Archive — US — techcrunch
The DeepMind trio who built a poker AI are now making money for quant hedgefunds
Former DeepMind researchers launched EquiLibre Technologies to apply poker AI technology to stock trading, achieving a $500 million valuation after a Series A round. Their AI, optimized for quant trading, has traded billions daily with no losses so far. Located in Prague, they aim to reduce competition risks while leading in reinforcement learning for finance.
Key points
- Three former DeepMind researchers founded EquiLibre Technologies after developing a poker AI called DeepStack.
- EquiLibre's AI applies reinforcement learning to stock trading, showing promising initial results in crypto markets in 2025.
- The company secured a $500 million valuation from Series A funding, the largest investment from Creandum at once.
- By leveraging Prague's talent pool and a defined lab-first approach, the startup differentiates itself from finance-driven firms.
- EquiLibre's AI has a record of zero negative months, indicating successful trading performance.
- Competition exists from established firms like Jane Street, which already utilize RL models.
Why it matters
- The trend of applying AI and machine learning technology to finance aligns with the 2025 IT trend of increased automation in financial markets.
- Reinforcement learning gaining traction in finance parallels advancements in AI across various sectors.
- The rise of AI-focused startups ties into broader investment trends in frontier AI technologies often led by DeepMind alumni.
EquiLibre Technologies exemplifies the evolving intersection of AI and finance, leveraging advanced learning techniques to drive profitability while navigating competitive pressures and positioning itself as a key player in the quant trading landscape.