A new AI math startup called Axiom has reportedly cracked four long-standing mathematical problems, demonstrating advances in artificial intelligence reasoning. The company's AI addressed challenges in areas like algebraic geometry and number theory that had puzzled mathematicians for years. This development highlights the growing capabilities of AI in tackling complex academic puzzles.
In a recent announcement, Axiom, an AI-focused startup, claimed that its technology has solved four previously unsolved problems in mathematics. These include issues in algebraic geometry, calculus, and number theory, fields that have long challenged researchers.
One specific example involves work from five years ago by mathematicians Dawei Chen and Quentin Gendron. They were exploring a complex area of algebraic geometry that incorporates differentials—calculus elements used to measure distances on curved surfaces. During their efforts on a particular theorem, they encountered a barrier: their proof relied on an obscure formula from number theory that they could not resolve or explain. As a result, Chen and Gendron published their findings as a conjecture rather than a full theorem.
Axiom's AI has now provided solutions to such conjectures, suggesting a breakthrough in automated reasoning for pure mathematics. The startup's achievement points to AI's potential to accelerate discoveries in theoretical fields, where human intuition has traditionally dominated. However, experts caution that while promising, these AI-generated proofs will require rigorous verification by the mathematical community to confirm their validity.
This event underscores the intersection of artificial intelligence and academia, with implications for how future research might integrate machine assistance. Published on February 4, 2026, the news has sparked interest in AI labs pushing boundaries in STEM disciplines.