The DeepMind trio who built a poker AI are now making money for quant hedge funds
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- EquiLibre Technologies raised a Series A at a $500 million valuation led by Creandum — which VP Cameron Sellers called the largest single investment the firm has ever made in one company — marking a sharp jump from its $140 million seed valuation reported by Dealroom.
- EquiLibre's algorithms are trading billions in daily volume across the S&P 500 and Nasdaq in partnership with Tower Research Capital, with the company claiming a "perfect record of zero negative months since inception" since rolling out on crypto markets in 2025.
- Martin Schmid (CEO), Rudolf Kadlec (CTO), and Matej Moravcik (CSO) built DeepStack — the first AI to defeat professional players at no-limit poker — as visiting PhD students at DeepMind's Edmonton office (shut down by Alphabet in 2023), and applied the same reinforcement-learning technique to markets because, as Schmid put it, "the scoring is super simple: how much money did the agent make?"
- EquiLibre is positioning itself as a research lab rather than a finance firm and plans to bring online what it expects will be one of the largest compute clusters in Central and Eastern Europe, competing against trading giants like Jane Street, which says it already uses RL with LLMs and operates "tens of thousands of high-end GPUs."
- Creandum's Sellers framed the bet as a play on frontier AI from DeepMind alumni — a hot category that also includes Ineffable Intelligence's recent $1.1 billion raise — while Schmid argued EquiLibre's four-year head start on RL-based trading means "we believe we are ahead."
- EquiLibre's 25-person team is based in Prague partly because, per Schmid, it's "much easier to keep the good people here, because there's not a new sexy AI thing happening every two months," and the startup shares a building with fellow Czech AI startup BottleCap AI.
Why it matters: EquiLibre's Series A at a $500 million valuation — roughly 3.6x its $140 million seed mark — shows VCs will pay a premium for DeepMind alumni applying reinforcement learning to quant trading, where even small edge gains convert directly to cash. The startup's self-claimed "zero negative months" track record trading billions via Tower Research Capital is its central pitch, but Jane Street's tens of thousands of high-end GPUs set a compute gap Schmid concedes the 25-person Prague team must close by "getting more from less."
Ask SkimNews
