The round
Harvey, which sells AI tools to lawyers, has raised $550m at a $15.5bn valuation, the company announced on 9 September. Diffusion and Lightspeed Venture Partners co-led the round.
Existing investors Sequoia, Kleiner Perkins, a16z, Coatue, Conviction, Elad Gil, Evantic, GIC, Goldman Sachs Alternatives, Verified Capital and WNDR also took part, Harvey said, with Sapphire Ventures and Whale Rock joining as new investors.
The valuation has climbed quickly. Harvey was valued at $8bn in December 2025 and at $11bn in March, TechCrunch reported, which means it has nearly doubled in about nine months. The new money takes the total Harvey has raised past $1.55bn, and TechCrunch counted at least eight priced rounds since 2023, five of them since the start of 2025.
Who uses it
Harvey says 80% of Am Law 100 firms, the largest US law firms by revenue, use its platform, along with in-house legal teams at five Fortune 10 companies. Those are the company’s own figures.
In a statement, co-founders Winston Weinberg and Gabe Pereyra said “the opportunity for companies to accelerate their competitive advantage with AI has never been higher.” The company said it will use the money to expand hiring and build out the teams that support clients.
Its own model, from an open-weight base
The round follows two product moves Harvey highlighted in the announcement. The first is Tenet, which Harvey calls its first post-trained open-weight model. TechCrunch reported that Tenet was built from Kimi K3, the open-weight model released by China’s Moonshot AI in July, and post-trained on legal data with help from the inference provider Fireworks.
The second is Harvey LAB, a Legal Agent Benchmark the company launched to measure how well AI agents handle legal work. As a benchmark built and published by a vendor, its results should be read as Harvey’s own evaluation rather than an independent one.
Tenet is the more significant of the two for the wider industry. Harvey built its business on top of frontier models from labs such as OpenAI. Moving to a model it post-trains itself, from an open-weight base, gives it more control over cost and behaviour. TechCrunch argued that Harvey shows how an industry can use AI heavily without relying on proprietary frontier labs.
Why it matters
Legal work has become one of the clearest early markets for generative AI: document-heavy, billed by the hour, and full of tasks that models can draft and humans can check. A $15.5bn valuation prices Harvey as a company that will own a large share of that market, not just resell access to someone else’s model.
The next thing to watch is how much of Harvey’s product actually runs on Tenet rather than on frontier models, and whether rivals in legal AI follow it towards post-training their own open-weight models.
