AI Startup Axiom Math has Emerged from Stealth with $64 Million Seed Funding Round

Axiom Math, building a self-improving superintelligent reasoner, starting with an AI mathematician, led by Carina Hong, Shubho Sengupta, and others, has emerged from stealth mode and announced a $64 million seed fund round led by B Capital and participation from Greycroft, Madrona Venture Group, and Menlo Ventures, valuing the company at $300 million.

Axiom is set to expand engineering and research talent, enhance reasoning engines, and test its system on benchmark problems across cryptography, physics, and advanced algorithms.

Led by Carina Hong, Axiom Math is building an AI system that solves complex math problems and generates new knowledge by proposing unproven conjectures. The model produces rigorous, step-by-step proofs verifiable with tools like Lean and Coq, translating textbook and research math into code to create and validate novel problems.

The team includes experts from Meta’s FAIR lab and notable mathematicians such as Francois Charton, Aram Markosyan, and Hugh Leather, combining AI and mathematical expertise to push the boundaries of discovery.

Axiom Math is developing AI models capable of discovering and solving new mathematical problems, with potential applications in finance, aircraft and chip design, and quantitative trading. The company emphasizes foundational scientific progress, reflected even in its Palo Alto office, where conference rooms honor mathematicians like Gauss and Lovelace.

By combining ambitious vision, deep expertise, and practical applications, Axiom is positioning AI to drive breakthroughs in both scientific discovery and industry innovation.

Mathematics has always been limited by scarce genius and siloed knowledge. Axiom is building an AI mathematician that discovers and proves theorems at superhuman speed. By translating intuition into formal proofs and back, it amplifies human reasoning while generating insights humans could never see.

Formal verification ensures every result is correct, while a conjecture-proving engine explores uncharted territory in a self-reinforcing loop: more proofs fuel better conjectures, faster discoveries, and deeper understanding.

This system isn’t just for math, it’s a general-purpose tool for modeling reality, from quantum physics to protein folding, chip design to economics. Axiom is accelerating a long-overdue mathematical renaissance.