Three former DeepMind researchers who built the poker-beating AI DeepStack have applied similar reinforcement-learning techniques to financial markets. Their Prague-based AI lab, EquiLibre Technologies, has been valued at $500 million following a Series A round led by Creandum, TechCrunch learned. The amount raised in the round was not disclosed.
Why poker and markets align
EquiLibre’s work is grounded in reinforcement learning (RL), where models learn through rewards. Martin Schmid, EquiLibre’s CEO, notes that market trading has a clear performance metric: how much money an agent makes.
Real capital, real volumes
According to the company, its algorithms began trading on crypto markets in 2025 and have since been deployed on stock exchanges. In partnership with quant firm Tower Research Capital, EquiLibre says its agents trade billions of dollars in daily volume across S&P 500 and Nasdaq instruments. The startup further claims a "perfect record of zero negative months since inception," meaning every month has closed with net gains.
Investment context and market potential
Creandum led the Series A; Cameron Sellers, vice president at Creandum, told TechCrunch this was the largest single investment the firm had ever made in one go, though the exact sum was not revealed. Sellers emphasized the enormous total addressable market in financial trading and noted that quant funds can generate profits on a scale that dwarf typical venture returns. He also stressed that EquiLibre positions itself primarily as a lab rather than a finance company.
Founders, origins and team
Founders Martin Schmid (CEO), Rudolf Kadlec (CTO) and Matej Moravcik (CSO) were visiting PhD students at Google-owned DeepMind’s first international research office in Edmonton, Alberta — an office Alphabet closed in 2023. There they developed DeepStack, the first AI to beat professional players at no-limit Texas hold ’em. Several academics who worked with them are now advisers to the startup, including Rich Sutton, who won the Turing Award in 2024 for his work on reinforcement learning.
The founders relocated to their home country, Czechia, and established EquiLibre in Prague. Schmid said returning home helped them recruit former colleagues: the company formed its initial team in 2022 and today has about 25 employees.
Funding history and growth plans
EquiLibre declined to disclose total funding to date but said it previously raised two other rounds. Pre-seed backers included Credo, a CEE-focused VC that has backed ElevenLabs and UiPath. Dealroom data indicates the company’s $10 million seed round was led by Blossom Capital at a $140 million valuation.
The startup now plans to scale its compute infrastructure and bring online what it expects will be one of the largest compute clusters in Central and Eastern Europe (CEE), aiming to increase capacity for training and running RL models.
Competition and risks
Reinforcement learning has become more accepted in trading since EquiLibre’s founding four years ago, Schmid said, but the field is competitive. Trading firm Jane Street, for example, says it already uses RL with LLMs and claims to operate tens of thousands of high-end GPUs. EquiLibre’s approach focuses on extracting more performance from fewer chips — "get more from less" — but that strategy risks being outpaced by larger players with vast hardware resources.
Schmid’s stated ambition is for EquiLibre to become "the AI lab in trading," though he acknowledges the market may not be winner-takes-all and multiple firms can benefit from advances in AI-driven trading.
Local ecosystem
Prague is becoming a hub for AI startups; EquiLibre shares a building with other emerging companies such as BottleCap AI. The founders argue that the regional talent pool and a steadier local job market help retain staff compared with faster-moving hubs like San Francisco.
In short: EquiLibre Technologies has leveraged founders’ DeepMind pedigree and reinforcement learning techniques to enter quantitative trading, secured a $500 million valuation in a Creandum-led Series A, and now faces the twin challenges of scaling compute and competing with established trading firms that already deploy RL at scale.
