AI verification startup Axiom Math announced Wednesday that its network of specialized AI agents has successfully translated and verified the Stacks Project, an iconic 7,654-page open-source mathematics repository. Founded by Columbia mathematician Johan de Jong in 2005, the project covers advanced algebra and geometry, which Axiom translated into 16 million lines of machine-checked Lean programming code.
The milestone aims to establish a machine-verified mathematical foundation that future AI agents can rely on to write software and execute complex scientific research without introducing errors or hallucinations. The startup deployed parallel AI agents alongside its own specialized models to formalize the highly interdependent theorems while preserving the logical structure of the library.
Founded in March 2025 by CEO Carina Hong, Axiom Math has raised $200 million in a Series A funding round led by Menlo Ventures, valuing the company at $1.6 billion. The company intends to use formal verification methods—long employed in hardware design—to ground advanced AI reasoning in provably correct logic.
Why it matters
Validates formal verification using automated agents to prevent AI hallucinations in high-stakes reasoning applications.
Provides a deterministic benchmark and logical foundation for next-generation scientific and software engineering models.
Demonstrates commercial investor appetite for mathematical correctness platforms, backed by a $1.6B valuation.
Source: techcrunch.com



