Algorithmic trading startup EquiLibre Technologies has closed a Series A funding round valuing the firm at $500 million (10,6 miliardy korun), according to co-founder Martin Schmid. The Prague-based firm, founded in 2022 by Martin Schmid, Rudolf Kadlec, and Matěj Moravčík, secured the investment to purchase thousands of graphics processing units to train proprietary financial models for global stock exchanges like NYSE and Nasdaq.
Series A Valuation and Nordic Backing
The investment round was led by Nordic venture capital fund Creandum, known for backing companies like Spotify, Bolt, Klarna, and Lovable. According to Creandum, the financing marks the firm’s historically largest single investment in a single company. Earlier pre-seed funding for EquiLibre came from domestic Czech fund Credo Ventures and individual investors. With a $500 million valuation, the startup now ranks among the most valuable young companies in the Czech Republic alongside online prodejce potravin Rohlík.
Did you know? EquiLibre’s founders worked at Google DeepMind, the elite artificial intelligence research laboratory owned by Alphabet, where they spent years before returning to Prague to launch the startup.
Hardware Scaling and Reinforcement Learning Strategy
EquiLibre plans to deploy the capital immediately into computing power rather than waiting years to fund hardware purchases from operational profits. “That hardware costs a lot of money, and we would have to earn money for a year or two before we could afford to buy it,” Schmid told Seznam Zprávy. By taking on investors, the company can buy equipment immediately, expecting the investment to pay off within two years. The startup utilizes reinforcement learning, a foundational branch of machine learning, to execute trades through a partnership with Tower Research Capital that sees the firm’s algorithms handle billions of dollars daily.
Financial Models Versus Large Language Models
Unlike generative artificial intelligence companies such as OpenAI or Anthropic, which face massive inference costs running text-based large language models for millions of users, EquiLibre deploys its models exclusively for internal trading. “We train our model, which is expensive, but then we deploy it on one exchange, so the inference is cheap because we don’t have thousands and thousands of customers,” Schmid explained. The company feeds publicly available historical market data—accessible to established competitors like Citadel and Optiver—into its systems.
Agility and Market Expansion Plans
Operating with a lean team of fewer than 30 people, EquiLibre builds its software completely from scratch rather than adapting legacy technologies used by older institutions. “Instead of us trying to understand the market ourselves, we are building models that try to understand it,” Schmid said, comparing the approach to building a system that plays chess rather than hiring a hundred chess players. Following its success on major American markets and profitability achieved at the beginning of last year, the firm aims to expand its algorithmic trading operations into Europe, Asia, and other global regions.
Frequently Asked Questions
What is EquiLibre Technologies’ valuation?
EquiLibre Technologies is valued at $500 million following its Series A funding round led by Creandum.
Who founded EquiLibre Technologies?
The startup was founded in 2022 by Martin Schmid, Rudolf Kadlec, and Matěj Moravčík, who previously worked at Google DeepMind.
What technology does EquiLibre use for trading?
The firm uses reinforcement learning to train proprietary financial models that trade on major global exchanges like NYSE and Nasdaq.
How does EquiLibre differ from AI companies like OpenAI?
Unlike large language model providers that incur high inference costs serving millions of users, EquiLibre deploys its models exclusively for its own proprietary trading.
Want to stay updated on the latest developments in European artificial intelligence and financial technology? Subscribe to our weekly newsletter or explore our latest industry reports.
Worth a look