From the Card Table to the Trading Desk
A poker-playing AI has a new opponent: the market. The three former DeepMind researchers who built it now run EquiLibre Technologies, a Prague lab valued at $500 million after a Series A led by Creandum. Their agents already trade billions a day, and so far they have not posted a losing month.
Stock trading – artistic impression. Image credit: Jason Briscoe / Unsplash
Key Takeaways
- A Creandum-led Series A lifted EquiLibre Technologies to a $500 million valuation, the venture firm’s biggest single-company investment to date.
- The lab’s agents trade billions daily on the S&P 500 and Nasdaq alongside Tower Research Capital and claim no losing month since they began.
- Founded by the team behind the DeepStack poker AI, EquiLibre runs on reinforcement learning, which treats profit as its only score.
One Skill, Two Games
The link between cards and markets comes down to how the AI learns. Reinforcement learning rewards a model for good moves and lets it sharpen through trial and error. Poker fits that method. So does trading, and trading keeps score in the plainest way possible. “The nice thing about trading and markets is that the scoring is super simple: how much money did the agent make?” said Martin Schmid, EquiLibre’s CEO.
The results are not hypothetical. Working with Tower Research Capital, EquiLibre’s algorithms move billions in daily volume across the S&P 500 and Nasdaq. The company says its agents held up after launching on crypto in 2025 and now on stock exchanges, pointing to “a perfect record of zero negative months since inception.” Researchers mapping out reinforcement learning for quantitative finance flagged this promise years ago.
The Round, and Why Creandum Went Big
Creandum led the raise, and the firm did not hedge on scale. Vice president Cameron Sellers confirmed it was the largest single investment the firm “has ever made in one go into a company.” He pointed to the size of the prize. “The potential total addressable market of trading in the financial markets is one of the biggest on earth, and there are countless funds over the years that have generated quantums of profit that make most venture-backed successes look small,” Sellers said. He added that EquiLibre bills itself as “a lab first, not a finance firm.”
Founders Who Would Rather Build
None of the founders came up in finance, and Schmid says that is the point. “I’m not doing this because I’m excited about making markets efficient. I’m doing this because we are all excited about building new things that have never been built before, and this is a lot of fun to build,” he said. Alongside CTO Rudolf Kadlec and CSO Matej Moravcik, he first met at DeepMind’s Edmonton office, later closed by Alphabet in 2023. There the group built DeepStack, the first AI to beat pros at no-limit Texas hold’em. Rich Sutton, winner of the 2024 Turing Award for his reinforcement learning work, now advises the company. Investors have been quick to fund labs spun out of DeepMind talent.
Scaling Up Against Deeper Pockets
The valuation climbed in stages. A pre-seed came from Credo, an early backer of ElevenLabs and UiPath. A $10 million seed led by Blossom Capital set the price at $140 million. The Series A now marks $500 million. With a team of 25, EquiLibre plans to expand its compute next and build one of the largest clusters in Central and Eastern Europe. Location helps, Schmid says. “It’s much easier to keep the good people here, because there’s not a new sexy AI thing happening every two months.”
The rivals are formidable. Jane Street says it already runs reinforcement learning with large language models, “or whatever else we need to train good models,” and claims “tens of thousands of high-end GPUs.” EquiLibre wants to wring more out of far fewer chips and “get more from less,” Schmid said. He is not framing it as a duel to the finish. “This is not a winner-takes-all market.”
A Pattern Investors Keep Backing
Money follows DeepMind alumni these days. Elite researchers keep leaving the big labs to start their own, and the checks arriving for those teams have grown enormous across the AI field, part of a broader surge in startup funding. EquiLibre’s founders describe trading as a place where the market updates its verdict every millisecond, with a research team pulled from DeepMind, Google, and leading trading houses. Big tech has noticed the appeal, funneling capital and tools toward startups built on frontier AI research.
Quantitative funds are drifting the same way. Machine learning keeps replacing older statistical models, and reinforcement learning has moved from fringe to standard in a handful of years. “When we started, people were skeptical,” Schmid said, though the mood has turned. “Because we started four years back, we believe we are ahead.” Whether that lead lasts is the open question, and the report that first detailed the round makes plain that well-funded rivals are training the same methods on far more silicon.