EquiLibre: The Poker AI Taking on Wall Street

A Transition from Poker to Stock Trading

Three former researchers from DeepMind, known for developing an artificial intelligence capable of beating humans at poker, have now redirected their expertise to the world of stock trading. Their new company, EquiLibre Technologies, based in Prague, has recently achieved an impressive valuation of $500 million. This valuation is the result of a fundraising round during a Series A, although the exact amount raised has not been disclosed, as reported by TechCrunch.

The funding round was led by Creandum, which confirmed that this is its largest single investment in a company to date. Cameron Sellers, vice president of Creandum, explained to TechCrunch that the link between poker and Wall Street lies in their propensity for reinforcement learning. This AI training technique relies on self-learning models motivated by rewards. Martin Schmid, CEO of EquiLibre, emphasized that the advantage of trading lies in the simplicity of its evaluation: it is simply about knowing how much money the agent has made.

Impressive Performance in the Markets

EquiLibre is not just playing with fake money. In collaboration with Tower Research Capital, their algorithms have been used to trade billions of dollars in daily volume on major indices such as the S&P 500 and Nasdaq. Since their launch in the cryptocurrency markets in 2025, EquiLibre's agents have shown remarkable performance, with a perfect record of zero negative months, meaning they have always finished each month with investments up.

By integrating its AI into the field of quantitative hedge funds, EquiLibre is venturing into a sector where automation is already well established. However, if successful, the improvements brought by their technology could quickly translate into significant financial gains. This prospect has made the startup particularly attractive to Creandum. Sellers noted that the total addressable market for trading in financial markets is one of the largest in the world, far surpassing most venture-backed successes.

An Innovation-Focused Approach

Although EquiLibre focuses on trading, Schmid insists that the company defines itself primarily as an innovation lab rather than a financial firm. Neither he nor his co-founders, Rudolf Kadlec and Matej Moravcik, have a background in finance. Their motivation lies in creating new, unprecedented technologies, not in optimizing financial markets. "I'm not doing this to make markets more efficient, but because we love building new things," Schmid stated.

Interest in AI startups led by former DeepMind members is strong among venture capital investors. Ineffable Intelligence, for example, recently raised $1.1 billion. While most of these companies are based in the UK, EquiLibre stands out by being located in the Czech Republic.

Strong Academic Roots

The founders of EquiLibre began their journey as PhD students at Google's international AI research office in Edmonton, Alberta, Canada, which was closed by Alphabet in 2023. It was there that they developed DeepStack, the first AI program to beat professional players at Texas hold 'em poker. They also collaborated with professors who are now advisory board members of EquiLibre, including Rich Sutton, who won the Turing Award in 2024 for his work on reinforcement learning.

To launch their startup, the founders chose to return to the Czech Republic. "That's where we had a lot of collaborators, and there was a large Czech diaspora at Google and elsewhere," Schmid explained. This return facilitated the formation of their initial team in 2022, which now consists of 25 people. According to Schmid, Prague offers a stable environment, away from the fleeting trends that characterize cities like San Francisco.

A Promising Future in AI

EquiLibre is not the only AI startup emerging in the region, but it stands out for the quality of its talent. The company plans to expand its computing infrastructure to become one of the largest computing clusters in Central and Eastern Europe.

Although the startup has not disclosed the total amount of its funding to date, Schmid mentioned that it has already raised two other funding rounds. Among the investors is Credo, a venture capital firm focused on Central and Eastern Europe, which has also supported ElevenLabs and UiPath. According to Dealroom, EquiLibre previously raised $10 million in a round led by Blossom Capital, reaching a valuation of $140 million at that time.

The current valuation of $500 million represents a significant leap, coinciding with the rise of reinforcement learning in trading. "When we started, people were skeptical," Schmid said. Today, this approach has become commonplace, and EquiLibre considers itself ahead due to its accumulated experience.

A Competitive Market

However, competition is fierce. Trading giants like Jane Street are already using reinforcement learning with large language models and have considerable computing resources. EquiLibre, on the other hand, seeks to maximize the efficiency of its infrastructure with a limited number of chips, according to Schmid.

To achieve its goal of becoming a leader in AI applied to trading, EquiLibre will need to navigate this competitive market skillfully. Schmid remains optimistic, asserting that the sector is not a zero-sum game: "This is not a market where the winner takes all."