Harmonic, a Palo Alto-based artificial intelligence startup backed by Robinhood CEO Vlad Tenev, said it raised $120 million in a Series C funding round that values the company at $1.45 billion post-money.
The round was led by Ribbit Capital, with significant participation from Sequoia Capital, Index Ventures and Kleiner Perkins. Emerson Collective joined as a new investor, the company said.
Harmonic, founded in 2023 and led by CEO Tudor Achim, is positioning itself around what it calls “Mathematical Superintelligence” (MSI), an approach designed to produce answers that can be checked rather than merely trusted.
At the center of that pitch is Aristotle, Harmonic’s mathematical reasoning model.
Unlike many general-purpose chatbots that can generate convincing but incorrect statements, Harmonic says Aristotle relies on formal verification in Lean4, a proof assistant used in academic and software-verification communities.
The goal, Achim said, is to eliminate hallucinations by forcing the system to express reasoning as verifiable code, not just natural language.
Harmonic said Aristotle recently reached “gold-medal level” performance on International Mathematical Olympiad problems and has made the model available to the public through an API.
The company said mathematicians and researchers have used early access to accelerate proof-checking and explore new results, though it did not disclose user counts or revenue.
Last week, Harmonic said it rolled out upgrades including plain-English input in addition to native Lean4, automated lemma generation and a simplified terminal interface, aiming to lower the barrier for non-experts.
Harmonic previously raised $75 million in a Series A in September 2024 and $100 million in a Series B in July 2025. Additional backers include Paradigm and Era Funds.
The financing underscores a shift in AI priorities from raw scale to reliability, with verification emerging as a potential differentiator for high-stakes fields like finance, engineering and scientific research.
Still, formal proof systems work best in tightly specified domains, and Harmonic will need to prove it can translate mathematical rigor into durable, paid adoption while larger AI labs pursue their own reasoning and safety approaches.