Photo by Behnam Norouzi on Unsplash
- EquiLibre Technologies closed a Series A at a valuation above $500 million as of June 30, 2026 — up from a $140 million seed valuation, a 257% jump in a single funding round.
- The company adapts the same reinforcement learning architecture behind DeepStack (2017) — the first AI to defeat professional poker players — to algorithmic trading across the S&P 500 and Nasdaq.
- Creandum led the round in what VP Cameron Sellers confirmed was the largest single investment the firm has ever made in one go; the company claims zero negative months since launching on crypto markets in 2025.
- The founding team's academic lineage runs through Rich Sutton, who received the 2024 ACM Turing Award for foundational reinforcement learning research and now sits on EquiLibre's advisory board.
What Just Happened
257%. That is the valuation markup EquiLibre Technologies secured between its seed round and the Series A it just closed, as reported by TechCrunch on June 30, 2026. The company — founded by Martin Schmid, Rudolf Kadlec, and Matej Moravcik, three researchers who built an AI that defeated professional poker players — is now valued at over $500 million after Creandum led the round. Creandum VP Cameron Sellers confirmed it was "the largest single investment the firm has ever made in one go into a company," a signal worth sitting with before reaching for any analyst's comparable set.
The origin story anchors the thesis. In 2017, the trio published DeepStack, the first AI system to defeat professional opponents at heads-up no-limit Texas hold'em. They were then visiting PhD students at DeepMind's Edmonton, Alberta office — the same office Alphabet shuttered in 2023 as part of a 12,000-person global layoff (roughly 6% of its workforce), confirmed by Global News Canada at the time. That closure accelerated the founders' transition from academic work to startup: CNBC's April 2022 coverage found the company already operational, suggesting the founders departed around January 2022. TechCrunch references a January departure without specifying the year — a minor gap in the public record that matters for anyone trying to reconstruct the founding timeline.
CEO Martin Schmid framed the pivot directly: "Our idea is that rather than playing poker, our algorithms will play algorithmic trading." As of June 30, 2026, those algorithms reportedly trade billions in daily volume across S&P 500 and Nasdaq indices through an exclusive partnership with Tower Research Capital — a relationship Bloomberg first covered in May 2024. That Bloomberg timing creates a notable divergence with TechCrunch's current reporting, which places the crypto market launch in 2025. Either the Bloomberg piece describes an earlier, narrower deployment and the 2025 figure marks a scaled expansion, or the record contains an inconsistency. For anyone conducting diligence on the company's performance timeline, that gap is worth resolving directly with the company.
The Pattern — Game Theory as a Trading Wedge
The underlying mechanism is reinforcement learning (RL) — a training paradigm where a model improves by receiving reward signals for correct decisions and penalty signals for wrong ones, iterated across millions of simulated scenarios. RL is what taught DeepStack to bluff, fold, and size bets optimally across hands a human would take years to accumulate. The translation to financial markets is not metaphorical. Both poker and algorithmic trading share a structural property that makes RL unusually well-suited: decision-making under incomplete information, where the agent must optimize against an environment that is simultaneously adapting to its behavior.
This is the playbook EquiLibre is running: identify a domain where RL has demonstrated superhuman performance at the academic level, locate a structurally analogous domain with institutional-scale commercial upside, and apply the same core architecture with domain-specific calibration. The same logic carried AlphaFold-derived research into drug discovery and LLM benchmarks into enterprise SaaS. What gives EquiLibre's particular instance unusual credibility is the advisory lineage. Rich Sutton — the trio's former professor — received the 2024 ACM Turing Award alongside Andrew Barto, per the ACM's official announcement, for developing the conceptual and algorithmic foundations of reinforcement learning itself. That is not a resume line; it is the field's highest credential sitting one degree of separation from the founding team.
The broader market trend EquiLibre is riding is documented: AI-powered funds generated 34% cumulative returns compared to 12% for the broader hedge fund industry between May 2017 and May 2020. The 22-percentage-point gap is what investors are pricing forward at the Series A level. For context on where AI-driven volume is concentrating today, the analysis at Smart Startup AI on the Nasdaq AI rally's second wind maps the same indices — S&P 500 and Nasdaq — where EquiLibre now operates at scale.
Chart: AI-led funds vs. broader hedge fund industry cumulative returns, May 2017–May 2020. Source: Market context data cited in research.
Photo by Nik Korba on Unsplash
The Valuation Sniff Test
A 257% markup between seed and Series A invites skepticism, and it should. The seed was $10 million at a $140 million valuation — already pricing in substantial future performance for a company with no disclosed revenue at the time. The Series A above $500 million still carries no public ARR figure. What is being valued is not a SaaS multiple; it is a combination of proprietary algorithm performance, the exclusive Tower Research distribution relationship, and structural defensibility of the RL approach itself.
That last point is the core argument for the premium. Quantitative alpha decays as competing capital pursues the same strategy. EquiLibre's defensibility claim rests on the nature of RL-generated strategies: the model's decision logic is not directly inspectable or rule-extractable in the way that a moving-average crossover strategy is. As the model continuously adapts to market microstructure, it becomes a moving target for reverse-engineering — a meaningful moat in a space where most edge commoditizes within 18 months of discovery. The zero-negative-months record since the 2025 crypto market launch, if independently verifiable, is the live proof point that anchors the thesis.
In my analysis, the valuation is aggressive but structurally justifiable — provided the Tower Research relationship is genuinely exclusive and the RL architecture continues adapting faster than competitors can replicate it. The risk that keeps me cautious is AUM scaling: as EquiLibre deploys larger capital, its own order flow becomes a market-moving signal, which creates a compounding information disadvantage that no RL model fully escapes at sufficient scale. That is the question I would put to the Series A lead before any follow-on consideration.
The Founder Move for Q3 2026
1. Source your wedge product from a published proof-of-concept. DeepStack was peer-reviewed, competition-tested, and publicly documented before EquiLibre existed. The startup commercialized a demonstrated capability, not a hypothesis. For founders translating AI research into products, a published result transforms your seed narrative from a technology bet into a translation bet — a meaningfully lower-risk framing for investors tracking AI investing tools and academic-to-commercial pipelines.
2. Treat the exclusive enterprise anchor as your go-to-market. EquiLibre did not launch a self-serve platform or pursue retail channels. It signed Tower Research Capital as both a distribution and validation partner, giving the technology immediate institutional-scale deployment and a credible third-party performance record. For AI-native startups operating in regulated or capital-intensive verticals — finance, healthcare, defense — one exclusive anchor partnership often unlocks Series A valuations that typical ICP-fit SaaS metrics would take three additional years to justify. Build that partnership before you raise, not after.
3. Recruit your academic advisors at founding, not post-funding. Rich Sutton's presence on EquiLibre's advisory board was part of the founding story. For founders building on deep technical foundations, a direct line to the originating researchers — or their close collaborators — is a credibility signal that no amount of accelerator branding or launch coverage replicates. In Q3 2026, as LPs increasingly scrutinize technical defensibility in AI rounds, that lineage is table stakes for a valuation north of $300 million. Identify who those advisors are for your domain and approach them in the research phase, not the fundraising phase.
Frequently Asked Questions
How does AI trading work for quant hedge funds using reinforcement learning?
Reinforcement learning (RL) trains algorithms by rewarding profitable trading actions and penalizing losses across millions of simulated and live scenarios. Unlike rule-based systems (which follow pre-coded instructions) or statistical models (which extrapolate from historical patterns), RL models adapt in real time to changing market conditions. EquiLibre's system draws directly from the same RL architecture behind DeepStack, applying it to the S&P 500 and Nasdaq through its Tower Research Capital partnership. The key advantage in markets, as in poker, is the model's ability to optimize decisions where it has incomplete information about other participants' intentions and positions.
Can AI hedge funds consistently beat the stock market over the long run?
Historical data shows a meaningful performance gap: AI-led funds generated 34% cumulative returns compared to 12% for the broader hedge fund industry between May 2017 and May 2020. EquiLibre claims zero negative months since launching on crypto markets in 2025. However, quantitative alpha tends to erode over time as competing strategies converge on the same signals, and no methodology has sustained indefinite outperformance across full market cycles. Evaluating claims like EquiLibre's zero-negative-months record requires understanding the market conditions during that period and the strategy's behavior under tail-risk scenarios. Nothing in this article constitutes financial advice regarding any investment portfolio allocation.
What is the performance of AI hedge funds compared to traditional funds?
The most widely cited data point covers May 2017 through May 2020: AI-powered funds averaged 34% cumulative returns versus 12% for the broader hedge fund industry during that window — a 22-percentage-point gap. More recent performance data for specific funds is generally proprietary and not publicly reported. EquiLibre has not disclosed AUM or audited return figures publicly as of June 30, 2026, beyond the zero-negative-months claim. Investors evaluating AI-native quant strategies for an investment portfolio should seek verified audited track records, not marketing-level performance claims.
How does poker AI strategy translate to algorithmic stock trading?
Both poker and financial markets are what game theorists call "imperfect information games" — environments where the agent must make optimal decisions without knowing the full state of the system. In poker, you cannot see your opponent's cards. In markets, you cannot see the full order book or other participants' intentions. Reinforcement learning excels in these environments because it learns strategies through repeated interaction rather than requiring complete state visibility. DeepStack (2017) proved RL could achieve superhuman performance in poker under exactly those conditions. EquiLibre's thesis is that the same architecture, trained on market microstructure instead of poker hand histories, can achieve a comparable edge in algorithmic trading — a claim its Tower Research partnership and track record are designed to validate.
Disclaimer: This article is for informational and editorial purposes only and does not constitute financial or investment advice. All valuations, performance figures, and market data cited reflect publicly available reporting as of June 30, 2026. Readers should conduct independent due diligence before making any financial planning or investment decisions. Research based on publicly available sources current as of June 30, 2026.