From poker to profit: DeepMind alumni’s AI lab is now quietly trading billions in stocks and crypto

BitcoinWorld

From poker to profit: DeepMind alumni’s AI lab is now quietly trading billions in stocks and crypto

Three former DeepMind researchers who built an AI that defeated professional poker players have turned their attention to financial markets — and the strategy appears to be working. Their Prague-based lab, EquiLibre Technologies, is now valued at $500 million after raising a Series A round led by Creandum, the venture capital firm confirmed to Bitcoin World. While the exact size of the round was not disclosed, Creandum partner Cameron Sellers described it as the largest single investment the firm has ever made in one go into a company.

Reinforcement learning finds a new home on Wall Street

The connection between poker and stock trading lies in a specific AI technique called reinforcement learning. In this approach, self-learning models are trained through a system of rewards and penalties — essentially learning optimal strategies through trial and error. According to Martin Schmid, CEO and co-founder of EquiLibre, financial markets offer a uniquely clear feedback signal. “The nice thing about trading and markets is that the scoring is super simple: how much money did the agent make?” Schmid told Bitcoin World.

This isn’t a theoretical exercise. EquiLibre has partnered with quantitative trading firm Tower Research Capital to deploy its algorithms across major U.S. stock indices, including the S&P 500 and NASDAQ. The startup claims its AI agents have been trading billions of dollars in daily volume and have posted “a perfect record of zero negative months since inception” — meaning each calendar month has ended with net positive returns. The same system has been active in cryptocurrency markets since 2025.

A lab-first identity in a profit-driven industry

Despite operating in the high-stakes world of quantitative hedge funds, EquiLibre explicitly positions itself as a research lab rather than a finance firm. Schmid emphasized that he and his co-founders — CTO Rudolf Kadlec and CSO Matej Moravcik — are not motivated by a desire to optimize market efficiency. “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.

This lab-first identity resonated with Creandum. Sellers noted that the total addressable market for AI-driven trading is enormous, with many funds having generated profits that “make most venture-backed successes look small.” However, he acknowledged that EquiLibre’s founders bring a rare combination of deep reinforcement learning expertise and a willingness to apply it to a domain where automation is already well established.

From Edmonton to Prague: building a team outside the AI hype cycle

The founding trio first met as visiting PhD students at DeepMind’s Edmonton office in Alberta, Canada — Alphabet’s first international AI research location, which was shut down in 2023. While there, they built DeepStack, the first AI program to defeat professional players at no-limit Texas hold’em. They also worked alongside Rich Sutton, who received the 2024 Turing Award for his foundational contributions to reinforcement learning and now serves on EquiLibre’s advisory board.

Rather than remaining in North America, the founders chose to relocate to their home country, the Czech Republic. “This is where we had a lot of people we had worked with, and there was a large Czech diaspora at Google and other places,” Schmid explained. “These were our friends, so we told them, ‘Hey, guys, we are moving back to Prague, do you want to join us?’” That decision helped the startup build an initial team of 25 people starting in 2022. Schmid believes the location remains an advantage: “It’s much easier to keep the good people here, because there’s not a new sexy AI thing happening every two months.”

Scaling compute and the competitive landscape

EquiLibre plans to use its Series A funding to scale its compute infrastructure, aiming to build what it expects will be one of the largest computing clusters in Central and Eastern Europe. The company previously raised a $10 million seed round led by Blossom Capital at a $140 million valuation, according to Dealroom data. Pre-seed backing came from CEE-focused VC Credo, which also backed ElevenLabs and UiPath.

Reinforcement learning has gained broader acceptance in trading since EquiLibre was founded. “When we started, people were skeptical,” Schmid said. “Because we started four years back, we believe we are ahead.” Still, competition is intense. Trading giant Jane Street has stated it already uses reinforcement learning alongside large language models and claims access to tens of thousands of high-end GPUs. EquiLibre aims to differentiate by extracting more performance from fewer chips — “get more from less,” as Schmid put it.

Schmid does not view the market as a zero-sum game. “This is not a winner-takes-all market,” he said. EquiLibre’s goal is to be recognized as “the AI lab in trading” — a reputation built on research output and consistent returns, not on capturing every dollar in the market.

Conclusion

EquiLibre Technologies represents a notable convergence of frontier AI research and practical financial application. Its founders bring a track record of building AI that outperforms humans in complex, imperfect-information environments — and they are now applying that same approach to markets where similar conditions apply. The startup’s $500 million valuation, its partnership with a major quant firm, and its stated goal of building one of the largest compute clusters in Central Europe suggest that reinforcement learning may have found a permanent home in quantitative finance. For now, the bet appears to be paying off.

FAQs

Q1: What is reinforcement learning, and why is it useful for trading?
Reinforcement learning is an AI training technique where models learn optimal behavior through a system of rewards and penalties. In trading, the reward signal is clear and measurable: profit and loss. This makes financial markets a natural fit for RL, as the AI can continuously refine its strategy based on outcomes.

Q2: Who are the founders of EquiLibre Technologies?
The company was founded by Martin Schmid (CEO), Rudolf Kadlec (CTO), and Matej Moravcik (CSO). All three were visiting PhD students at DeepMind’s Edmonton office and were part of the team that built DeepStack, the first AI to beat professional poker players at no-limit Texas hold’em.

Q3: How does EquiLibre make money?
EquiLibre applies its reinforcement learning algorithms to trade stocks and cryptocurrencies in partnership with quantitative hedge funds, including Tower Research Capital. The AI agents execute trades across major indices like the S&P 500 and NASDAQ, generating returns based on the strategies learned through reinforcement learning.

This post From poker to profit: DeepMind alumni’s AI lab is now quietly trading billions in stocks and crypto first appeared on BitcoinWorld.