Axiom, a startup that builds AI systems able to mathematically verify software and scientific reasoning, announced on 12 March 2026 that it had raised $200m in Series A funding. Menlo Ventures led the round, which SiliconANGLE reported values the company at $1.6bn. Pulse 2.0 described the valuation as exceeding $1.6bn.
Existing investors Madrona, Greycroft, B Capital, Toyota Ventures and Triatomic Capital also took part, according to Pulse 2.0. SiliconANGLE gave the company's legal name as Axiom Quant Inc.
The company first drew attention in October 2025 when it raised $64m in seed funding, SiliconANGLE reported.
“LLMs are statistical by nature — they produce plausible outputs, not provably correct or safe ones”Matt Kraning and C.C. Gong, Partners, Menlo Ventures
What Axiom does
Axiom calls its approach Verified AI. SiliconANGLE explained that the company trains AI systems to generate outputs that are formally verified in Lean, a programming language designed for mathematical proofs, so that each step of a model's reasoning is machine-checkable. Deterministic proof verifiers flag wrong outputs, which Axiom says can give mathematical certainty that a generated function returns the correct answer and does not add hidden vulnerabilities.
The pitch is aimed at AI-generated code. Matt Kraning and C.C. Gong, partners at Menlo Ventures, wrote in a blog post cited by SiliconANGLE that code that "frequently works" is a "terrifying standard" for critical infrastructure. They argued that hallucinations and unsafe code from large language models are architectural problems, not bugs that the next model generation will fix.
Pulse 2.0 noted that Axiom sees the same two-step process of hypothesis and proof in science generally, and wants to automate it for software and other technical fields.
Milestones and team
SiliconANGLE reported that in December 2025 Axiom's system achieved a perfect score on the Putnam Competition, a demanding undergraduate mathematics exam, and that it verifiably proved a 20-year-old number theory conjecture that its founding mathematician Ken Ono had been unable to solve.
The company was founded by Carina Hong, a Stanford PhD student and MIT graduate, who is its chief executive. The chief technology officer is Shubho Sengupta, formerly a research director at Facebook AI Research, and the team includes Ono and François Charton, according to SiliconANGLE.
What the money is for
Axiom said the new capital will expand its team and accelerate development of Verified AI. SiliconANGLE said the next step is to scale up training infrastructure and grow its team of math experts, with the aim of making formal verification fast and affordable for every company using AI.
Holger Mueller of Constellation Research told SiliconANGLE that the risk of models hallucinating is dangerous when it comes to code, and that the challenge now is to scale the solution to meet the demands of the coding world. The sources reviewed did not report headcount or revenue, and Axiom's headquarters is not stated in them.
Company website: axiommath.ai