DeepMind poker AI trio valued at $500M after Series A, now trading billions

What happened

EquiLibre Technologies, a Prague-based AI lab founded by three former DeepMind researchers, raised a Series A at a $500 million（約800億円） valuation led by Creandum. The founders—Martin Schmid (CEO), Rudolf Kadlec (CTO), and Matej Moravcik (CSO)—previously built DeepStack, an AI that defeated professional poker players; they are now applying reinforcement learning to stock and crypto trading in partnership with quant firm Tower Research Capital.

Why it matters

The startup's trading algorithms have generated "a perfect record of zero negative months since inception" across billions in daily volume on the S&P 500, NASDAQ, and crypto markets since their 2025 rollout. For quant hedge funds, where automation is already standard, faster or more profitable AI models can quickly translate to cash—making EquiLibre's approach attractive to VCs pursuing frontier AI from DeepMind alumni. The reinforcement learning technique the founders pioneered in games now underpins a much larger financial-markets use case.

What to watch

EquiLibre plans to scale its compute infrastructure to build one of the largest compute clusters in Central and Eastern Europe. The startup faces competition from established trading firms like Jane Street, which already deploys reinforcement learning and claims "tens of thousands of high-end GPUs"; EquiLibre is aiming to "get more from less" compute. Schmid stated the market is "not a winner-takes-all" despite these pressures.