CuspAI: When AI Meets Carbon Capture

In October 2024, as hurricanes batter the US East Coast as a symbol of climate change and AI is validated by Nobel Prizes for Physics and Chemistry, we spoke with Professor Max Welling, the distinguished AI pioneer and Research Chair at the University of Amsterdam, about the future of the company he co-founded, CuspAI, to help combat global warming.

Welling’s reputation alone should get the company far. He supervised many impactful PhDs at the University of Amsterdam, including ones by AI household names such as Durk Kingma, Taco Cohen, and Thomas Kipf. “And more recently, I expect the work of T. Anderson Keller and Sindy Löwe on how oscillators can make neural networks more brain-like, will provide the next wave of innovation,” says Max.

Emerging from stealth in June 2024, CuspAI secured €28 million in seed funding. With Meta as a partner and angel investments from Google DeepMind operators, CuspAI is now hiring top-tier global talent. In fact, the recent Nobel Prize Winner – and “godfather of AI” – Geoffrey Hinton is a CuspAI board advisor.

In short, CuspAI can now move forward in creating AI-designed materials to capture and store carbon – with Amsterdam Science Park acting as a natural home base.

How do you describe your job to those outside the field?

I’ve spent my career developing new methods to make machines smarter. And recently, I’ve been more inspired to use these smart machines to help solve societal problems, with climate change now being the most urgent. With only 12 to 13 years before retirement, this is really what I want to focus on now.

We are entering an era of materials-on-demand – both Google DeepMind and Microsoft have developed algorithms to generate materials with specific properties. How does that relate to what CuspAI is doing? And what are you doing that’s next-level?

These efforts are quite similar to what we do. But we’re also building an efficient search engine around it and an extensive database that captures all the things being computed so they can work to improve our models again. We are also building a platform for evaluating these materials. And we’re specialised in a particular class of materials for carbon capture: metal-organic frameworks. But these materials also have many other potential applications, including energy storage, catalysis, gas and water purification, and semiconductors.

Why did you set up the office here at Science Park?

We’re close to the sciences here. This is where the knowledge and talent are. We have very close ties to the informatics and chemistry institutes. And it’s those students in machine learning and chemistry that we eventually want to inspire and recruit. It’s ground zero for this kind of research, with several programmes that align very well with us, such as the Molecular and Material Design Technology hub and the AI for Sustainable Molecules and Materials. These are all things happening in parallel and synergistically with what we do. So, we feel that we’re in an excellent ecosystem here.

What’s the big message you want to spread?

AI has proven itself. But we are behind. We must invest in AI in the sciences at the university, governmental, and European levels. This technology can help solve urgent societal problems in climate change and healthcare.

Any message for students?

If you’re worried about the environment and sustainability and you work in AI, you don’t have to make any sacrifices. You can be well-compensated financially while working on truly meaningful things.