Axiom

What Axiom does

Axiom Math (legally Axiom Quant Inc.) is a Palo Alto-based AI research company building a self-improving AI mathematician on a formal-verification stack. Founded in early 2025 by Carina Hong, a Stanford math PhD dropout and Morgan Prize winner, the company develops systems that translate natural-language mathematical statements into Lean 4.26.0/Mathlib and produce machine-checked proofs. The platform has three interlocking products: AXLE, a managed Lean verification and manipulation API exposing primitives such as verify_proof and extract_theorems; AxiomProver, the core autonomous theorem prover that has scored a perfect 12/12 on the Putnam 2025 exam and autonomously settled multiple open research conjectures (Fel's, Partial Vandiver, Ramanujan-tau, lattice triangles); and Axplorer, an open-source generative tool based on a re-engineered PatternBoost loop for combinatorial discovery.

Axiom's underlying architecture pairs a Conjecturer that proposes out-of-distribution mathematical statements with a Prover that verifies them in Lean, generating training signal from successes, failures, and counterexamples — a self-reinforcing loop the company frames as the path toward mathematical superintelligence. The technical team includes CTO Shubho Sengupta (formerly Meta AI Research Director), Founding Mathematician Ken Ono, Discovery Lead François Charton (co-creator of PatternBoost), and Prover Lead Chris Cummins, with a 17-person team as of December 2025. Investors include B Capital, Menlo Ventures, Greycroft, Madrona, and Toyota Ventures.

Axiom is pre-revenue. AXLE is offered as a free public API and Playground with a separate "capacity interest form" gating production access; no published price points, paid tiers, or enterprise SKUs have been disclosed. The intended future monetization is Verified AI services for AI-generated code in hardware/chip design, software engineering, quantitative finance, and cryptography — domains where formal verification of AI output commands a premium. To date the company has raised approximately $264 million across a $64M seed (October 2025) and a $200M Series A (March 2026) at a $1.6 billion post-money valuation.

Name

Axiom

Legal name

Axiom Quant Inc.

Website

https://axiommath.ai

Company type

Private

Founded year

2025

Operating status

Operating

Headcount range

11–50 employees

Short description

Axiom Math (Palo Alto, founded 2025) builds an AI mathematician on Lean 4 formal verification, offering the AXLE verification API, the autonomous AxiomProver theorem prover, and the open-source Axplorer combinatorial discovery tool to researchers and, prospectively, to enterprise customers in hardware, software, quantitative finance, and cryptography.

Ownership category

akta.pro rank

akta.pro primary industry

Intelligent Practice & Mastery Learning Platforms (Spaced Repetition, Skill Drills) (EDAFAFAD)

akta.pro secondary industries

Mathematics Tutoring (Arithmetic–Calculus & Beyond) (EDANAAAA), Game-Based Learning & Gamification Authoring (EDAFAHAF)

Keywords

• AI theorem proving
• Formal verification
• Mathematical AI
• Automated reasoning
• Verified AI systems

Headquarters

HQ city

Palo Alto

HQ country

United States

HQ region

North America

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Infrastructure, Operations, Marketing or Sales

Revenue model

1. AXLE API access: Free public tier of the AXLE Lean-engine API with a separate capacity-interest form gating production/higher-volume access, suggesting a future usage-based commercial tier for verify_proof and extract_theorems calls.
2. Verified AI solutions for AI-generated code: Future-facing revenue targeting the broader market of all AI-generated code requiring formal verification, especially in hardware/chip design, software engineering, quantitative finance, and cryptography; today largely pre-commercial but framed as the company's primary monetization direction.
3. Open-source tooling (Axplorer): No direct revenue; open-sourced codebase functions as a developer-acquisition funnel and community-building mechanism rather than a commercial product line.

Pricing tiers

Go-to-market motion1 record

Core offering

Axiom builds self-improving AI mathematicians that solve complex mathematical problems and produce formally verified proofs using Lean 4 and Mathlib. The company offers AXLE, a managed API and infrastructure layer for proof verification, AxiomProver, an autonomous AI theorem prover, and Axplorer, an open-source discovery tool that lets users explore verified solutions to competition-level mathematics problems.

Product overview

Axiom Math operates as a single integrated platform for verified, formal mathematical AI built around three interlocking products. At the foundation, AXLE (Axiom Lean Engine) is the managed infrastructure layer — a public API and Playground that exposes proof verification and manipulation primitives (verify_proof, extract_theorems, etc.) so external systems can run Lean reasoning without managing the Lean runtime themselves. Layered on top, AxiomProver is the core AI mathematician: an autonomous theorem prover that takes natural-language mathematical statements, formalizes them into Lean 4.26.0/Mathlib, and produces machine-checked proofs — demonstrated by a perfect 12/12 Putnam 2025 result and autonomous proofs of multiple open research conjectures. Complementing the proof side, Axplorer is the open-source discovery tool (built on the PatternBoost generative-modeling technique) for generating novel mathematical constructions such as extremal graphs. Together, AXLE provides the verification substrate, AxiomProver delivers formal reasoning, and Axplorer powers discovery — with Verified AI applied downstream to hardware verification, software/code verification, quantitative finance, and cryptography.

Differentiator

Problem solved

Functional benefit

Brands

• AxiomProver: Autonomous AI theorem prover that produces formal Lean proofs to mathematical problems, achieving a perfect 12/12 score on the Putnam 2025 exam.
• AXLE (Axiom Lean Engine)
• Axplorer

Products and services

• AXLE (Axiom Lean Engine) Managed API and infrastructure for formally verified mathematical proofs, exposing AXLE capabilities to enterprise customers for chip design, software engineering, quantitative finance, and cryptography workloads.
• AxiomProver Autonomous AI theorem prover that solves advanced mathematics problems and produces complete formally verified proofs without human assistance. Targeted at research labs, formal methods teams, and enterprises requiring rigorous mathematical guarantees.
• Axplorer Open-source discovery tool that lets users browse and explore formally verified solutions to competition-level mathematical problems such as Putnam and IMO questions.

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

AI capability9 records

Feature6 records

Recent moves6 records

Expansion highlights5 records

Broad incumbents

• OpenAI: OpenAI's o-series reasoning models have demonstrated strong performance on math benchmarks (AIME, competition math) and the company has invested in code-generation tooling that intersects with software verification. As a frontier lab with massive model-training resources, OpenAI represents a broad threat to Axiom's positioning if reasoning capability generalizes to formal proof.
• Google DeepMind (AlphaProof): DeepMind's AlphaProof system achieved silver-medal-equivalent performance on the International Mathematical Olympiad using Lean-based formal reasoning — overlapping directly with AxiomProver's Lean 4/Mathlib stack. As a hyperscaler-backed research effort, DeepMind represents Axiom's most capable potential competitor in AI theorem proving at frontier scale.
• Anthropic: Anthropic develops Claude, which has been evaluated on formal-verification and reasoning benchmarks and which the company is positioning as a tool for software engineering and code correctness. Its focus on agentic workflows and verifiable outputs places it adjacent to Axiom's software-verification ambitions.
• Meta AI (FAIR): Meta FAIR is where Axiom's CTO, Discovery Lead, and several researchers previously worked and where PatternBoost was originally developed. FAIR's broader research portfolio includes Llemma and other math-focused foundation models, making Meta both Axiom's most prolific talent supplier and a strategic competitor with vastly greater compute resources.

Others

• Lean FRO / Mathlib community: The Lean theorem prover and its Mathlib library form the foundational open-source ecosystem on which AxiomProver and AXLE are built. While not a commercial competitor, the Lean FRO governs Axiom's underlying infrastructure and any strategic dependency on its roadmap directly affects Axiom's product velocity.
• DARPA / NSF funded formal-methods groups (Isabelle, Rocq): Academic and government-funded formal-verification efforts using Isabelle, Rocq/Coq, ACL2, and HOL4 represent the broader ecosystem of theorem-proving research that Axiom's commercial stack will compete with or partner against. NSF, DARPA, and university groups continue to push formal-method capabilities and shape the talent and toolchain landscape Axiom operates within.

Direct peers

• Imandra: Imandra provides formal verification and reasoning-as-a-service for finance and regulated industries, built on the Rocq/Coq theorem prover. It is one of the few commercial vendors in formal verification with paying enterprise customers, making it a direct analog for Axiom's downstream monetization ambitions in quantitative finance and cryptography.
• Harmonic: Harmonic is explicitly building 'mathematical superintelligence' using AI and formal methods, sharing Axiom's mission of automating advanced mathematical reasoning. The two companies are the most direct competitors in the AI-formal-math startup category, both pursuing foundation-model approaches to theorem proving and verification.
• AISLE (AI Safety via Lean): AISLE applies Lean-based formal verification to neural-network reasoning and AI safety use cases, sharing Axiom's Lean-first stack and theorem-proving orientation. As a smaller, focused peer, AISLE represents both a technical competitor in Lean-based reasoning and a potential partnership candidate in verified-AI safety applications.

Emerging players

• Cognition (Devin): Cognition Labs builds autonomous AI software engineers (Devin) and competes in the AI-driven code-generation market that Axiom aims to serve with verified-generation capabilities. As an emerging player in AI software development, Cognition represents a partial-overlap competitor for Axiom's downstream verified-code and software-verification ambitions.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat4 records

Key risks7 records

Key highlights7 records

Customer concentration

• Xx.com/axiommathai
• GitHubgithub.com/AxiomMath
• LinkedInlinkedin.com/company/axiommath

Revenue estimate

Valuation estimate

Number of profiles

Profiles7 records

Funding overview

Funding rounds3 records

Investors8 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Axiom