Two Palo Alto Startups Raise Hundreds of Millions to Build AI Systems for Mathematical Proofs

Axiom Math and Harmonic, two start-ups based a few doors apart in Palo Alto, California, have each raised hundreds of millions of dollars to develop AI systems that generate and verify mathematical proofs. Axiom Math was founded by Carina Hong, a former student of mathematician Ken Ono at Stanford University. Ono left his professorship at the University of Virginia in 2025 to join the company.

Harmonic is led by chief executive Tudor Achim. Both firms occupy nondescript offices in Palo Alto. Rooms at Axiom Math are named after mathematicians including Carl Friedrich Gauss and Ada Lovelace.

The reporter visited the companies in April. Axiom Math’s AI tools have produced five papers accepted in mathematical journals. The company aims to generate dozens of papers by next year. Its systems have generated checked proofs in algebraic geometry and number theory.

Harmonic’s tools have also produced checked proofs in the same fields. Achim said verification is becoming more valuable as AI writes increasing amounts of code. OpenAI chief scientist Jakub Pachocki said mathematics is useful for AI development because results are measurable.

Sébastien Bubeck, also at OpenAI, said recent models no longer produce nonsense in fields where earlier versions failed. The most recent AI models won gold medals at the International Mathematical Olympiad and disproved an 80-year-old conjecture. Epoch AI asked Ono to help create hard math problems to test AI systems.

Ravi Vakil at Stanford University said current funding levels will not last. He noted there is little commercial return from solving problems such as the Riemann hypothesis. Shubho Sengupta at Axiom Math said some mathematical modeling is already paywalled by hedge funds.

He added that advancing the bounds of mathematical knowledge should remain free. Achim said users may eventually pay for useful math tools. He stated the company will continue to support mathematicians if it views mathematics as central to future technology.

Ono compared the current moment to the arrival of Srinivasa Ramanujan in the early twentieth century. He recalled his father, a Japanese mathematician, discussing the need to embrace unexpected advances.