The headline number is $11 billion. The interesting part is what it is a bet on, and it is not the model.
What the valuation is actually pricing
Harvey sells AI to lawyers, the most risk-allergic, change-resistant buyer in the enterprise, and it convinced half the Am Law 100 to buy. It went from a 2022 experiment by two roommates, ex-litigator Winston Weinberg and ex-DeepMind researcher Gabriel Pereyra, to $100M ARR in three years, roughly $350M by mid-2026, and a $200M raise at $11 billion in March 2026, co-led by GIC and Sequoia. Over 142,000 lawyers use it.
That valuation is not a bet on a model edge. Harvey runs on outside frontier models and always has. It is a bet on vertical depth: the legal-specific workflows, the trust that lets a partner put their name on the output, the flagship co-development with Allen & Overy that opened the rest of Big Law, and the land-and-expand where median seats double in a year. In a domain where a wrong answer is malpractice, the moat is trust and distribution, not capability, and those are exactly the things a frontier lab cannot ship in a weekend.
It is the opposite answer to the same threat Cursor faced. Cursor went horizontal and got so exposed on model supply that it took a $60 billion acquisition to fix it. Harvey went so deep into one vertical that the model commoditizing underneath it barely matters. Own the lab, or own a domain so specific the lab will never go there. For the full breakdown, read the Harvey playbook and the Cursor teardown.