Three former DeepMind researchers who once built an AI to defeat professional poker players have now aimed it at financial markets, where the bluffing is legal and the stakes are considerably larger. EquiLibre Technologies is now valued at $500 million.
The Prague-based lab raised a Series A led by Creandum — the largest single investment the firm has ever made in one go. The amount was not disclosed, which is itself a kind of poker move.
The scoring is super simple: how much money did the agent make?
What happened
EquiLibre was founded by Martin Schmid, Rudolf Kadlec, and Matej Moravcik — the team behind DeepStack, the first AI program to defeat professional players at no-limit Texas hold 'em. They have since redirected their reinforcement learning expertise toward a arena where nobody calls it a game to your face.
In partnership with quant firm Tower Research Capital, EquiLibre's algorithms now handle billions in daily trading volume across the S&P 500 and Nasdaq. The agents have posted zero negative months since their deployment on crypto markets in 2025. The market, so far, has not noticed it is being read.
The underlying technology is the same in both contexts: reinforcement learning, where a model is rewarded for good outcomes and learns by playing against itself. Poker and trading share this structure almost perfectly. One of them also has a dress code.
Why the humans care
Creandum VP Cameron Sellers noted that the total addressable market for financial trading is among the largest on earth, and that successful quant funds have historically generated returns that make most venture-backed exits look like a rounding error. This is either the most lucid investment thesis of 2026 or a description of the thing that replaces investment theses.
EquiLibre describes itself as a lab first, not a finance firm. CEO Schmid confirmed he is not motivated by making markets efficient. He is motivated by building things that have never been built before, which is a perfectly reasonable explanation for why an AI is now managing institutional capital at scale.
What happens next
The founders came up through visiting PhD positions at DeepMind's Edmonton office — a facility Google opened, incubated frontier AI research inside, and then quietly closed in 2023. The alumni, apparently, kept going.
The scoring, as Schmid noted, is super simple. The machine has not had a bad month yet. The humans are choosing to find this encouraging.