DeepMind Poker AI Trio's EquiLibre Hits $500M Valuation In Creandum Series A
EquiLibre Technologies, the Prague AI lab spun out of DeepMind's DeepStack poker team, closed a Creandum-led Series A at a $500M+ valuation to scale reinforcement-learning trading agents.
Key Takeaways
EquiLibre Technologies, the Prague AI lab founded by ex-DeepMind researchers Martin Schmid, Rudolf Kadlec and Matej Moravcik, closed a Creandum-led Series A at a valuation topping $500 million (about €438 million), reported July 9, 2026.
Creandum VP Cameron Sellers confirmed it is the largest single-company investment the firm has ever made; the round size itself was not disclosed.
The founders created DeepStack at DeepMind's Edmonton office, the first AI to beat professional heads-up no-limit hold'em players (published in Science, 2017), and have ported that reinforcement-learning stack to trading.
In partnership with Tower Research Capital, EquiLibre's agents trade billions of dollars in daily volume across the S&P 500, Nasdaq and crypto venues, with no negative months reported since inception.
Proceeds will fund one of the largest computing clusters in Central and Eastern Europe, expand the research team, and extend trading into futures and options; the company previously raised a $10 million seed led by Blossom Capital.
Kaan Tınmaz
EquiLibre Technologies, the Prague-based AI lab founded by former DeepMind researchers Martin Schmid, Rudolf Kadlec and Matej Moravcik, has closed an undisclosed Series A round led by Creandum at a valuation topping $500 million (roughly €438 million), AI Insider reported on July 9, 2026. Creandum vice president Cameron Sellers confirmed it is the largest single-company investment the firm has ever made.
From DeepStack to live markets
The trio built their reputation as visiting PhD students at DeepMind's former Edmonton office, where they authored DeepStack — the first AI to defeat professional heads-up no-limit hold'em players, published in Science in 2017. They have since ported that reinforcement-learning stack to financial markets. In partnership with quant firm Tower Research Capital, EquiLibre's agents now trade billions of dollars in daily volume across the S&P 500, Nasdaq and crypto venues, with the company reporting no negative months since inception.
Prague as a compute hub
The three founders relocated from Canada to their native Czechia to launch the company in 2022, citing access to a strong local RL talent pool. EquiLibre previously raised a $10 million seed round led by Blossom Capital. Sellers noted that trading is one of the largest addressable markets globally, though the firm still describes itself as a research lab rather than a finance company. Proceeds will fund one of the largest computing clusters in Central and Eastern Europe.
Physical-vs-financial RL
The round arrives days after Mistral released Robostral Navigate, its first RL-trained robot navigation model — underscoring that the same simulation-heavy training loop the DeepStack trio pioneered is now the dominant frontier both in trading and physical AI. It also lands alongside a fresh wave of European funding rounds targeting decision automation, including Taktile's $110M Series C and Trase's $107M seed.
What's next
EquiLibre plans to expand its research team of deep-learning engineers and stand up new production strategies before extending Tower Research's book into futures and options. Sellers said Creandum sees the company as a rare crossover bet capable of matching DeepMind's frontier RL research standard while running self-funding revenue in a live market. The founders declined to break out either the round size or projected AUM, calling both premature.
Reporting based on coverage from AI Insider, TechCrunch, EU-Startups and SiliconANGLE.
EquiLibre is a Prague-based AI research lab founded in 2022 by former DeepMind researchers Martin Schmid, Rudolf Kadlec and Matej Moravcik, the creators of the DeepStack poker AI. It applies reinforcement learning to financial trading in partnership with Tower Research Capital, though it describes itself as a research lab rather than a finance company.
How much is EquiLibre worth after the Series A?
The Creandum-led Series A values EquiLibre at more than $500 million (roughly €438 million). The exact round size was not disclosed, but Creandum said it is the largest single-company investment in the firm's history.
How does EquiLibre make money?
Its reinforcement-learning trading agents, run in partnership with quant firm Tower Research Capital, trade billions of dollars in daily volume across the S&P 500, Nasdaq and crypto venues. The company reports no negative months since inception and plans to expand into futures and options.
What will EquiLibre do with the new funding?
The company plans to build one of the largest computing clusters in Central and Eastern Europe, hire more deep-learning engineers, stand up new production trading strategies, and extend Tower Research's book into futures and options.
