Applied Deep Tech and the European Founder Revolution with NP-Hard Ventures' Paul Veugen

Paul Veugen is co-founder and partner at NP-Hard Ventures, an Amsterdam-based early-stage venture capital firm investing in pre-seed and seed companies solving the hardest problems across robotics, defense tech, and foundational infrastructure.

You've had a long career as a founder. How did that experience lead you to start NP-Hard Ventures?

I started Usabilla 16 years ago as a university student. It was a product that helped companies collect feedback on live webpages. I accidentally became an entrepreneur. I always joke that I raised capital from the only two active angel investors in the Netherlands at the time, and I showed up to pitch them only to see them shake hands to close the deal. At that point I realized I was actually doing this.

Four years in, I realized the business needed enterprise sales. I handed over the keys to my co-founder, Marc, who initially joined as CTO, and turned it into a massive success. Usabilla eventually sold to SurveyMonkey. But that moment of shaking hands with my first  angel investors planted a seed. I knew that someday I wanted to help founders accelerate their careers in a similar way.

After Usabilla, I started Human, one of the first all-day activity trackers on the App Store. We looked at the Fitbit and the Nike FuelBand and thought: nothing in those devices isn't also in your iPhone. So we built a tracker with a gameplay layer on top. That company was acquired by Mapbox, and as part of the deal I moved my family to San Francisco. I came in at about 150 people and left at roughly 400. I learned what it meant to scale an organization, going from inheriting a team of five to running a 60-person division.

While working at Color, a DNA testing company, I was sitting behind my desk when my first company got acquired and my payout came through. It gave me the freedom to think about what I really wanted to do. I stayed on at Color because it's a fabulous company, but I also started angel investing, primarily in European companies. In Europe, only about 7 or 8 percent of investors have any operational or founder experience. That's a combined bucket. As a founder, I'd experienced firsthand how hard it is to find capital from people who actually understand what it takes to build a company.

I started with about 30 angel deals, first by myself, then with Micha, one of the other partners, and later Anke. The angel checks were 15,000 or 25,000 euros. There's only so much impact you can make with small check. So we formalized it into NP-Hard Ventures: the fund we wished we'd had on our own cap table as founders. Anke leads the fund full time as managing partner. Micha and I are building our own companies alongside the fund. My wife says I have two full-time jobs now, but the setup gives us a unique lens. We're building every day with the technology we invest in.

What is NP-Hard's investment thesis?

Our first fund was a 12 million euro pre-seed and seed fund. About 70 to 80 percent of our capital went into European companies. Our second fund, which we're currently raising and is already operational, will be 25 million euros. We invest in what we call "applied deep tech," meaning deep technology with a clear line to business today. Our focus areas are robotics, defense tech, and foundational infrastructure.

We've invested in Monumental, a robotics company using swarms of autonomous robots to lay bricks on construction sites. We invested in Ark Robotics, which builds autonomous and semi-autonomous land drones that help protect borders in Ukraine and defend Europe. And Mirai helps developers build with hyper-optimized LLMs that can run fast on any type of device. These are the core layers of infrastructure we'll build on in the next decade, and we view most of them through a European lens.

You've seen the European venture ecosystem from both sides, as a founder and an investor. What's driving the shift you're seeing today?

The companies we invest in all have wildly ambitious, global goals. They happen to be incorporated in Europe, but they have worldwide visions. What I'm seeing is more founders who combine bold missions with an insane set of skills and a different perspective on the world they want to build. Salar at Monumental is a good example. He observed the world around him, picked one of the biggest problems he could think of, and is now tackling it step by step. He dares to be open about his ambition and turn it into action. I'm seeing more of that unapologetic ambition in Europe, and it makes me very bullish.

I spent three years living and working in San Francisco, and I experienced an incredible talent pool there. But I've experienced the same in Europe. We don't lack talent. I'd argue we don't lack ambition either. We're just more down to earth about it, and we complain more. We have this Calimero complex that we need to shake off. That's starting to happen now, partly driven by broken trust and geopolitical shifts. There's a growing realization that we can't just keep looking to others. We have challenges we need to solve ourselves. As Europeans, especially in the Netherlands, we tend to be quiet about our ambitions. I enjoy seeing more people being vocal about what they want to achieve.

There's a perception that talent, customer, and capital pools are all deeper in the U.S. Do you agree?

On the capital side, yes. European capital is more risk-averse. Investors here tend to judge risk, while in the U.S. they judge opportunity. That creates a very different dynamic for founders. But it's improving. We're seeing bigger rounds happen, smarter capital being deployed, and real momentum building.

On talent, I push back. If I put teams side by side, there are amazing people on both sides of the Atlantic. If you look at venture capital returns and account for the handful of super outliers, I don't think the gap is what people assume. I invest mostly in Europe and I get excellent returns. The main thing is that we're not good at selling ourselves. Our storytelling is terrible. European founders sometimes lack pitching skills. In the U.S., founders learn to sell their vision so convincingly that even if they're destroying the earth, they might still make you believe it's the most noble cause you could chase. That's a real skill.

We also have a language handicap. Someone else put it well: "I'm losing 30 IQ points the moment I have to switch to a different language." That matters when you're building truly global companies and you need to craft a compelling pitch narrative. A big part of building a company from nothing is aspirational. You need to make people believe. Maybe we should also just stop comparing ourselves. That would be a good starting point.

How do you think about the go-to-market motion for robotics companies? Is the lab-to-field gap getting smaller?

Our first fund focused more on software infrastructure, dev tools, and SaaS. But we've seen a clear trend: the best robotics teams are tackling their challenges with a software-like iterative approach. Fast loops, quick iterations, constantly evolving the product directly in the field rather than in a lab.

Monumental started with technical Lego to prototype, then moved to 3D-printed prototypes that could actually lay bricks but would break down after three or four uses, then to self-produced machines that now operate at scale on real construction sites. They deploy teams of robots as subcontractors. Salar's model of owning the entire stack and getting paid per brick enabled them to build a fundamentally different type of robotics company.

We see the same pattern with Ark Robotics. The companies making the biggest progress are the ones that innovate in the field. The companies with superior technology on paper sometimes fail when put to a real operational test. In Ukraine, by the time you ship something to the battlefield, the battlefield has changed. The most successful approaches are almost counterintuitive: high tech with low tech components and rapid iteration. It's the companies that literally innovated on the battlefield that have been most effective. That's why I make the comparison to software iterations.

If you zoom out further, the software interface itself is disappearing. AI is likely becoming the interface to our products. If the cost of building software drops close to zero, what you're left with is protocols and foundational infrastructure layers. The models powering all of this are still taking a weird approach. We act as if compute is free. The LLMs are amazing, but we're basically sending a McKinsey consultant to the local bakery to help with their bookkeeping. I'm interested in how you orchestrate that more intelligently. A lot of it will involve physical components steering objects in the real world, and it won't be at the scale of a massive Tier 1 LLM. It will take different shapes: speech to text on a local device in your meeting room, on-device models powering your iPhone experience, simple apps.

How do you evaluate founders and opportunities at the pre-seed stage?

Most of the founders we talk to are very early. They usually have a prototype or an early product. Today, you should be able to show something. We often ask people in the first or second conversation: can you show us something? Pull something up. In some cases we see a terminal window and get excuses. But that's the part we're looking for. I don't need a fully fledged product. We want to reason through the product together. Most of our conversations go deep on product and go-to-market strategy because that's essentially all we can judge.

We're evaluating the team, their experience, what drives them to get out of bed every day, combined with whether they can articulate what's unique about their approach. What do you know about the market that other people don't? What's your technical insight that others aren't aware of or don't share? That's basically everything we have to work with at this stage. There's almost never significant revenue. We don't require repeat founders, though we have quite a few. And we don't wait for other investors. The biggest frustration as a founder is the dance where nobody is interested until someone is, and then everyone piles in. We try to commit early, sometimes before we know exact terms. We help founders shape their round and bring in people around us who we trust.

Is there something you believe as an investor today that's contrarian but will be accepted as obvious truth in the near future?

I think we're seeing the very early starting point of everything with big lab AI. I don't believe the current world where OpenAI and Anthropic are dominant represents more than a fraction of what's possible. The products they're shipping are still primitive. We're only now seeing early product-market fit on specific challenges. I've experienced it myself with Claude Code for coding. But if that shape and form can be applied in so many different directions, then the race to build the smartest model to execute the most ambiguous task in the most ambiguous environment starts to look less like a winning strategy. We've essentially built Swiss Army knives. They're wildly expensive and not optimized.

There are so many other races that still need to be run. You can compete with these companies. Not easily, but I see no reason why you can't build a better product. The benchmarks might be three or four months behind, but consumers are slower than that. Building a higher-quality product with a model that's six months behind becomes a product challenge, not a compute challenge. Anthropic is showing good examples of this by hiring designers like Mike Krieger to build interfaces. A person on the team built a solution that's now used by half the engineering workforce in the world. Those are examples of the actual thing that happened. You can do that with different types of models too.

How do you define deep tech?

We have a simple take on it. We don't define deep tech the traditional way. We put the label "applied" in front of it. We're looking for long-term technical solutions to complicated problems that can be adapted and brought to market in a shorter time frame. Most of the companies we invest in can go to market within 12 to 18 months. In the traditional sense, that's not even deep tech. But I think development cycles are getting shorter for these types of technologies.

We're also looking at areas like quantum computing, where you have a nonexistent market on a nonexistent product category with very different timelines. Even then, we'd judge a company on its ability to execute in the ecosystem, not on the market, which may not exist yet. Their execution would need to be much more iterative. I would define deep tech for us as really hard, NP-hard problems.

SPONSORED

Infinite Frontiers is made possible by HAUS, the leading public relations agency for deep tech startups and scaleups.

Learn more