EquiLibre Technologies Valuation: $500M+ Series A and Creandum's Biggest-Ever Check

EquiLibre Technologies just closed a Series A at a valuation above $500 million — Creandum's largest single check ever, and more than 3.5x the company's $140 million seed valuation from nine months earlier. That's the short answer. The longer answer is that a 14-person-founded, three-person-led Prague startup is now pricing itself against Numerai, a decade-old AI hedge fund with a public token and a JPMorgan check.

EquiLibre was founded in January 2022 by Martin Schmid, Rudolf Kadlec, and Matej Moravcik, three former DeepMind and IBM researchers who spent years before that building DeepStack, one of the first AI systems to beat professional poker players heads-up. The pitch is straightforward on paper and hard to execute in practice: apply reinforcement learning and game theory — the same toolkit that solved imperfect-information poker — to financial markets, where the "opponents" are other trading algorithms and the stakes are billions of dollars a day. The Series A, reported in early July 2026, is the clearest signal yet that investors believe that bet is working.

EquiLibre Technologies Valuation: What the Series A Actually Values the Company At

EquiLibre Technologies is valued at more than $500 million (roughly €438 million) following a Series A round led by Creandum, announced in early July 2026. Creandum has called it the largest single investment the firm has made into any one company, a notable signal given the firm's history backing Spotify, Klarna, and Small Giant Games. The round follows a $10 million seed raise led by Blossom Capital roughly nine months earlier, which had valued EquiLibre at just $140 million — meaning the Series A represents better than a 3.5x markup in under a year.

Existing seed backers Blossom Capital and Credo Ventures — an early investor in ElevenLabs and UiPath — both returned for the Series A, alongside new investor Miton. EquiLibre's total disclosed funding now stands at $53.3 million across pre-seed, seed, and Series A, per PitchBook data, a relatively capital-efficient path to a half-billion-dollar valuation compared to most AI infrastructure startups raising nine-figure rounds at similar marks.

Creandum's own history makes the size of this check more meaningful than a headline number alone. The Stockholm-founded firm has backed Spotify, Klarna, and Small Giant Games across two decades of European venture investing, and typically writes Series A and B checks in the low-to-mid eight figures rather than concentrating capital into a single position. Choosing EquiLibre for its largest-ever check, ahead of any consumer or enterprise SaaS deal currently in its pipeline, signals that Creandum's partners see a faster and more direct path to outsized returns in AI-native trading than in most categories the firm has historically prioritized.

EquiLibre vs. Numerai: Two Very Different AI Trading Models

The closest public comparable for EquiLibre's valuation is Numerai, the decade-old AI-powered hedge fund that also sits at roughly $500 million after a Series C round that included up to $500 million in backing from JPMorgan. The two companies reached similar valuations through almost opposite models: Numerai crowdsources signal from thousands of outside data scientists worldwide into a blended "Meta Model" that drives its equity strategies, while EquiLibre keeps its reinforcement-learning agents built entirely in-house by its three founders and a growing research team.

Figures are 2026 estimates blended from PitchBook, Tracxn, Bitget News, and company disclosures. Numerai AUM and performance figures per its public fund reporting; EquiLibre figures per Tech Funding News and The Recursive, checked July 2026.

What EquiLibre's AI Trading Agents Actually Do

EquiLibre's agents apply reinforcement learning and game-theoretic optimization — the same class of techniques used in DeepStack, the poker AI its founders built years before starting the company — to execute trades across the S&P 500 and Nasdaq. Rather than predicting prices directly, the approach treats markets as a multi-agent game where the system learns a strategy that performs well against other adaptive participants, including rival algorithms, which is conceptually closer to how AI solved poker than to how most quant funds have historically built statistical arbitrage models.

Since late 2023, EquiLibre has held an exclusive agreement to supply its trading algorithms to Tower Research Capital, a prominent New York proprietary trading firm — giving the startup an established, revenue-generating distribution channel well before this Series A, rather than the pre-revenue profile typical of most $500M-valued AI companies. Most of the new capital is earmarked for compute: EquiLibre has said it plans to build one of the largest AI compute clusters in Central Europe, a signal that the next phase of competition in this category is as much about training infrastructure as it is about trading strategy.

Why AI Trading Agents Are Attracting Venture Capital in 2026

EquiLibre's round lands inside a broader 2026 pattern: AI startups raised roughly $202 billion in 2025, nearly half of all global venture capital, and financial services has become one of the categories where investors are most comfortable paying growth-stage prices for early-stage companies, because the revenue path is direct rather than speculative. Unlike a foundation-model lab burning cash on training runs with no near-term monetization, an AI trading startup can point to actual assets under management or licensing revenue almost immediately, which is exactly the argument Creandum is underwriting with its largest-ever check.

That said, the category carries a structural risk generalist AI investors sometimes underweight: trading performance is adversarial and non-stationary — a strategy that works today can decay the moment competitors adapt or market regimes shift — which makes durability of edge much harder to underwrite than durability of a software moat. Track how AI-native valuations are trending across categories on our AI valuations dashboard, and see how EquiLibre's step-up compares to other 2026 AI rounds in our AI funding surge breakdown.

What the EquiLibre Round Means for Investors

For VCs and family offices watching the AI-in-finance category, the most important signal in EquiLibre's round isn't the headline valuation — it's that Creandum, a firm with a two-decade track record and Spotify and Klarna on its résumé, chose this as its single largest bet ever, ahead of any consumer or enterprise SaaS opportunity currently in its pipeline. That's a strong endorsement of the thesis that reinforcement-learning-native teams, not repurposed LLM wrappers, will capture the most durable value in AI-driven trading.

The comparison to Numerai is instructive for allocators building a category thesis rather than picking one company: both firms reached roughly $500 million valuations, but through different distribution models — Numerai's crowdsourced Meta Model versus EquiLibre's proprietary agents licensed to an established prop shop. That suggests the category can support more than one winning architecture, at least at this stage, which is a more favorable setup for follow-on investors than a single-winner-take-most market structure would be.

The open question is durability. EquiLibre hasn't disclosed a 2025-2026 performance figure the way Numerai has with its 25.45% net gain, and trading-agent moats are inherently harder to diligence than SaaS retention metrics — reinforcement learning that beats the market today can be arbitraged away as competitors train similar systems. Investors underwriting this category should weight the Tower Research Capital relationship and the planned compute buildout as evidence of a real, compounding technical edge, not just a growth-stage AI valuation riding the broader 2026 funding wave. Compare EquiLibre's pricing against other pre-revenue and early-revenue AI companies on our AI valuations dashboard.

There's also a talent-market signal worth watching: EquiLibre's three co-founders came directly out of DeepMind's research organization, the same talent pool feeding foundation-model labs, robotics startups, and now quantitative trading shops all at once. As reinforcement-learning researchers become one of the most contested hiring categories in AI — alongside frontier-model pretraining and agentic-systems engineers — expect more spinouts applying the same core techniques to adjacent high-stakes, adversarial domains: fraud detection, algorithmic procurement, and automated negotiation are the next logical targets for teams with this specific research background.