Axiom Math’s AI System Solves Unsolved Problems and Advances Automated Proofs
At a glance
- Axiom Math was founded in March 2025 by Carina Hong
- The startup raised $64 million in seed funding in 2025
- AxiomProver solved four previously unsolved math problems
- Ken Ono joined Axiom Math as founding mathematician
Recent developments in AI-driven mathematics have been marked by Axiom Math’s achievements in automated problem solving and proof generation. The company’s AI system, AxiomProver, has addressed several longstanding mathematical challenges using formal verification tools.
Axiom Math was established in March 2025 by Carina Hong, who previously studied at Stanford and was recognized as a Rhodes Scholar. The startup secured $64 million in seed funding during 2025, with B Capital leading the investment round and additional participation from Greycroft, Madrona, and Menlo Ventures.
Ken Ono, who previously held a tenured academic position at the University of Virginia, joined Axiom Math as its founding mathematician. His move to the startup brought additional expertise to the company’s focus on mathematical discovery using artificial intelligence.
The company’s AI system, AxiomProver, uses the Lean proof assistant to both generate and verify mathematical proofs. This approach allows for rigorous checking of solutions and supports the automation of complex mathematical reasoning.
What the numbers show
- Axiom Math was founded in March 2025
- $64 million in seed funding was raised in 2025
- AxiomProver solved all 12 problems from the 2025 Putnam Competition
- Four previously unsolved mathematical problems were addressed by AxiomProver
AxiomProver has solved four mathematical problems that had not previously been resolved. These included a conjecture by Dawei Chen and Quentin Gendron, Fel’s Conjecture involving Ramanujan’s formulas, a probabilistic number-theory problem referred to as “dead ends,” and a problem that utilized techniques from Fermat’s Last Theorem.
The proof for the Chen-Gendron conjecture was made publicly available by posting it to the arXiv repository. This step provided the mathematical community with access to the AI-generated proof for further review and verification.
In addition to addressing unsolved problems, AxiomProver produced Lean-verified proofs for all 12 problems from the 2025 Putnam Competition. These results were shared through Axiom’s blog and discussed in online forums.
Axiom Math’s use of AI and formal verification tools demonstrates a procedural approach to mathematical discovery. The company’s activities have centered on generating and verifying solutions to recognized mathematical challenges using automated systems.
* This article is based on publicly available information at the time of writing.
More on Technology
-
Smartphones Share User Data Overnight Despite Idle Status
Smartphones transmit system and tracking data overnight while idle, including device IDs and location signals, raising privacy concerns for users.
-
House of Lords Supports Ban on Smartphones in UK Schools
A recent amendment proposes a smartphone ban during school hours, according to reports. The decision faced opposition from the government.
-
Project Vault Launches as US Expands Critical Minerals Strategy
The US initiated Project Vault, a $12 billion minerals stockpile effort, with $10 billion financing from the Export-Import Bank, according to reports.