How did you react to the news that CuspAI had raised such an enormous amount?

“It’s surprising how much money is flowing into this field. By Dutch standards this is a huge amount — maybe the biggest Series B round ever. On the other hand, I’m also aware that our biggest competitors, like Periodic and Lila, are raising even more. So in that sense it’s not an absurd amount. I think it’s fair that they’re giving it to us, because I think we can make the best use of it and have the most impact. But it’s a lot of money, and a lot of responsibility to handle it well.”

When you were younger, did you ever expect you’d found such a successful company?

“As a student this wasn’t even on my radar. When I was young — back when the dinosaurs still roamed the earth — there were really only two options: you went into science, or you went to work for a company. Starting your own company wasn’t on anyone’s radar. That’s something I’d really like to change in the Netherlands. If I had to give myself one mission, it would be to at least get everyone to consider that a startup is a fantastic choice.”

How did you come up with the idea yourself?

“I first had an academic career. When I came back to the Netherlands, we started our first startup: Cypher. That was a very different kind of startup back then. That’s when I first tasted how fun it was — that positive energy and dynamism. After that I spent about seven years in Big Tech, at Qualcomm and Microsoft. That’s really industry; I learned a lot there, but it wasn’t for me. This startup is really Silicon Valley-style, and I find that a fantastic experience.”

What’s that experience like?

“Well, it’s moving faster than I expected. My co-founder Chad is incredibly successful at raising money and has a great network. Once you have momentum, it creates a kind of pull on talent and new investment. You get a snowball effect. Going from nothing to a $2.6 billion valuation in two years is of course beyond expectations — those are crazy numbers. But that we, especially in the configuration of Chad and myself, could pull this off really well — I’ve always had a lot of confidence in that.”

How will the money be used?

“We have a lot of plans. First, we need to expand the team in various places; we’re going to hire people in Singapore and Silicon Valley, among others. We’re also going to train more models, like our own agents and LLMs, which costs a lot of money. We’re going to generate data too, both digitally and through experimental facilities and labs. And we’ve just set up the Foundry, an ecosystem of fifty companies, where we want to create all kinds of value. That also requires an initial investment.”

What exactly does the Foundry involve?

“It’s a network in which you very quickly bring together the different parties you need to develop a new material. Think compute partners, AI partners, labs, and a customer with a specific problem. That way you can quickly kick off projects — sometimes pre-competitive, sometimes competitive — so that everyone can get a taste of AI and achieve successes together.”

What’s the goal?

“We want to bring about a paradigm shift. The field is still very traditional and conservative: groups of scientists reading papers and running experiments, but still making very little use of AI. Through this network we want to bring everyone along on the AI train and show that things can go much faster if they start using our platform.”

How did this funding round differ from the previous one?

“There are definitely new aspects. We want to significantly expand the agentic frameworks, take bigger steps in digital-twin-like simulations, experiment more, and generate more data. Those are things that are simply more expensive, and that we couldn’t do yet in the previous round.”

Do you think Europe is becoming too dependent on American tech companies?

“Looking at Europe and ourselves, I think we need to make Europe less dependent on other countries. We need to want to develop this kind of technology ourselves, and our company helps with that. But I see no problem at all in collaborating with American companies. Companies consist of many different groups, and within them are a lot of well-meaning people who want the best. You just have to pay close attention to make sure you’re working on the right problems, and not go along with a narrative you don’t agree with.”

What should Europe do, in your view, to keep up with American frontier labs like OpenAI and Anthropic?

“That’s a complicated issue. We now have a number of companies in England, France, and Germany building LLMs, like Mistral, Aleph Alpha, and Black Forest Labs, but American investment is enormous and hard to match. It’s important that Europe builds its own capacity, because there’s a chance that at some point both China and America will flip the switch. Then you need to have a good starting point of your own.”

Where do you see opportunities for Europe to keep up with these companies?

“In physical AI — an area where Europe is well positioned and where we should focus. We need to build our own compute infrastructure so we’re not one hundred percent dependent on other countries, and support our own AI frontier labs. That funding doesn’t even have to come entirely from Europe; it can partly come from America too, as long as we in Europe create the conditions in which these kinds of companies can grow and thrive.”

An enormous amount of money is currently being poured into AI. How do you make sure expectations around your company stay realistic?

“Unlike a lot of other LLM companies that promise a lot but haven’t delivered anything yet, we’re already very busy with real customers. We’ve signed contracts and there are already revenue streams coming in. That keeps us very grounded and well-rooted in the real economy — it’s not a castle in the air. Of course investing is risky, but companies like OpenAI and Anthropic are also making a lot of money. It’s all very rationally calculated and financially substantiated based on market size and the speed at which we can penetrate that market.”

What does CuspAI need to have achieved a year from now for this funding round to be called a success?

“We’re hard at work searching for a breakthrough material: a useful material that could never have been discovered without AI, and that the market is really waiting for. If we can achieve that in a number of areas together with our partners, the round will be a success. It’s also a success if we’ve significantly grown our number of paying customers. Ultimately it’s a matter of continuing to build: the platform needs to get bigger, more data needs to be gathered, and the whole machine needs to run faster.”

Where would you like to see CuspAI in five or ten years? What’s the ultimate dream?

“My dream is for CuspAI to function as an autonomous search engine for materials. That as a user you can have a frictionless conversation about what you’re looking for, after which calculations and experiments are automatically launched all over the world to produce that material. The pace at which new materials are developed, tested, and applied in real objects needs to go up enormously. My dream is that in a few years, when you buy a new chip, battery, or solar cell, it comes with “Cusp inside” as standard — meaning a Cusp material has been incorporated that makes the product fundamentally better.”

And how long before this becomes reality?

“The first commercial materials will likely be here within the next few years. What I do already see is that the time our company needs to arrive at a candidate material is getting shorter and shorter. We can already do that very quickly now.”

Interview by Bodiel Siebers