Potential Sciences Information

Potential Sciences is a frontier lab for chemistry. AI models in chemistry are starved of data: they all learn from the same finite record of past experiments, or from simulations too crude to trust. We generate new training data from the laws of physics, at the accuracy of a laboratory measurement. Our models improve with compute rather than with lab time, so scientists know how a molecule will behave before anyone makes it. We start with drug discovery, where the best models have already learned everything the existing record can teach, and every failed experiment costs years. Hatem Helal founded Potential Sciences in October 2025 to build the lab this problem needs. We are mathematicians, physicists and chemists trained at Caltech, Cambridge and CERN, who went on to build AI systems for science at Graphcore, Recursion and Dayhoff Labs. We are based in Cambridge, England, with researchers in Bristol and Tübingen, and advised by Andrew Fitzgibbon FRS (Graphcore, Microsoft Research) and James Hawkins (co-founder and co-CEO of PostHog). We have raised eight figures in capital and compute, with the UK's Sovereign AI Fund and Twin Path Ventures as key investors, alongside Blue Wire Capital, Lightfield Equity, Nauta Capital, Odyssey, Portfolio Ventures and Rule30. The Sovereign AI Fund backs us as both investor and compute partner, with access to national supercomputing through the AI Research Resource (AIRR). Get in touch: hello@potentialsciences.com