a little faster? Get our Pathfounders WhatsApp Announcements (appears in your main WA feed) OR get the Pathfounders Whatsapp Public channel (same content).

This newsletter is brought to you in partnership with: 

The Latest on Pathfounders:

• Our first mixer London Mixer “AI, Power & the Road to 2026” featuring a stellar VC and expert panel was held on Thursday, Feb 26, thanks to our partners Orrick UK and Nebius.

You can read all about it here:

Geopolitical tensions mean AI, dual-use, and 'sovereign tech' are here to stay, say investors at Pathfounders London event

The panel also identified a new generation of AI-native entrepreneurs, operating in palpably different ways from the generation that came before them.

The panel featured:

• Sitar Teli, Managing Partner, Connect Ventures

• Alex van Someren, recently exited as the UK's Chief Scientific Adviser for National Security and former VC 

• ⁠Manjari Chandran-Ramesh, Partner, Amadeus Capital Partners

• ⁠ Ana Barjasic, Connectology, European Innovation Council

Europe, it's time to put your money where your mouth is

In an open letter (exclusive to Pathfounders), the partners of three major VC funds argue that Europe — after proving it can create world-class startups — must now prove it can scale them with growth capital. That is, if it wants technological sovereignty from the US.

BeyondMath raises $18.5m to train AI that understands physics

The deeptech startup's simulation technology is already being used by Formula 1 teams and aerospace manufacturers to accelerate engineering design. By Amelia Isaacs

Podcast: Can Italy surf the AI wave?

Italy has long been synonymous with incredible design, engineering and industrial giants, but can it take advantage of the new AI era? Featuring Vento VC CEO Diyala D'Aveni. By Mike Butcher

Op-Eds and Contributed Columns

We’re ready to start receiving pitches for Op-Ed columns. These can be on tech, startups, investing, AI, the future, frankly, you name it. So long as it tech starup or VC-related and bloody interesting.

The rules are: A short paragraph outlining the piece. Followed by three or four bullet points about the main issues you’d cover. The argument must be backed up by third-party references. We’ll only take pieces we green-light. Nothing submitted ‘on-spec’. We will build a form for these submissions, but in the meantime, you can pitch us at contact@pathfounders.com. Pitching us more generally? Here’s all you ned to know about that subject. 

Hot Tips and News 

Please send all tip-offs, suggestions, and news to contact@pathfounders.com

SUBSCRIBE TO THIS NEWSLETTER

That’s all, folks

As always, you can interact with Pathfounders via the email contact@pathfounders.com and, even, via an anonymous form if you want to tell us something you think is important, but would prefer to keep some distance. You can even subscribe to Pathfounders WhatsApp Announcements (appears in your main WA feed) or the Pathfounders WhatsApp Public channel.

Follow Pathfounders:

Pathfounders WhatsApp Announcements (appears in your main WA feed)

Pathfounders WhatsApp Public channel

We are not on Substack (sorry), but we will publish headlines from here.

LinkedIn

X (Formerly Twitter)

BlueSky

Telegram

Facebook

Instagram

Luma (events)

RSS Feed for all posts

Podcast: Spotify

Spotify RSS

Async

Async RSS feed

YouTube

TikTok

Contact Pathfounders Anonymously

Speaking at the inaugural Pathfounders London Mixer  — the first in a series — a panel of VCs and experts tackled the thorny subjects of the white-hot era of AI, what’s next for deeptech, and the future of so-called ‘sovereign’ technologies — all amid the background of, at times, quite chaotic geopolitical tensions. 

Sitar Teli, Managing Partner, Connect Ventures, commented on the recent effect of AI on public markets: “What we're seeing is a lot of excitement [about AI] in the private markets, and a lot of panic in the public markets. And sometimes it’s hard to square those two things. But there is a lot of valid correction in the public markets when it comes to, particularly, SaaS. Not because it's suddenly worthless, but because it's worth less than it was before? The multiples are coming down. It doesn't mean it's going to zero. There is a correction in terms of what the value of those businesses is.”

Of the private, VC-driven markets, she added, “This is one of the most exciting times to be a VC. AI is one of the most fundamental technologies we've seen, and it genuinely feels very exciting. On the flip side, I have never seen so many bad ideas in my life.”

Teli described venture investing right now as looking like a “barbell” in terms of deployment: “You have some VCs that are deploying a lot, and then you have others [deploying] very little. We're definitely in the second bucket. We’re seeing a lot of companies that are in red ocean areas. For example, Avatar-based language learning. And you go to the App Store, and there's 30 of them. We're seeing a lot of bad ideas, and no one has any differentiation or defensibility.”

On the issue of sovereignty Ana Barjasic, founder, Connectology  — who was a four-year board member of the European Innovation Council (at 10 billion euros, the largest Innovation Fund in Europe) — said the EU was — suddenly faced with the full scale invasion of Ukraine — had to deal with the prospect of needing to add ‘dual-use’ technology to its investment thesis: “This was a very dramatic moment. However, [the question arose] who decides what is ‘sovereign’ tech?. And if you are engaging with public funds like the EIC, it's very important to understand that potentially, the growth of your company will be limited as a result of whatever the definition of the ‘sovereign’ tech is.”

Alex van Someren, of Capricorn Consulting, a former VC who recently exited as the UK's Chief Scientific Adviser for National Security, said that in the era of increasing geopolitical threats from the likes of Russia and China, “if you're an entrepreneur, it’s stopped being optional to deal with governments. You are going to have to confront having a better understanding of how governments work, how governments think, and build government relations. That is a pain in the ass, and I assure you it will be torturous and annoying. But at the same time, I think it will serve you very well to spend the time trying to understand how that works, because actually, you do not have a choice.”

He added that export regulations, dual-use regulations, sources of funding, access to supply chain,s and sovereignty, “are all things which governments have the ability to control.” Entrepreneurs are increasingly likely to find themselves faced with legal issues around their technologies. “That is actually quite a profound change, and it wasn't the case a few years ago,” he said.

On Deeptech and dual-use investing, ⁠Manjari Chandran-Ramesh, Partner in the Early Stage Fund of Amadeus Capital Partners, said: “I’ve been doing deep tech investing for the last 11 years, and Amadeus Capital has been doing it for 30 years, way before it became cool. And now it's actually cool! We have generalist investors coming to us and saying, ‘Hey, I love this quantum’ or ‘I love this robotics. Let me invest in this. Teach me more’.”

However, she said startups needed to be wary about how much money they get from government institutions. “It's possible to manage it such that you are still able to exit to the US, or sell to other countries, but it requires very careful handling,” she said.

The panel also concurred that there was now a generation of founders who were ‘AI-native’ — many being on the younger side who had simply grown up developing with AI tools and behaved and produced products in a fundamentally different way from previous generations of founders.

“The last three deals we did were built by people who just grew up with AI,” said Teli. “ They're in their late 20s, and they started coding when they were teenagers. They read the original AI paper that came out in 2017, and they started working with the technology super, super early, back when it was deep learning, and then transitioned into generative AI. And they've never known a world of building product without it.”

The panel also emphasised how Europe could have the ‘edge’ in the realm of quantum computing.

Later, during round table networking sessions, attendees chewed the fat over subjects such as  “What does an exit look like in 2026” and “What to build in the era of AI”.

The entire panel discussion will soon be available on the Pathfounders YouTube channel, and the audio will be available via the Pathfounders Podcast (available on Apple iTunes and Spotify).

The event was partnered with international law firm Orrick and AI cloud provider Nebius.

François Robinet, Managing Partner at AVP
Olivier Tonneau, Founder and Partner at Quantonation 
Andreas Schwarzenbrunner, General Partner at Speedinvest 

After ten years of ecosystem building, European tech must now prove its capacity and commitment to execute at scale.

The continent stands at an inflection point, with entrepreneurs building globally competitive companies in critical and fast-growing industries. Across AI, quantum computing, life and health sciences, robotics, and energy, Europe has the opportunity to build and scale critical businesses, essential to Europe’s economic resilience and technological sovereignty.

Collaboration is key to unlocking this future.

Improving access to capital through coordinated action is now central to Europe’s ability to compete for global leadership in innovation. Initiatives such as Build in Europe, Project Europe, the European Commission’s renewed EU-Inc agenda, and EU-backed vehicles like the ScaleUp Europe Fund signal growing alignment across founders, policymakers, and capital providers.

Finding common ground and forming partnerships is essential to bridging Europe’s existing paradox: that world-class science, innovation, and talent are hindered by a lack of scale capital. As venture capitalists, we must help build the structural conditions that allow European companies to scale globally. This moment demands cooperation among European investors willing to pool their expertise, capital, and conviction across markets, stages, and sectors to unify a fragmented ecosystem.

Europe’s Structural Scale Gap

Europe is home to some of the world’s most promising technical talent and scientific research, yet we consistently underinvest in scaling the outcomes, particularly in R&D-heavy sectors like AI, defense, and deep tech. Europe invests up to 3x less per capita in late-stage VC than the US and allocates a tiny amount of pension funding to startups.

The well-publicized growth capital gap must be addressed for Europe to succeed. Funding the future at scale can only be achieved through tackling regulatory fragmentation, talent access gaps, and fundamentally changing how long-term institutional capital is allocated to innovation.

Leading US venture platforms combine scale, integrated capital, and long-term institutional backing, supported by deep public markets and a mature base of pension and endowment capital, to support companies from inception to global expansion. Their ability to deploy large, conviction-led rounds at critical stages reinforces a powerful cycle of ambition, talent, and capital. Europe doesn't need to replicate the US model, but it must learn from what works: building larger pan-European platforms, aligning policy and private capital, and convincing long-term institutions that Europe is ready for capital at scale.

Particularly in capital-intensive and long-cycle innovation areas like deep tech, the need for European growth funding has never been more pronounced. Unlike later-stage financial capital, venture and growth investors are built to support ambition, uncertainty, and rapid technological change. Too often, Europe’s most promising companies face a binary choice: slow down or raise large late-stage rounds from non-European capital, with an inevitable shift in strategic gravity.

As venture capital investors, operational expertise meets risk appetite for innovation. Beyond balance sheets, European VC is uniquely positioned to identify and scale with the continent’s most ambitious founders.

According to Atomico’s State of European Tech Report, the number of startups in Europe has increased fivefold since 2015. Recent figures show that European venture investments exceeded $85 billion across more than 8,600 deals in 2025, one of the strongest years on record.

Momentum across sectors and geographies suggests a window for decisive action. Alongside the growth in the number of startups, the venture capital ecosystem is also thriving with new, emerging managers competing with larger and more storied institutions, increasing the value proposition for founders. This geographic expansion is visible across Central and Eastern Europe, Southern Europe, and the Nordics — regions rich in talent but historically less served by venture capital.

Addressing the continent’s digital and capital divide is now central to the European Commission’s policy agenda. It must also be central to investment strategy: looking beyond traditional hubs, backing scale-ups from less capital-dense ecosystems, and working through strong local partners. This demands a genuinely pan-European approach, reducing fragmentation, strengthening cross-border platforms, and building alliances across capital and industry. Sovereignty, in this context, is achieved through coordination and scale.

Why Growth Capital Is a Sovereignty Issue

Speaking of sovereignty, many of Europe’s most promising technologies are inherently dual-use. Quantum computing, robotics, space technologies, and advanced semiconductors sit at the intersection of civilian innovation and defence, resilience, and security. These sectors demand long time horizons, substantial capital, and alignment with public policy frameworks. Europe-based growth capital can better support multi-year R&D and industrialization cycles, align investment strategies with EU defence and security priorities, and act as a stable, long-term partner for companies operating in sensitive technological domains. Anchoring these investments in Europe helps ensure that strategic technologies remain aligned with European interests over time, without compromising commercial ambition.

Coordination at Scale

Europe has the talent, research, and capital, but it currently lacks alignment between countries, institutions, and ambition. What it needs is the financial infrastructure to translate innovation into leadership, and leadership into sovereignty. For investors, this represents a compelling opportunity: deep innovation, attractive relative valuations, and improving exit pathways as pan-European capital scales.

To succeed, these efforts must continue to deepen pan-European coordination, reduce structural fragmentation, and strengthen alliances among investors willing to back Europe’s next generation of global champions. The opportunity and ambition are there; it is now crucial to scale it up.

By:

François Robinet, Managing Partner at AVP
Olivier Tonneau, Founder and Partner at Quantonation 
Andreas Schwarzenbrunner, General Partner at Speedinvest 

The BeyondMath team

Modern engineering software has long been built the same way: one model, one problem. Early large language models, for example, were typically designed to do one task — translate text, answer questions or search. In physics and engineering, the same logic applied. One tool for airflow, another for heat or electromagnetics, each tool highly specialised to an individual use case.

Now, with GPT-style models and a foundational approach, one model can be applied to a wide array of problems, and, for BeyondMath, this looks like a foundational model of physics that it believes could transform how engineers work.

The Cambridge-founded deeptech company is building AI models that learn the laws of physics directly, enabling design iteration at much higher speeds. 

By training a single, foundational model to understand how fluids, heat and forces behave, the company believes it can apply the same system across automotive, aerospace and energy — from Formula 1 cars to aircraft components and renewable infrastructure — dramatically reducing the time it takes to go from design to production.

“We are building a foundational model of physics that can be applied to a wide range of problems, as opposed to a model that can do one problem in a narrow way. And we are quite differentiated in the market in that sense,” co-founder and CEO Alan Patterson tells Pathfounders.

The result should be faster simulation — enabling thousands of iterations in seconds rather than days — reducing costs, increasing efficiency and reducing the environmental impact.

That approach has now attracted fresh capital. BeyondMath has closed a $10m seed extension led by Cambridge Innovation Capital, alongside existing investors including UP.Partners, Insight Partners and InMotion Ventures. The round brings its total funding to $18.5m, following an $8.5m raise in 2024.

But the ambition extends far beyond speeding up simulations.

“It's not just the number of iterations and optimisation, but how fast you can go from paper to production. Why does it take five years? Why doesn't it take six months? And how do you make that happen? And when you go into how these companies work and how we build and engineer and manufacture, bringing down these timelines at every point of the chain is hugely important,” Patterson says.

That capability is already being deployed at scale. BeyondMath’s customers include major automotive, aerospace and electronics manufacturers, and it has partnerships with NVIDIA and AWS. In the US, it is working with engineering firm Honeywell on a $19m, three-year project with the firm to reduce the time it takes to simulate thousands of iterations of complex aircraft components from days to seconds.

In Formula 1 — where competitive advantage is measured in fractions of a second — BeyondMath supports real-time testing of thousands of design combinations. Engineers can simulate changes down to the literal nuts and bolts to optimise aerodynamics.

“Almost everything we build or manufacture deals in some way with this problem,” Patterson says.

The company was founded by AI industry veterans Patterson and Darren Garvey, marking their fourth time working together. They first crossed paths in 2010 at Evi Technologies (then True Knowledge), which was acquired by Amazon and went on to form the basis of Alexa. Later roles at eBay and HomeX followed, before they left in 2022 to launch BeyondMath.

Having studied electromagnetic simulation in the 90s, Patterson remembers it taking a week to run a simulation. Fast forward to 2022 and he saw it was still a huge bottleneck. But with advances in AI and machine learning over the past five years, a lot more has become possible when it comes to training AI to conceptualise physics to this level.

“I think even now the hardware is really only just catching up to the scale of the problem. And also the techniques that you have to employ means it's only really possible to do this now,” Garvey says.

There was some scepticism when they started given the level of AI adoption, but now, “it feels like this year is going to be a huge year for physics AI,” Garvey says. “I think the markets now are starting to recognise the value it can have. People are starting to really see it.”

“BeyondMath is tackling one of the hardest and most valuable problems in engineering,” Edward Inns, Principal at Cambridge Innovation Capital, said in a statement. “By combining first-principles physics with modern AI, the team has built a platform that can redefine how complex systems are designed across multiple industries.”

The new funding will be used to scale commercial deployments, increase its customer base across Europe, the US and Japan, and nearly double the company’s team by the end of the year. And, as Patterson notes, it will also mean “as ever with AI companies: more compute, more GPUs”.

Company info:

• Founded: Cambridge

• Office: London

• Team size: 18 people, with the aim of increasing this to between 32-40 by the end of the year

• Capital raised: $18.5m total

  • A $10m seed extension led by Cambridge Innovation Capital, alongside existing investors including UP.Partners, Insight Partners and InMotion Ventures

  • An $8.5m raise in 2024

According to the latest figures, Italy’s tech sector raised $768 million in the first half of 2025, blowing past figures for previous years. Healthtech alone pulled in $126m, fintech $91.6m, energy tech $78.9m, and enterprise software funding has nearly doubled since 2021 (figures from Vento VC). 

In theory, Italy is literally riding a wave. So is it finally having its ‘Tech Renaissance’? 

Startups seem to be on a roll. For instance, Milan-based AI security firm Aindo, recently closed a $20 million Series A, and Bologna-based distributed cloud storage provider Cubbit won EU backing.

The European Commission forecasted Italy’s GDP would grow by 0.9% in 2026, while inflation is expected to hold steady under 2%, which could establish a stable environment for investment.

And Italy is the EU’s second-largest manufacturing base, with dense industry, resilient supply chains, and deep engineering talent. For instance, STMicroelectronics recently launched a new €5 billion fab in Catania.

Now, the camera is cutting to Turin, the birthplace of Fiat, which annually hosts Italian Tech Week. Last year it drew over 15,000 attendees, headlined by Jeff Bezos and the EU head Ursula von der Leyen.

The event was always primarily organised by Vento, founded by Fiat scion John Elkann, one of Italy’s leading early-stage funds. But now it’s rebranding as ‘Wave, by Vento’ making the provenance more explicit. 

But can the event maintain its independence, just as Italy generates some tech momentum?

In this week’s Pathfounders podcast, Vento CEO Diyala D'Aveni told me: “Of course, we are still kind of behind compared to other European ecosystems, but we are on the right trajectory. And the numbers… are a validation of that. That's why we started to invest in this ecosystem in the first place a few years ago, because we were expecting this growth. And now for the first time, we see some new companies raising Series A, Series B, and doing well in this ecosystem.”

She said the name change to Wave reflects the increasingly international nature of the event: “One thing that I think wasn't clear is that it was always us organizing the event. [But] we always have other VCs on stage because our objective with the event is to bring the best people to the conversation. So we don't care if it's us or if it's someone else.”

The aim is “always to have the best founders and the best audience. And that's what we are working on at the moment and do,” she said.

With the name change, they are building “something that goes beyond Italy and possibly beyond Europe. So we need a name that can represent that idea and that intention.”

She added that although Italian founders are spread internationally, increasingly many are returning: “For instance, 50% of our companies are Italians outside of Italy. So for sure we have a brain drain problem. But what is good is that even those Italians that are not in Italy anymore are supporting the ecosystem, [they] are coming back to help other founders when they are growing. They are opening new legal entities in the country, hiring people, creating new jobs,” said D'Aveni.

How has the government of Georgia Meloni affected the ecoystsem: “The government is more aware of the importance of this ecosystem. For sure, there are things that we could do better, but I think we are going in the right direction. And also on the European level, I think there are very good signals. As you know, last year also the president of the European Commission was on stag,e and she said very important things. Of course now we need to see if there will be actions, but for sure we are on a good trajectory also on that side, I think.”

She added that the 28th Regime initiative promoted by EU Inc “would be very important because, you know, opening a company in Italy is still kind of cumbersome. But more than that, when they want to expand to a new country in Europe, there's a lot of things they have to do. And I think this is one of the main drawbacks of building a company in Europe. Every country is kind of a different world. So we need to have better legislation and a better way to operate throughout Europe and not only in your initial market.”

You can watch or listen to the full interview below:

YouTube

Spotify

Apple 

From the Editor’s desk:

• Please join for our upcoming Mixer featuring a VC panel — including breakout roundtables and networking drinks — Thursday, Feb 26, London “AI, Power & the Road to 2026”

• Please take our SHORT, three-question, reader survey here.

• Any other thoughts, comments, criticism: contact@pathfounders.com

Sign up to be the first to hear about the launch of our member network here.  Make sure you’re a subscriber to this newsletter and follow all our channels below. To partner with the site, newsletter, and/or events, contact our head of partnerships: Ana Williams, ana@pathfounders.com

Like your news and updates from us a little faster? Get our Pathfounders WhatsApp Announcements (appears in your main WA feed) OR get the Pathfounders Whatsapp Public channel.

The Latest on Pathfounders:

• JOIN our first mixer: “AI, Power & the Road to 2026” featuring a VC panel — including breakout roundtables and networking drinks — Thursday, Feb 26, London

Get tickets here.

Featuring:

• Sitar Teli, Managing Partner, Connect Ventures

• Alex van Someren, recently exited as the UK's Chief Scientific Adviser for National Security and former VC 

• ⁠Manjari Chandran-Ramesh, Amadeus Capital Partners, Partner in the Early Stage Fund focusing on AI, Robotics and Quantum

• ⁠ Ana Barjasic, Connectology, European Innovation Council 

We’ll explore the major forces shaping tech and venture in 2026, including:

​• AI’s next phase
​• What’s next for Deeptech
​• Geopolitics and the future of sovereign technologies

​Agenda:
​6:00pm – Arrival & Networking Reception
6:30pm – Panel Discussion
7:15pm – Pathfounders Speed Tables
8:00pm – Open Networking

​Thank you to our Partners:

Orrick UK 
Nebius

The latest on Pathfounders:

Quantonation raises the largest ever quantum fund
The Paris and New York City-based venture capital firm is backed by Novo Nordisk, Toshiba and Vertex Holdings
by Amelia Isaacs

Pathfounders Interview: Pierre Festal on Kembara's billion euro growth fund
Spain-based Mundi Ventures recently launched the Kembara Venture Fund to address this with a €750 million first close, and the ultimate target of a billion euros. Partner Pierre Festal joined Pathfounders to unpack what this means for Europe. 
• Watch on YouTube
• Listen on Spotify 

Pathfounders Interview: EdTech investment is changing fast
‘Edtech’ is morphing into ‘Learning & Work’ according to a new report from Brighteye Ventures. Pathfounders caught up with report author Rhys Spence to unpack where VC investors are putting their money in Europe.
• Watch on YouTube
• Listen on Spotify 

Pathfounders interview: Philip Bahoshy, CEO of Magnitt, on the AI investment boom in MENA
Venture capital in the MENA region is moving decisively toward AI-native platforms, which captured 69% of AI funding. The UAE and Saudi Arabia jointly attracted 87% of all AI capital deployed. Pathfounders caught up with  Philip Bahoshy, CEO of Magnitt, a key research house in MENA, to discuss how investment AI is shaping the region's tech industry.
• Watch on YouTube
• Listen on Spotify 

Who is David Silver?

Pathfounders did an AI-assisted deep dive into the man of the week, David Silver, who left Google DeepMind late last year to launch a London-based start-up, Ineffable Intelligence.

This week, the startup raised raising $1bn at a reported $4bn pre-money valuation, led by Sequoia, with big tech also in discussions to join the round. (Financial Times).

If you want the short version, it’s this: Just creating bigger chatbots via LLMs is a dead end. His idea is to create AI agents that can create experiences and environments in which they themselves can learn more than what human knowledge can teach them.

So Ineffable Intelligence will be a lab built around Reinforced Learning as the scaling paradigm, aiming at “superhuman intelligence” as an explicit target.

If he’s right, whoever cracks scalable “experience” loops gets a compounding platform that isn’t bottlenecked by human knowledge. Hence, the investor appetite. 

What we know: 

• David Silver (b. 1976) is one of the best-known reinforcement learning (RL) researchers of his generation: He’s a DeepMind “original,” UCL professor, and a key figure behind AlphaGo, AlphaZero, and AlphaStar—systems that hit superhuman performance in games via deep RL.

  Why is this important? 

• There’s been a capital stampede into “founder-scientist” AI labs aiming to go beyond today’s LLM paradigm, especially by building agents that learn by doing, not just by reading.

What is his core thesis?

Silver’s big idea (formalised with Richard Sutton) is that AI is shifting from systems trained mainly on human-produced text/images toward systems that improve mainly through experience:

• We are now in the era of Human Data: models absorb “everything humans wrote down,” then get nudged by human preference. Powerful, but bounded by human knowledge. (Here’s his paper on the idea)

• We are entering the era of experience: “a new generation of agents” will become superhuman by learning predominantly from experience—i.e., by interacting with environments, generating data, and iterating. Experience will dwarf the scale of human data used today. 

He claims AlphaZero has already demonstrated this, the lesson being: “remove the human ceiling”

What’s his story?

• AlphaGo (2016) used human expert games to get started, then reinforcement learning to improve.

• AlphaGo Zero / AlphaZero showed something more radical: human data wasn’t necessary and could even be constraining; self-play + RL could blast past the best human play and keep improving. (This is where he leans on the “bitter lesson” framing.) (Deepmind Interview)

• He treats “Move 37” as the symbolic moment: “One of the biggest moments of the Alph Go story was move 37 that everyone always references. Move 37 was a move that happened in the second game of Alph Go against Lisa Dal. Alph Go played a move that defied everyone's expectations. an “alien” move the human community didn’t expect—evidence that experience-driven systems can produce novelty that isn’t just “average human internet text.”

What he thinks is missing in LLMs:

Silver’s critique is not that LLMs are useless, but they are based too much on us:

• LLMs + RLHF (Reinforcement learning from human feedback) are great at human-aligned fluency and usefulness, but anchored to human judgments—so they struggle to reliably discover strategies humans wouldn’t recognise as good ahead of time.

• He draws a sharp distinction between:

  • Ungrounded preference (a rater saying “this cake recipe looks good”),

  • versus grounded feedback (someone actually bakes/eats it and the outcome is good/bad).
    He argues the latter is what unlocks open-ended improvement and genuine discovery.

In short: human feedback provides scaffolding but not a jumping off point to go somewhere new.

He thinks AI agents are the commercial bridge:

• Silver/Sutton explicitly bet on agents learning from interaction as the route to “superhuman capabilities.”

• Ineffable is there positioned as building on Silver’s RL work, training systems through interaction with environments, not only static text. 

• This matches the broader industry thesis: agents are how you operationalise intelligence into workflows (planning, acting, iterating), not just chat. 

His view of creativity and “beauty” as pertains to AI:

• Creativity is not mystical; it’s what happens when a system performs massive trial-and-error and accumulates “a million mini discoveries.”

• He’s motivated not just by utility, but by the idea that intelligence can produce beauty: the “sense of beauty” Go players reported when AlphaGo revealed new possibilities. 

Safety, risk, and alignment (as he frames it):

• Silver is concerned about the unintended consequences of humans (climate, pathogens, etc.) and sees AI as potentially a tool to avert disasters—if regulations prohibit unacceptable uses.

• In the “experience era” framing, he acknowledges serious risks in “untethering” systems from human data and says the transition needs to be taken seriously (part of why he wrote the position paper at all).

• He also implicitly runs into the classic specification/metric problem (“paperclips,” tyranny of metrics): he argues real-world environments contain many signals and that systems could adapt goal proxies over time, with humans as part of the environment, but he does not claim alignment is solved. (Alignment Forum)

What Ineffable Intelligence likely represents:

Based on news reporting and his published thesis, Ineffable looks like a very specific bet:

• Post-LLM frontier: not “bigger chatbots,” but experience-generating, environment-interacting agents.

• A lab built around RL as the scaling paradigm (“sustainable fuel” vs mining human data), aiming at “superhuman intelligence” as an explicit target.

• If he’s right, whoever cracks scalable “experience” loops (in sims, tools, code, robotics, science) gets a new compounding curve that isn’t bottlenecked by the web. Hence, the investor appetite. 

Op-Eds and Contributed Columns

We’re ready to start receiving pitches for Op-Ed columns. These can be on tech, startups, investing, AI, the future, frankly, you name it. So long as it tech starup or VC-related and bloody interesting.

The rules are: A short paragraph outlining the piece. Followed by three or four bullet points about the main issues you’d cover. The argument must be backed up by third-party references. We’ll only take pieces we green-light. Nothing submitted ‘on-spec’. We will build a form for these submissions, but in the meantime, you can pitch us at contact@pathfounders.com. Pitching us more generally? Here’s all you ned to know about that subject. 

Hot Tips and News 

Please send all tip-offs, suggestions, and news to contact@pathfounders.com

SUBSCRIBE TO THIS NEWSLETTER

That’s all, folks

As always, you can interact with Pathfounders via the email contact@pathfounders.com and, even, via an anonymous form if you want to tell us something you think is important, but would prefer to keep some distance. You can even subscribe to Pathfounders WhatsApp Announcements (appears in your main WA feed) or the Pathfounders Whatsapp Public channel.

Follow Pathfounders:

Pathfounders WhatsApp Announcements (appears in your main WA feed)

Pathfounders Whatsapp Public channel

We are not on Substack (sorry), but we will publish headlines from here.

LinkedIn

X (Formerly Twitter)

BlueSky

Telegram

Facebook

Instagram

Luma (events)

RSS Feed for all posts

Podcast: Spotify

Spotify RSS

Async

Async RSS feed

YouTube

TikTok

Contact Pathfounders Anonymously

Olivier Tonneau, Will Zeng and Christophe Jurczak / Quantonation

Investment into quantum technology is surging, yet there are few VC funds solely focused on the sector.

Quantum computation, communication and sensing combined could generate up to $97bn in revenue worldwide over the next decade and Quantonation is set to help fuel that with the €220m close of its second fund.

The Paris-based early-stage VC fund investing in deep physics and quantum technologies has raised the largest ever quantum fund backed by Novo Nordisk, Toshiba and Vertex/Temasek, among others.

The European Investment Fund (EIF) also invested €30m into the fund, which is set to expand into advanced materials, sensing and next-generation computing.

“Quantum has spent decades being described as five years away,” said Quantonation co-founder and managing partner Christophe Jurczak. “That wasn’t a failure of physics, but of ecosystems. What’s changed is alignment: hardware, software, supply chains, and industrial demand. Quantum is no longer a race to build one machine. It’s an interlocking stack, and that’s where durable value now sits.”

Closing its second fund positions Quantonation as the largest dedicated quantum investment firm globally, by AUM. While generalist VC funds are increasingly looking into quantum companies, only a few are dedicated to them entirely.

The second largest quantum fund is 55North, which launched last October with the first close — €134m — of what is set to be a €300m fund.

The Denmark-headquartered VC has backed Finnish startup IQM as part of its €275m Series B and German Kiutra in its €10m Series A.

In London, “quantum-first” VC firm Firgun Ventures launched with its first close of $70m back in November. It aims to raise $250m. Similarly to Quantonation, it focuses on early-stage startups, and this week made its first investments into Los Angeles-based Quantum Elements and Canadian startup Photonic.

Quantonation has invested in companies across quantum computing, networking, sensing, materials, adjacent supply chain and broader deep-physics domains since launching in 2018.

To date, the first fund has invested in 27 companies worldwide, including Nord Qauntique, Multiverse Computing and Pasqal, which doubled down on its presence in Korea with a $52m investment last October, as Pathfounders exclusively revealed.

The second fund has invested in 12 companies so far, with expectations of building out a portfolio of around 25.

“With Quantonation I, our mission was to back the pioneers. These were the teams turning quantum physics into working machines, and to help the first leaders emerge,” Olivier Tonneau, Quantonation co-founder and partner, said. “With Quantonation II, the focus shifts to utility and scale: demonstrating practical advantage, building robust products, and preparing for industrialisation.”

Company info:

• Headquartered: Paris, France and NYC, USA

• Co-founders: Olivier Tonneau andChristophe Jurczak

• Recent investments include: Diraq (quantum computing), Pioniq (advanced materials), Chiral Nano (nanomaterials), Qblox (supply chain) and Project11 (security)

Technology development and defense, and of course war, are inseparable. That means tech matters will be in the background of nearly every discussion at the Munich Security Conference, which begins today. Given that the conference’s chairman, Wolfgang Ischinger, wrote that we live in a time of “wrecking ball politics” in a foreword to an MSC report, we can expect the conference to be dramatic. And this means pivotal, if not world-changing tech matters could be in play.

AI is pretty much the leading candidate for world-changing tech right now. And perhaps the leading AI topic is about AI sovereignty—as in, a nation (or supranational body like the EU) developing and having access to its own cutting-edge AI system. In terms of geopolitics AI is in focus because many of the cutting-edge innovations are happening in the U.S. and China. Historically, from a European point of view, this might have been tolerable, but as Ischinger said, the “recalibration of US foreign policy” under the Trump administration has cast doubt on the reliability of the U.S. as an ally…hence the interest in sovereign AI developments, and a focus on regional data and infrastructure security.

France has been a leader in this direction. This week, French naval defence design outfit Naval Group formed a joint venture with defence company cortAIx, the AI accelerator from France’s Thales Group, to help “accelerate the industrialisation of AI solutions applied to defence systems.” It’s all part of an effort to develop sovereign AI solutions for the defence industry in France. 

But France’s push to move out of the United States’ tech shadow goes much further than this: in February last year the nation announced a €109 billion AI infrastructure investment, which some commenters suggested was the most “ambitious” sovereign AI program outside the U.S. and China. And just a few weeks ago the French government announced it was barring official use of U.S. software platforms like Zoom and Microsoft’s Teams and replacing them with its own, locally-developed alternatives. According to David Amiel, minister for the civil service and state reform, the move is to “guarantee the security and confidentiality of public electronic communication.” We can imagine that AI and vibe coding, perhaps from notable French startup Mistral, might be used to help code the French Zoom alternative. 

Industry watchers wondered, of course, if other non-U.S. nations might follow suit, including its neighbor Canada. A report in the Independent even reminded us that Austrian soldiers have adopted open-source software as an alternative to U.S.-sourced services from Microsoft and Google, and German politicians have switched to using free alternatives to carry out administrative duties. 

As a side note, you’ve got to wonder if the sudden rise of AI-coded solutions may threaten the entire global software-as-a-service model (dominated by U.S. companies) with AI stepping in to craft custom solutions for customers on the fly. And if this happens, will some of these new-gen AI-SaaS solutions be Europe-based?

There are thousands upon thousands of defence tech topics that will surely surface in the Munich debates, but we can be certain that plenty will center on reliance on actual defence hardware manufactured in the U.S. 

The impressive, futuristic, but expensive F-35 stealthy strike fighter aircraft may well come up. The U.S. has been attempting to sell these “fifth generation” aircraft, which cost up to $100 million per vehicle, to overseas partners—contracts exist with eight NATO nations and five non-NATO partners. But the program has been hit with cost overruns and technical snags (last week it emerged that manufacturer Lockheed Martin was delivering newer F35s to the USAF with dummy weights in the nose in place of a cutting edge, but very delayed next-gen radar system). Meanwhile, overseas F35s will remain tied to Lockheed Martin due to the complexity of their hardware and software design, raising concerns that a fractious U.S. government could implement a form of “kill switch” by simply withdrawing support. 

Canada’s government, which has a decades-old deal (the North American Aerospace Defense Command) with the U.S. to buy 88 of the jets, has recently been publicly considering an alternative. A few weeks ago Sweden’s Saab proposed Canada buy 72 Gripen jets and eight long-rance early warning GlobalEyes drones as an alternative. Gripens are considered advanced fourth, or fourth-and-a.half generation fighters, and cost about as much as F-35s, but the design is long-established, and their per-hour running costs may be about a quarter of the American jets’. CNBC news said the Saab deal would guarantee some 12,600 jobs in Canada.

Industry watchers will be interested to see drones mentioned here, given the dramatic deep tech innovations happening in real time as Ukraine uses drones in ever-more innovative ways against invading Russian forces. It’s timely, too, that Ukraine may soon receive up to 150 Gripen jets of its own (alongside 100 French-made Rafale jets).

Interestingly, Canadian PM Mark Carney and Foreign Minister Anita Anand are expected to take part in the Munich conference. Meanwhile, the U.S. postured in reaction to Canada’s possible Saab deal by threatening, in effect, to intervene in defence matters in Canadian airspace more. The U.S. ambassador to Canada, Pete Hoekstra, warned that “NORAD would have to be altered,” if Canada opted for the Saab jets which are, he said, not as “interchangeable, interoperable as what the F-35 is.”

Of course AI, jets, and drones are just the posterboys for a lot of tech-centric issues that will be lurking in the background at Munich. Pick a hot topic like data centers in space, the overall space industry, sovereign data control, health tech, and pretty much anything else you can think of, and it may play a part.

All of this boils down to one conclusion: Europe, and other non-U.S. allies, may be attempting a “big unplug” from U.S. defence giants and from other U.S. technology providers too. The moves could have trillion-dollar-scale, long-term implications.

And what’s driving all this fuss? A single person. 

From the Editor’s desk:

The important bits first:

• Please apply for our upcoming Mixer featuring a VC panel — including breakout roundtables and networking drinks — Thursday, Feb 26, London “AI, Power & the Road to 2026”

• Please take our SHORT, two-question, reader survey here.

• Any other thoughts, comments, criticism: contact@pathfounders.com

Pathfounders will run a programme of curated events this year, alongside a membership network for founders and investors. Sign up to be the first to hear about it here.  Make sure you’re a subscriber to this newsletter and follow all our channels below. To partner with the site, newsletter, and/or events, contact our head of partnerships: Ana Williams, Ana[@]pathfounders.com 

Like your news and updates from us a little faster? Get our Pathfounders WhatsApp Announcements (appears in your main WA feed) OR get the Pathfounders Whatsapp Public channel.

Upcoming Events

• Catch Mike Butcher, Founder & Editor of Pathfounders, at: Mobile World Congress and subscribe to our Luma to be alerted to our upcoming events. He’s on mike@pathfounders.com

The Latest on Pathfounders:

Pitch Pathfounders
In our new feature, Pitch Pathfounders, we spoke to first-time founder Aaron Devitt, a 23-year-old Dublin-based founder who dropped out of college to build an AI-native startup he’s convinced will supercharge a hard-to-crack problem in proptech.

Exclusive: Former MI5 engineer secures £2m to keep agentic AI in check 
Overmind is building the supervision layer for autonomous AI agents in high-risk industries.
By Amelia Isaacs

Occam raises €3m to bring autonomous drones to Ukraine’s frontline 
Backed by European defence investors, the startup is betting the next phase of the war will be decided not by the number of men but by technology, innovation, and autonomous systems.
By Amelia Isaacs

Can Multiverse’s new acquisition give it the growth it needs?
Founder and CEO Euan Blair tells the Pathfounders Podcast what his next strategic moves will be, and why the US didn’t work out as a potential market.
By Mike Butcher

Cities are the key to ClimateTech
As 2150 closes its €210m Fund II, co-founder Christian Hernandez tells the Pathfounders Podcast that cities are the answer to decarbonisation.
By Mike Butcher 

Mike’s Drop

Our first event!

First of all, please apply for our first-ever curated Pathfounders Mixer featuring a panel discussion, bringing together founders, investors, policymakers, and ecosystem operators for a sharp on-stage conversation, followed by structured, high-signal networking. ​We’ll explore the major forces shaping tech and venture in 2026, including: ​Geopolitics and capital flows; ​AI’s next phase of growth; and what’s next for strategic technologies across the UK and Europe.

The first speakers to be announced are (more to come):

• Sitar Teli, Managing Partner, Connect Ventures
• Alex van Someren, recently exited as the UK's Chief Scientific Adviser for National Security and former Managing Partner at Amadeus Capital Partners
• More to come!

2026: The year Europe grows up?

• And can Europe create its own social media? 

In attempted number 197 (it might as well be at this point), a startup called “W” launched at Davos, attempting to become a sovereign social platform for Europe. W forks the BlueSky protocol ATT, but you can only join a waitlist for now. 

It's backed by an advisory board, former ministers, and businesses, most of which are based in Sweden. The two Vs in W supposedly stand for "values" and "verified."

W is designed to remain under European ownership. (Let’s also note that Mastodon is moving to a European nonprofit structure after initially operating as nonprofit in Germany.)

‘W’ CEO, Anna Zeiter, who recently left her job as chief privacy officer at eBay said there’s an “urgent need” for a new social media platform built, governed, and hosted in Europe.

We’ll see what the product is like before we comment further. 

Davos Redux: Is Sovereignty the answer to would-be Kings?

The technology industry is in a state of heightened awareness at the start of 2026. The moniker for Davos was “the spirit of dialogue.” But as I pointed out in my video piece short from there, the actions and stance of the Trump administration created something much closer to a “spirit of destabilisation.” This meant that, as one founder who specialised in sovereign compute put it to me over a dinner, “Trump just made me a billionaire.”

Emerging from the frenetic Davos/WEF period, European entrepreneurs and investors were ‘cock-a-hoop’ that the ‘EU Inc’ project to create a Delaware style company structure for the whole of the EU had been mentioned on stage by Ursula von Der Leyen at the WEF. Only to then hit reality when people started asking how long it would take…. 

The brief Davos download: 

VCs were unusually present during the Davos gang-bang. Lightspeed had an entire ‘house’ - dubbed (tres originally, ‘Lighthouse’), while First Minute Capital’s Brent Hoberman and Balderton Capital’s Bernard Liautaud were glad-handing all and sundry at various events and receptions. This rarely used to happen at Davos (and I’ve been going for a number of years). The main constituent population at Davos is bankers and financiers, NGOs, governments, Crypto kids and Nepo-babies. Rarely VCs and startup founders, whose budgets are generally leaner. What was attracting them? Well, it’s pretty simple: The world has become nightmarishly more febrile since Trump returned, so the only real way to get a sense of what is actually going the f*ck on was to get in amongst the Davos hordes and suck up contacts, and the atmosphere.

Another indication that Davos/WEF will be in everyone’s diary for next year was the sheer amount of AI companies present - its effect on jobs and economies will no doubt draw them back for years to come, Trump or no Trump. 

Palantir’s Karp said humanities graduates are doomed, but BlackRock COO Robert Goldstein said it was recruiting graduates who studied “things that have nothing to do with finance or technology,” and McKinsey’s global managing partner said the company is “looking more at liberal arts majors, whom we had deprioritized, as potential sources of creativity.”

The CEO of Google’s DeepMind Technologies was more optimistic. Hassibis said internships would likely be replaced by interns learning AI tools instead. However, the job market would be in “uncharted territory” after the arrival of AGI.

Jensen Huang, Nvidia CEO didn’t denying the existence of an AI bubble, but said it was a misreading of the needed investment into AI infrastructure, such as compute. Larry Fink, CEO of Blackrock, agreed with him. Huang added that Europe was well placed for this moment.

Microsoft CEO Nadella said AI had to grow beyond tech otherwise the bubble would burst. 

Kristalina Georgieva, head of the International Monetary Fund, said AI would be a “tsunami hitting the labour market”, with young people worst affected.

Elon Musk was, as usual, full of… superlative comments: "We might have AI smarter than any human by the end of this year,” he claimed, not unexpectedly. We’d also all have robots and “abundance for all" later on. Of course, none of his robots are actually working in his own factories yet. 

Yann LeCun just dumped on LLMs as usual.

In a spooky prediction of the Moltbot craze, which took off with vengeance in the last couple of weeks, Canadian computer scientist and one of the so-called 'God-fathers of AI', Yoshua Bengio, warned at Davos that today's systems are trained too closely to be like humans. Well, guess what, that’s almost exactly how these Moltbots were behaving on Moltbook. 

Also at Davos, Dario Amodei, Anthropic CEO, said humans could lose control of AI and “not selling chips to China is one of the biggest things we can do to make sure we have time to handle this.” Guess what happened the week after? China gave the green light for Chinese AI companies to buy Nvidia H200 AI chips. So that prediction lasted a matter of days.

And pretty much everyone was excited about Agentic AI.Pitch Pathfounders

Pitch your startup to Pathfounders. Fill out this form, and you may get featured.

Op-Eds and Contributed Columns

We’re ready to start receiving pitches for Op-Ed columns. These can be on tech, startups, investing, AI, the future, frankly, you name it. So long as it tech starup or VC-related and bloody interesting.

The rules are: A short paragraph outlining the piece. Followed by three or four bullet points about the main issues you’d cover. The argument must be backed up by third-party references. We’ll only take pieces we green-light. Nothing submitted ‘on-spec’. We will build a form for these submissions, but in the meantime, you can pitch us at contact@pathfounders.com. Pitching us more generally? Here’s all you ned to know about that subject. 

Hot Tips and News 

Please send all tip-offs, suggestions, and news to contact@pathfounders.com

SUBSCRIBE TO THIS NEWSLETTER

That’s all, folks

As always, you can interact with Pathfounders via the email contact@pathfounders.com and, even, via an anonymous form if you want to tell us something you think is important, but would prefer to keep some distance. You can even subscribe to Pathfounders WhatsApp Announcements (appears in your main WA feed) or the Pathfounders Whatsapp Public channel.

Follow Pathfounders:

Pathfounders WhatsApp Announcements (appears in your main WA feed)

Pathfounders Whatsapp Public channel

We are not on Substack (sorry), but we will publish headlines from here.

LinkedIn

X (Formerly Twitter)

BlueSky

Telegram

Facebook

Instagram

Luma (events)

RSS Feed for all posts

Podcast: Spotify

Spotify RSS

Async

Async RSS feed

YouTube

TikTok

Contact Pathfounders Anonymously

Overmind co-founders Sam Brunt, Akhat Rakishev and Tyler Edwards

As autonomous AI agents begin to move from experiments into real-world deployment, a growing problem is emerging: once these systems are live, it’s increasingly difficult to understand what they’re doing, why they’re doing it and whether they can be trusted to behave safely in high-stakes environments.

It’s not just personal assistants like OpenClaw (/Clawdbot/Moltbot), agents are rolling out across verticals, and in highly regulated industries like legal, defence, healthtech and fintech, this can pose operational and security risks.

This is the space where London-based Overmind positions itself, providing a “supervision layer” for agentic AI.

The startup, which launched just seven months ago, raised a £2m seed round led by specialist cybersecurity investors, Osney Capital, with participation from 14Peaks, Portfolio Ventures, Antler and Endurance Ventures.

Overmind, self-described as “supervision for super intelligence”, is taking a different approach to monitoring agentic AI. While traditional approaches to AI security were built for deterministic software, agentic systems are inherently non-deterministic by design — they adapt, learn from new data, update their memory and change behaviour constantly — which, according to co-founder and CEO Tyler Edwards, requires new, adaptable layers of protection.

“It's about moving security from these kind of static fire-and-forget tools to something that is constantly running, constantly learning, constantly understanding,” Edwards tells Pathfounders. “Because this is software that isn't static. It changes, it adapts, it learns new contexts, updates its memory. It's not the same as you found it yesterday.”

Because there are so many potential risks, it is impossible to prepare an agent for all of them, but it is possible to assess what it is doing when it’s in production, Edwards explains. Just like with a human, while you can’t prepare for everything that might possibly go wrong, you can put fail safes in places to prevent critical errors.

Edwards, who spent eight years building AI systems for British intelligence agencies including MI5, MI6 and GCHQ, has co-founded Overmind alongside CTO Akhat Rakishev and CRO Sam Brunt.

Rakishev previously led machine learning infrastructure at Monzo and Lyst, while Brunt has scaled go-to-market at unicorns Funding Circle, Pipe and Vertice.

Having traded in the stress of building AI systems for British intelligence agencies — “You come in one day and they say, we've got intelligence that within the next week Russia is going to invade Ukraine. Here's what we need to do before that happens” — for, as Edwards puts it, “a slightly different flavour” of stress in the form of a startup, his background in defence has shaped the company’s focus on solving problems for complex, high-stakes industries, but without the constraints of civil service.

Entering a rapidly emerging market, companies like AgentOps.ai (the creators Agency AI raised $2.6m in 2024) are targeting agent observability and testing, while Zenity, an agent-focused security startup, raised $38m last year.

Overmind specialises in AI models for specific tasks and sectors, enabling fine-tuning down to the individual company level, dependent on the data available. The startup plays in a space where security is key and generalist models might not cut it.

“Say for example, you're building in medicine,” Edwards says. “Your model doesn't need to know the full works of Shakespeare. Maybe it just needs to know specifically what a melanoma is and how to diagnose it.”

The AI can learn the specific tasks and teams can ship smaller, more specialised models that are cheaper, faster and easier for the teams to build themselves.

“In the new frontier of autonomous AI, agent security, performance, and execution are the ultimate competitive advantages,” Osney Capital partner Adam Cragg said in a statement. “Overmind provides businesses with truly differentiated technology that monitors and secures agentic AI while iteratively improving model performance, enabling teams to scale with confidence.”

The developer-focused software development kit (SDK), integrates into existing codebases and is designed to enable AI and agents to learn from production data — it can learn from its surroundings and specialise in specific tasks. 

“We kind of just want to get out of the way,” Edwards says. “Here's the tech. Go crazy and we'll get out of the way. But here are the tools for you to just do really incredible things.”

The team is also building a telemetry stack to improve reinforcement learning by collecting more and higher quality data and the fundraise will go towards technical hires for further product development.

Company information:

• Six staff, with plans to double headcount by the end of the year

• Based in London

Occam’s journey started with a prototype in a London park where Guiam Wainwright walked his big, white, furry samoyed. Except at the end of the lead in his hand wasn’t his dog walking, but a drone flying in the sky.

Guiam Wainwright was testing an early version of what is now a fully autonomous drone defence system. Back then, each time the drone’s primitive AI spotted a trigger — a passing dog walker, for instance — it would lock on, try to chase after it and be reeled back in. Within months, versions of this system would be assessed for use on the Ukrainian frontline.

All of this testing was happening before Wainwright had legally incorporated Occam as a company, or even settled on its original name. The startup was first called Cakeshop AI, a title he later abandoned after discovering it provoked a Marmite-like reaction among army generals on the frontline — and didn’t quite sit alongside other military-grade tools.

In addition to working to the deadline of a war, Wainwright, now co-founder and CEO of Occam, had his own smaller, self-imposed version — or, as he jokes, one imposed by his wife — in the form of his newborn daughter.

The idea for the startup emerged around ten months ago, when Wainwright was about 30 days into paternity leave. “I got a bit bored,” he says. Conversations with his wife’s cousin — a former Director General of MI6 — soon turned to Ukraine, and specifically to drone commanders operating on the frontline.

What began as a way to fill time while helping a good cause quickly evolved into a full-scale effort to automate cheap drone warfare, tackling what Wainwright describes as the pilot-to-drone mapping problem.

By that point, drones had already become the defining weapon of the conflict. Ukrainian forces were increasingly reliant on them to compensate for shortages in manpower and conventional firepower — but the systems available to them did not scale.

“The problem the Ukrainians laid out was essentially that the drone program was keeping them in the fight. But the drone program had limitations due to the technology they had available to them,” Wainwright says. “The reality of it is you need to put mass in the air to have a wall. If you can only put a couple of drones up in the air, you’ve got a mesh. It's not like an impenetrable barrier that can hold anyone back.”

Integration testing

The scale of the challenge is stark. Last year, approximately 54,700 drone attacks were recorded in Ukraine, compared with about 11,000 in 2024. Drones accounted for roughly 96% of all aerial weapons deployed — up from around 30% at the start of the air war in 2022.

That context has shaped Occam’s rapid trajectory. The company’s system has now completed assessment as fit for integration testing by Brave1, the Ukrainian defence tech cluster.

The Brave1 field test

“AI is one of our key priorities, as it is fundamentally transforming the nature of combat operations,”  says Andrii Hrytseniuk, CEO at Brave1. “We highly value the fact that British technology companies like Occam are working directly with Ukraine to respond to real frontline needs — an experience that strengthens not only Ukraine, but Europe as a whole.”

Alongside this validation, Occam has raised a €3 million (£2.6m) pre-seed round led by Presto Tech Horizons, a resilience-focused VC fund backed by defence-industrial partner CSG (Czechoslovak Group), with participation from Antler, Freedom Fund, TYR.vc and a group of experienced defence and security industry angels.

What began as a paternity-leave side project has rapidly become something far bigger.

Day one at Antler

Wainwright co-founded Occam alongside COO Daniil Bash, whom he met on day one of Antler. Wainwright arrived with a single black slide bearing white text: ‘War is here’. It followed a stream of more familiar fintech and crypto pitches, making it a somewhat polarising introduction. Bash was wearing a t-shirt reading ‘slava Ukraine’ and stood up and applauded Wainwright’s pitch.

“For the rest of the cohort, I was the crazy guy trying to kill people with AI.”

CTO Aniketh Ramesh joined after meeting Wainwright at a defence hackathon. Wainwright himself brings a background spanning startups and machine learning, starting off his career supporting NATO troops with AI and ML, founding multiple startups and helping lead AI and robotics at Ocado.

Now, working in Ukraine, “it is simultaneously amazing and awful to see such a large number of people come together to collectively defend each other,” he says.

Until now, there had never been a real need for this kind of technology. Military systems were typically cloud-based and required access to high levels of compute and expensive hardware. Ukraine has forced a shift: operations must run at mass, cheaply and autonomously.

An automated frontline

With drones now inflicting around 90% of battlefield damage, an automated frontline increasingly means automated drones. Occam’s approach involves installing its software onto existing Ukrainian drones, paired with a low-cost compute unit that operates autonomously — without relying on the cloud or GPS.

Previously, one pilot would operate one drone and would have to be within a roughly 20-mile radius and there were high failure rates caused by GPS spoofing and signal jamming. Occam’s answer is autonomy on minimal hardware.

The company has engineered a sequence of algorithms capable of running on a Raspberry Pi Zero — an ultra-compact computer smaller than a credit card and costing just £10–15. The board is mounted on the drone, connected to the camera and flight controller, and given a mission: where to go and what to target. From there, it operates independently.

Close-up of Occam technology on a drone at Brave1 field test

Safety measures are built in, including battery constraints to prevent drones leaving designated areas, alongside direct Ukrainian military oversight. Occam is now at the integration stage, with Brave1 pursuing collaboration between the startup and select Ukrainian manufacturers to bring the technology to market.

The team itself is spread across integration engineers — many ex-infantry — in Ukraine, and core technologists based in Tallinn and London.

“We're going to have to win this based on tech and economies, and so if we invest in automation, if we can create battlefield systems where the brunt of all the fighting is held by automated systems, we can send robots to die,” Wainwright says. “And that means that our casualty count drops really low.”

Occam is now scaling its Ukrainian operations, with the ambition to be deployed across every drone in the fleet this year, continually adapting to Russian tactics. The company is also running a small number of paid adoption projects with European defence primes — which Wainwright cannot yet name — that will “fly by the end of the year”. A second raise is planned for the end of the year to support further expansion.

Matej Luhovy, Partner at Presto Tech Horizons, says: “They deploy, learn, and iterate under conditions dictated by the front line. Gui has an exceptional ability to rally people around a clear mission, and the team executes on it with urgency and discipline. Combined with a software-only autonomy stack that works without GPS or external connectivity, this is a solution that is not just novel, but fundamentally different.”

Beyond Ukraine, Occam has strategic partnerships and collaborations already underway with European defence primes seeking to reinforce NATO capabilities.

“This war will not be won not by the weight of numbers but by the power of creativity and innovation. Trusting good people to come up with good ideas, and creating good systems to apply them,” says Sir Alex Younger, former Director General of MI6 and Occam advisor. “Occam is a great example. It will confer asymmetric advantage in a fight that we simply must win.” 

Guiam Wainwright, Occam

What will training and ‘upskilling’ look like in the future? That’s the puck that Multiverse, the UK-based startup, has been trying to skate to for a number of years. The arrival of Generative AI both as a tool and a skill that must be mastered by the modern worker has made that mission ever more urgent.

Last week, Multiverse acquired Berlin-based data and AI training provider StackFuel, a small startup that had raised only €1.5m in total, from a handful of seed investors. 

Multiverse now plans to use that acquisition as a wedge into the German market, and subsequently other European markets. 

In a wide-ranging conversation on this week’s podcast (available on YouTube and Spotify), I asked Founder and CEO Euan Blair what his next strategic moves would be, and why the US didn’t work out as a potential market for Multiverse's expansion a couple of years ago.

The deal brings StackFuel’s local market expertise and enterprise customer base, including Mercedes-Benz, IAV, and Telefónica, into Multiverse’s expanding European operations. 

StackFuel was founded by Leo Marose and Stefan Berntheisel, both of whom will now join the senior Multiverse leadership team. The company has previously reported a 92% programme completion rate for its courses, which is pretty impressive.

The acquisition comes amid a period of consolidation in European edtech, following a €500m round that pushed Brevo to unicorn status.

Against that backdrop, Multiverse is doubling down on workforce upskilling as governments and corporates desperately try to turn their investment in AI into productivity gains.

London-based Multiverse has so far raised more than $400m from investors including Index Ventures, Lightspeed Venture Partners, and General Catalyst, reaching a $1.7bn valuation in 2022 after a $220m funding round. 

However, the company has also faced financial pressures. Accounts filed at Companies House last year showed Multiverse had recorded a pre-tax loss of £63.3m for the year ending March 2025, up £2.6m year-on-year, despite revenues rising more than a third to just under £80m. The filings also show the company made 55 redundancies over the same period.

Blair told me the StackFuel deal works well for the company because Germany has a stated aim to become an “AI nation.”

“Germany is Europe's largest economy, yet one of its least digitally transformed. They've got an incredibly skilled workforce up to a point, but digital skills are something the government is really trying to encourage… For us to kind of get a foothold in the German market, take our AI platform and curriculum to ramp up in Germany was incredibly exciting,” he said.

“If you look at the German government, as part of the coalition agreement, it said that it wanted to become an AI nation. It will only do that by skilling huge volumes of people on the job, and that's where we want to help,” he added. 

Multiverse has been operating in Germany since 2025 and holds AZAV accreditation for its courses, but this new acquisition will no doubt be viewed as having the potential to boost Multiverse’s much-needed growth. 

Shlomie Liberow, Aisy

Cybersecurity companies have long measured their worth in big numbers. Dashboards boast millions of vulnerabilities discovered, thousands of alerts generated, and endless tickets created — proof, at least on the face of it, that the tools are working. The more problems surfaced, the better the product appears. For the teams on the receiving end, however, it often means the opposite: an ever-expanding to-do list with a lack of clarity on what to prioritise.

Datadog’s recent State of DevOps Report found that only 18% of critical vulnerabilities are actually worth prioritising, yet teams spend around 130 hours per week monitoring threats and more than 20 minutes of manual effort on each individual vulnerability. If you’re starting out with one million tickets, even cutting a queue by 90% could still leave tens of thousands of problems demanding attention.

Aisy, an AI-native threat prioritisation startup emerging from stealth today, is approaching the problem “backwards”. Instead of starting at the ticket level, it looks at business objectives and threat models first, figures out the big-picture problems, and uses context to assess what steps need to be taken from there.

The company is emerging from stealth with $2.3m in seed funding from Flying Fish Ventures, Osney Capital, and 6DC to tackle what founder and CEO Shlomie Liberow sees as cybersecurity’s volume trap. He says it all comes down to simplifying things and reducing security tool overwhelm.

“This is a very challenging area to get right, because you can't really get it wrong,” Liberow tells Pathfounders. “You can't get it right 50% of the time or even 80% of the time and be okay with that. And so the whole theme throughout is, how do we keep it really simple? It doesn't feel like there's a lot going on, but there's a lot of work that we have to do in the background to make it feel that way.”

This idea is clearly a core tenet of the company, even down to the design aesthetic — futuristic, calming landscapes that seem more likely to feature as the background to a meditation app than a cybersecurity platform.

“The focus, the branding, the vibe is all about bringing a bit more simplification to the world that’s usually lots of flashing lights and scary colours,” he adds, “but doing so in a confident way.”

With more than eight years of experience in the bug bounty space, Liberow is a self-taught hacker and bug bounty hunter who previously led hacker R&D at HackerOne. He left his role early last year to start working on Aisy, which was inspired by his own experience in the field, and the platform uses the same reconnaissance techniques bug bounty hunters would use for discovery.

“I realised that it's very difficult for people to operate when they don't have the headspace and the bandwidth to look at every single ticket. [...] So we're able to say, these are the three things you need to worry about. And this is why. And here's the story. And if you want to dig in deeper, we can do that.”

Differentiating in a crowded space

Rather than approaching security from a ticket level, which Liberow describes as “essentially whack-a-mole”, the platform asks users what problem they want to solve — that could be key products, customer journeys, sensitive systems — and then searches for anything that threatens those priorities. 

The startup is not trying to replace the many background scanners and security products companies already rely on. Its target is the layer where a user has to log in, interpret alerts and decide what matters. 

The company is entering an increasingly busy field. British startup Maze raised $25m Series A last June in a round led by Theory Ventures that added to a $6m seed round, while Cogent Security raised $11m last July and Zafran Security, slightly further along on its growth journey, secured $60m Series C funding last month.

“This space generally is crowded, no doubt,” Liberow says, “especially in the last two years, where you can just say anything and claim anything and do a good proof of concept. And so it was more about how we convey the background to what we're doing and why we're doing things differently.”

He began working on Aisy in early 2025 after a series of advances in machine learning made him realise that his bug bounty experience could be translated into software. Funding followed at the end of the year, and Liberow intentionally looked outside of the cyber space, hoping a more AI-focused fund would help extend access to more AI specialists, though they also have one cyber VC and one more traditional European firm for “access to old school markets”.

“Companies are facing an explosion in their attack surface while security teams are overwhelmed with data,” Flying Fish Ventures partner Vanessa Pegueros said in a statement. “Aisy’s attacker-driven approach transforms vulnerability management from reactive triage into proactive risk reduction. Aisy gives CISOs and other security leaders a far more accurate view of their real exposure with the context security leaders need to make better decisions and move faster.”

Aisy’s backers also include advisers and angels from DeepMind, Cisco, HP and Trail of Bits. Liberow plans to use the funds to double down on ML, bring in hacking experts as contractors, add three more people to his current team of three (Liberow, an ML specialist, and a core engineer,) and open a London office this quarter.

Owning the context layer

The security and vulnerability management market is projected to grow from $17.55bn in 2025 to $24.7bn by 2030, but Liberow believes success will depend less on new detection techniques than on scaling human expertise.

“We're not using AI for creativity — we're using it to scale the creative expertise that we already have,” he says. “The more niche the knowledge, the better, because that's where new companies can offer deep value alongside AI, by bringing deep domain understanding that off-the-shelf AI models can't match.

Recent breakthroughs in AI’s reasoning capabilities, context management, and compute costs will allow the startup to scale what has previously been incredibly time-consuming manual work.

“I feel like if we wait another year, who knows where things will go. I think this is the right combo of good enough capabilities to be able to start scaling up that knowledge and then tweaking it from there.”

Long term, Liberow describes this as owning the context above the tools.

“I want to own the context layer more than anything else,” he says. “Vulnerability management is a stepping stone to solving a problem. But what I really want to do is understand what matters to the organisation, who's involved, who's talking to who, why things are important, and when you do that you can unlock vulnerability management, but you can also unlock a lot of other things…eventually.”

Check out our podcast episode and video interview above

Climate tech VC 2150 has closed its second fund at €210m, lifting assets under management to €500m and reinforcing its bet that cities are actually the best place to apply carbon-fighting technologies.

In a statement, the London-based firm, the venture arm of urban-focused investor Urban Partners, said the fund has attracted a broadened LP base spanning Europe, Asia, and North America. Backers include Viessmann Generations Group, Novo Holdings, Denmark’s sovereign fund EIFO, and US-based Church Pension Group.

Fund II stays tightly focused on technologies reshaping cities and the heavy industries that power them, from industrial heat and cooling to circular economy and lower-carbon materials. The fund has already made seven investments, including electrified industrial heat specialist AtmosZero, refurbished electronics marketplace GetMobil, metals recycling platform Metycle, and direct air capture startup Mission Zero Technologies, alongside three unannounced deals.

Since launching its first fund in 2021, 2150 has built a 27-company portfolio generating more than $1bn in combined annual revenues, with climate impact measured at megatonne scale. Like its predecessor, Fund II is classified as an Article 9 vehicle. Article 9 in the Paris Climate Agreement covers investment products targeting sustainable investments and those that have sustainable investment as their objective.

Co-founder Christian Hernandez Gallardo told Pathfounders the strategy remains unchanged: backing cost-competitive technologies that make cities “more efficient, resilient and sustainable,” even if climate policy momentum has been hit by ructions from the US administration.

Fund I investments include home decarbonisation platform 1Komma5º, e-mobility leasing and battery-swapping network Vammo, and ultra-efficient cooling company Blue Frontier.

Faculty Founders: Marc, Angie and Andy

Since the rise of Generative AI, the consultancy sector has felt increasing heat coming from a tech industry encroaching upon its largely human-powered services industry, forcing incumbents to acquire technical depth before their own clients get wise to the whole thing. The solution, perhaps, was to acquire AI startups, and that is exactly what has just happened in the UK.

Accenture has agreed to acquire Faculty, one of the UK’s most prominent “AI-native” services firms, in a deal that underlines a deeper shift underway in the global consulting industry. Faculty had raised a total of £38 million since 2014, with investors including Apax Digital (at the series A) and Mercuri on the Secondary Market, according to CrunchBase and Dealroom. LocalGlobe, co-founded by Saul Klein, also backed Faculty since its seed round in 2016.

The transaction (with terms undisclosed and still subject to regulatory approval) will see Faculty’s 400-plus AI specialists folded into Accenture. Accenture was also, no doubt, attracted by Faculty’s Fellowship Programme, which converts PhDs and post-docs into industry AI operators.

As part of the deal, Faculty CEO and co-founder Marc Warner will become Accenture’s Chief Technology Officer and join its Global Management Committee, putting him at the heart of the firm’s decision-making process. 

Warner is a former Harvard quantum physics researcher and ex-member of the UK’s AI Council.

In a statement, Accenture CEO Julie Sweet described the acquisition as a way to “bring trusted, advanced AI to the heart of our clients’ businesses.” 

Faculty, founded in 2014, has built its reputation not on generic ‘AI strategy’ pitch decks, but on deploying applied machine learning in mission-critical environments. The UK National Health Service’s COVID-era Early Warning System, used daily by NHS Gold Command to predict hospital demand and allocate ICU capacity, was one such example.

Faculty’s core asset is Faculty Frontier, a decision-intelligence platform that combines data, AI models, simulation, and optimisation into a single system. Some industry observers have even described Faculty to me as ‘the UK’s Palantir’.

Accenture is already deploying it with clients such as Novartis to re-engineer clinical trial planning — an area where marginal efficiency gains can translate into hundreds of millions in value.

AI safety as a commercial moat

The move is also significant in that Faculty has positioned itself as a leader in “AI safety by design”, working with frontier model labs and institutions, including the UK’s AI Security Institute. In the consultancy industry, “safe AI” is crucial as both a compliance requirement and a sales differentiator, especially in the era of the EU AI Act.

By bringing AI capability in-house, Accenture may now have a crack at insulating itself from the problems presented by opaque US AI models, which could create legal and reputational risk, especially in the era of the Trump administration.

Rule 30 co-founders (Guy Conway, Felix Winckler and Damian Cristian) and Boardy co-founder and CEO Andrew D’Souza

It’s no secret that a lot of investors, like everyone else in the tech industry, are using AI to speed up workflows. From recording and transcribing meetings to researching markets and summarising diligence, tools like Fireflies, Granola and ChatGPT have quietly become part of the venture capital stack.

But some companies are pushing far beyond workflow automations and even having AI run their funds entirely.

Two not-so-traditional venture funds, Rule30 and Boardy, have both embedded AI so firmly in their deal making process that it’s not just advising humans — it’s shaping, and in some cases determining, where capital is deployed. 

Rule30, an AI research lab for venture that now has its own fund, describes itself as having a “fully algorithmic approach”. Investing in pre-seed and seed-stage startups across Europe and North America, typical check sizes range from $100k to $300k per investment and are decided entirely by the startup’s probabilistic decision engine that the co-founders, Guy Conway and Damien Cristian have spent the last five years building.

The two previously ran an accelerator, Zero 1, and were frustrated by the traditional process of startups getting financed by VCs.

“In our view it was way too much gut feel, way too much intuition by quite often very junior analysts who were gatekeeping a process,” Conway tells Pathfounders. 

“Personally, we feel like meeting people to judge personalities is a very bad way of doing it, unless you’re specifically trained in assessing personalities,” he adds. “And we’ve never met a VC who has done any sort of psychology or personality training.”

The pattern matching problem

Venture has long had a problem with “pattern matching”, that is, investors using patterns and experiences from the past to make decisions about new investments, and often using patterns and experiences from their past. They might be more inclined to back founders who look like them, went to their university or worked where they worked.

In 2013, Y Combinator co-founder Paul Graham was quoted in the New York Times saying “I can be tricked by anyone who looks like Mark Zuckerberg”. He was, he says, clearly joking, but there’s a reason the internet ran with the meme. 

A Founders Forum study found 78% of venture capitalists use “founder pattern matching” as a factor in their decision making. That, combined with the fact that 82% of venture deals come through warm introductions, creates a barrier for founders that fall outside existing demographic patterns. 

The promise of algorithmic investing is that data, not intuition, could flatten these barriers.

But if AI systems are trained on historical venture data — itself shaped by decades of bias — can they really produce fairer outcomes?

Building an algorithmic VC

Conway and Cristian have spent the last five years building data sets and data pipelines, scraping reams and reams of data from Pitchbook, Crunchbase, LinkedIn and Deel to build their own algorithmic or quant VC fund that uses data models and algorithms to make investment decisions. 

The startup saw its first close earlier this year and has made 12 deals so far. Backed by Isomer Capital as anchor LP, it is two thirds of the way towards its $15m target.

Originally the algorithm was identifying the top 10% of founders, but now it can identify the top 3% and top 0.1% of startups it looks at and then sort them into bands before backing them systematically.

Conway says they’re on the lookout for “outlier founders”: “the founders that are going to build businesses that are billions of dollars in value”, as he puts it.

In addition to opening up the decision engine to founders so they can enter their own information and see whether Rule30 would place them in their top 0.1% of founders, they analyse thousands of profiles of founders who are listed as being “in stealth mode” on LinkedIn to figure out who to invest in before anyone else.

“Our conviction is that research, data and compute will power the next generation of venture funds,” Cristian said in a statement. 

The AI analyses the founder’s experience, how that relates to the company they’re building, how their career “accelerates or deviate from the normal average”, at what point in time on a company’s journey they worked there — were they at Airbnb from day one or did they spend a month there as an intern last year — their educational background, their network and how it has evolved, the list goes on.

This is where the critical tension emerges: with the algorithm trying to pattern match based on the data it’s been trained on, will investments remain skewed to certain demographics based on the data even with the “gut feel” decision making process removed?

It’s hard to say without seeing the data, and Conway notes the algorithm doesn’t look at names, photos or other aspects of a founder’s identity, but it does look at education and geography, both factors that have historically correlated with demographic patterns in VC. 

Rule30’s process 

Rule30 looks at a few hundred profiles every week pulled from web scrapers, third party data and suggestions from angels. These are passed through models that create a prioritised list of targets, and it is only at this point that a human joins the loop — but not for long.

There are three reasons that a person enters the process at this stage.

One is that founders want to know the investors. Two is to better understand the fundraising plans of the founder. Three is to then plug the transcript of that half hour chat with the founder back into a model and refine their score to come up with a prioritised list of the best founders and deals and then to try to win allocation in the deal.

The goal is not for the team to talk to the founder and see if they personally think it would be a good investment.

“We have no human in the IC [Investment Committee]. The humans can’t say to do a deal even if the models say don’t do a deal. The human’s job is to execute the transaction,” Conway says.

“If you’re bringing it to an IC, what you find is they will follow the data points where it agrees with what people want to do, and they’ll discard those data points where it tells them they can’t do something that they’re trying to do. We’re quite adamant we will only deploy where the algorithm tells us to deploy.”

A more human-like approach 

At Boardy, there’s a bit more of a human in the loop, but it’s still largely up to “Boardy” himself, an AI “superconnector”, to make investing decisions.

The startup has been around for more than a year — it raised $3m pre-seed last October and $8m from Creandum just months later — but has only just developed from a straight up networking startup to now making its own investments as well.

Boardy is designed to be as much like a human as possible, but the most useful human you could possibly hope to come across. There’s no app, no login or dashboard, just a number on WhatsApp you can text or, ideally, call and talk to directly.

Voice based from the start, Boardy has somehow managed to get people having one hour phone call conversations in a time where people are less inclined to speak on the phone than ever before.

“Maybe it’s in some ways easier to talk to an AI on the phone because you don’t feel like you’re, you know, taking [time] away from somebody else,” Boardy co-founder and CEO Andrew D’Souza tells Pathfounders. “You don’t feel like you’re being judged as much.”

Boardy is designed to be the ultimate, zhuzhed up, non-judgemental board member (hence the name, but I enjoy that it feels a bit like a surfer too), connecting founders, investors and creators. 

It feels weird to attribute he/him pronouns to an AI that’s a suit with a box for a head, but that’s how D’Souza talks about him, and Boardy even sounds distinctly un-AI when you give him a ring. In fact, he sounds like a cheerful Aussie man.

As it turns out, this is reminiscent of D’Souza’s first boss at McKinsey in Toronto — something he didn’t realise until six months into building Boardy. “He was just a charming Australian guy and all the clients loved him,” he says.

After more than a year of connecting founders and investors and finding the best startups for funds to invest in, it seems obvious that the next step would be for Boardy to get involved in these deals. You can even speak to Boardy about it.

“You can actually call Boardy and talk to him about the fund and the plans,” D’Souza says. “He might hallucinate a few things, but I think for the most part he’s got most of the answers there.”

Boardy Ventures

When I spoke to D’Souza about a month ago, I asked if he’d thought about having Boardy make his own investments. A few weeks later and the startup has launched Boardy Ventures, a flexible fund targeting early-stage founders.

“We’ll probably have conversations with 50 to 100,000 founders over the next 12 months who are raising and even if we invest in the top 100 or few hundred, I think it’ll be a pretty good returning fund,” D’Souza says. 

Similarly to Rule30, Boardy will not be leading rounds and, in partnership with AngelList, will be writing checks around the $100k mark. The company says it has had more than 1,500 inbound requests to be LPs. To expand its reach beyond typical VC networks, Boardy Ventures is recruiting 1,000 “deal partners” (previously venture scouts) globally, offering them 50% carry to help identify founders outside the usual channels.

Where the startup is already talking to founders and connecting them to investors, it makes sense that it would be involved in the ones it thinks are the best. 

Once Boardy has a conversation with a founder looking to fundraise that he thinks the company should be a part of, he escalates that to the humans in the room.

Boardy

The plan is to put together a group of investing partners — a human investment committee that will probably be different for each startup Boardy considers investing in — and Boardy will go to a few humans in his network and ask for their opinions.

There will be a call, Boardy will listen in, form an opinion and while a person on the team will be the official decision maker, they’ll “go largely based on Boardy’s recommendation”, D’Souza says. 

Currently Boardy Ventures is not the main focus of the company — Boardy’s main goal is still to try to make the best connections, but sometimes that will be himself.

“I think one of the nice things is that, yes, Boardy can invest and he can put skin in the game,” D’Souza says. “But the primary objective is actually to introduce you to the best possible investor for you and then sort of co-invest with that best possible investor in some cases.” 

Separately, more than 700 firms have apparently reached out to have Boardy join as a venture partner — including Creandum, who led their last raise, but Boardy wouldn’t tell me who else was locked in when I called him to ask. Firms can hire Boardy to be on their team and then  “pay him a six figure salary the same way you would pay a non executive director or a VP of biz dev”.

Can algorithms outperform intuition?

As AI-powered funds like Rule30 and Boardy emerge, they're forcing the venture capital industry to confront uncomfortable questions about its own decision-making processes. If most VCs admit to pattern matching and the majority of deals come through warm introductions, can data-driven approaches level the playing field?

The answer may depend on how these algorithms are built and what data they're trained on.

Rule30's purely algorithmic approach could, in theory, remove human bias entirely, but it risks encoding historical patterns into its models. Boardy's hybrid model keeps humans in the loop, potentially preserving some of the relationship-building that makes venture capital work — or reintroducing the biases it aims to avoid.

What's clear is that both approaches will generate data that traditional VCs can't: concrete evidence of whether algorithms can consistently identify successful founders better than human judgment. In an industry built on instinct and reputation, that might be the most disruptive outcome.

Company data:

Rule30:

Founders: Guy Conway and Damien Cristian

Portfolio: Aiomics, Autonomous Minds, Emerge, Gyre Energy, Jiro, Kiin Bio, Optavex, Ploy, Stack8s, Synkka, Unbound

Boardy:

Founders: Andrew D’Souza (CEO, former founder and CEO of Clearco), Matt Stein, Shen Sivananthan and brothers Ankur Boyed and Abhinav Boyed (former CTO)

Team: Around 25

Photo by Planet Volumes on Unsplash

Just as 2025 shudders to a close, there’s been a dramatic flurry of developments in one of the most niche, most nerdy, and potentially most world-changing technologies, one that’s still waiting for its time in the spotlight: Quantum computing. This, some pundits argue, is the inevitable next big revolution in computing tech, and it’s one that could create computers so mind-bogglingly powerful they’d be hard to describe in comparison to today’s machines. 

This week, an analyst at global investment bank Jefferies lit a fire under the quantum computing world when he predicted that in 15 years the market for quantum computers would rocket from today’s $1 billion annual revenue figure to around $198 billion, with a short list of decade-old startups to watch, one of which is Maryland-based IonQ. This medium-sized firm landed itself in the spotlight back in June with a $1.075 billion (around £800 million) purchase of U.K.-based Oxford Ionics, with a goal of combining the two firms’ technologies to develop quantum computers that can rival, and then beat, today’s traditional computer chips.

I’ve put the cart before the horse a bit, so here’s a two-sentence summary of what quantum computing means. Traditional computer chips rely on electrical signals that are on or off, a binary system that is efficient, but which means there’s a limit to how much data can flow through a processor at any one time, since signals can only be in two states. Quantum computers work on signals that can be in multiple states all at once, all merged together thanks to the logic-defying physics of quantum mechanics: this means vast amounts of data can, in theory, be worked on inside the chip at any one time. It sounds like magic. And if all the bizarre and expensive tech can be made to work, these computers may be exponentially faster than today’s simple binary ones. 

And lest you think this is all niche, far-future, high-tech startup dream stuff, this week Amazon (whose cloud infrastructure quietly props up most of the internet and many of the apps and services you use for fun and work) reorganized some of its divisions, combining its chip-design unit with its artificial general intelligence unit and its quantum computing unit into one new entity. This may be seen as very much a long-term maneuver, positioning Amazon in the right place when quantum computing achieves the same kind of reliability and power as today’s “traditional” silicon chip computers. And AI is in the mix for one very obvious reason: AI processing requires truly vast computer power, which is currently expensive (in mid 2024 it was estimated that it cost Google around $191 million just to train the AI computer model, a price that excludes the cost of the supercomputer infrastructure it runs on).

Also this week, the U.K. government’s research and innovation funding wing UKRI announced it was allocating a special strategic budget of over £1 billion (around $1.34 billion US) for spending in the 2026-2030 window to bring “quantum technologies” from the lab into real-world commercial products. In this case, this is a special application of quantum computer technology for defence purposes like quantum sensing and secure comms, and also cash to help quantum startups to found and scale in the U.K.

You may think £1 billion is small change, the kind of money the U.K. government would easily lose down the back of the sofa at Number 10 Downing St., and in a way you’re right. In August, Microsoft revealed it was spending $1 billion with Danish partners to develop what may be the world’s most powerful commercial quantum computer. And that’s just one U.S. tech giant’s effort. But the UKRI’s cash is a welcome step in the right direction, nonetheless. Britain isn’t exactly known as a hotbed of traditional semiconductor innovation—maybe it can position itself better for the next revolution.

Meanwhile, there’s been a huge development in that other nerdy, niche, deep-tech innovation: AI. Lovable, the Stockholm-based vibe coding outfit which hit the headlines this year as the fastest-growing software startup in history, just raised $330 million in fresh capital, giving it a valuation of $6.6 billion—making it a hexacorn (?!) mere months after it raised $200 million at a $1.8 billion value. 

I guess the excitement about making everyone a coder, in line with Lovable founder Anton Osika’s assertion that “humans were born to build,” per the New York Times, far outstrips the notion of being able to process trillions of bits of computer data in the quantum blink of an eye. 

Courtesy of Enlightra

As artificial intelligence models grow larger and more complex, the industry’s biggest constraint is no longer how quickly chips can compute — but how quickly they can talk to each other.

Training today’s large language models requires thousands of GPUs to work in parallel, often spread across multiple racks or even entire data centres. While chipmakers have poured enormous resources into making individual processors faster and more powerful, the infrastructure that connects them has struggled to keep up. The result is a widening communication bottleneck that slows performance and drives up energy consumption.

“While everyone was fighting for the faster and bigger GPUs or TPUs [tensor processing units], this part was not completely omitted, but it was kind of overlooked,” Maxim Karpov, co-founder and co-CEO of Enlightra, tells Pathfounders. “And the gap between how fast you can compute, and how fast you can communicate the data between these GPUs has been growing in the past few years.”

Most of today’s GPU-to-GPU communication still relies on copper cables. As data rates increase, those connections become less energy efficient and work only over short distances. At extreme speeds, copper links can connect only a handful of GPUs at once — far short of what modern AI clusters require.

The alternative is light. Optical fibres can transmit far more data, over much longer distances, all while consuming less power. 

Enlightra, a startup that has been working on this problem since 2021, is now emerging from stealth to tackle this issue of data transfer.

The startup, founded by two long-time collaborators and researchers, has quietly raised $15m over the past few years to commercialise a new class of chip-scale photonic lasers designed to remove one of the key remaining bottlenecks in optical interconnects. Investors include Y Combinator, Runa Capital, Pegasus Tech Ventures, Protocol Labs, Halo Labs, Asymmetry Ventures and TRAC VC. 

Both academics by background, Enlightra’s founders went through Y Combinator and Intel Ignite in 2022 and 2023 respectively to learn as much as they could about how to turn their scientific discoveries into a viable startup.

At the core of Enlightra’s technology is a multi-wavelength laser architecture — sometimes referred to as “comb lasers” — that can generate many different colours of light from a single device. These colours appear like the teeth of a hair comb and allow multiple data streams to be transmitted simultaneously over the same optical fibre.

Today, optical systems typically rely on dozens of discrete lasers to achieve this effect, driving up cost, complexity and power usage. Enlightra replaces those components with a single silicon photonic chip that converts light from one laser into multiple wavelengths.

“Essentially, more colors is better,” Karpov says. And it helps with energy efficiency which is, he says, “a big issue for these AI clusters [because] they’re super energy hungry and they want to make sure that they are using their energy as efficiently as possible”.

The market opportunity is substantial. McKinsey estimates that the market for energy-efficient interconnects will reach $24bn by 2030, driven by AI workloads and the rapid expansion of data centre infrastructure. Major technology companies including NVIDIA, Google, Meta and Broadcom are all investing heavily in optical interconnects as data growth accelerates.

“AI is driving an optical revolution,” Dmitry Galperin, general partner at Runa Capital, said in a statement. “Enlightra’s multiwavelength lasers are a foundational technology for the next decade of high-performance computing, bringing the efficiency and scalability the industry urgently needs.”

While Karpov wouldn’t name Enlightra’s current customers, he said the company is already working with leading chipmakers. “I can tell you they are technology leaders developing chips for AI computing,” he said. “If you look at the top five companies developing AI chips, you’ll definitely find our customers.”

Operating in the same space but across the Atlantic, Xscape raised $44m Series A led by IAG Capital Partners last year, while Celestial closed a $175m Series C in March and Ayar Labs completed a $155m funding round that valued the company at over $1bn last December.

Company data:

• Founders: John Jost (co-CEO) and Maxim Karpov (co-CEO). Jost’s work has contributed to two Nobel Prize–winning advances in quantum computing and optical frequency combs. Karpov was honored in Photonics100 (2024) and named an MIT Innovator Under 35 in 2025.

• Founded in 2022 in Lausanne, Switzerland

• Team size: 25