AI
Harvey Just Hit 15.5 Billion Dollars, and Law Firms Are Training Their Own Models Now
The most telling number in AI this month is a valuation: $15.5 billion, reached by legal AI startup Harvey on September 9 with a $550 million raise led by Diffusion and Lightspeed Venture Partners. That is nearly double the $11 billion valuation from its $200 million round in March, which itself followed an $8 billion round in December. Harvey has now raised more than $1.55 billion in total, and annual recurring revenue sits above $400 million.
The customer list explains the price tag. Harvey says 80 percent of the Am Law 100 firms use its platform, alongside in house legal teams including five of the Fortune 10. Its customer count has grown from roughly 1,300 organizations in March to more than 3,000. The platform automates contract analysis, due diligence, compliance, and litigation work, and its Vault repository lets firms store up to 100,000 documents with search driven by AI across them.
Here is the part that matters more than the money. A couple of weeks before the raise, Harvey shipped Tenet, its first in house model, built from the open weight Kimi K3 model and post trained on legal data with help from inference provider Fireworks. Harvey is now encouraging its customers to adopt and post train their own open weight models. As TechCrunch put it, an entire industry is learning to use AI heavily while relying less on proprietary frontier labs like OpenAI and Anthropic. The company's own press release says the funding will help firms 'build and own their intelligence at scale.' Own is the key word.
Harvey is moving fast on every front. The raise came alongside its fourth acquisition of 2026, Guardrails AI, an AI security platform for agents whose team now joins Harvey's product and engineering organization. Co founder Winston Weinberg framed the logic in the acquisition announcement: 'Every firm we work with asks the same question before they let an agent near real client work: How do you know what it will do?' Putting safety know how underneath every shipped agent is the answer. The company also launched Harvey LAB, a Legal Agent Benchmark, to measure agent performance in legal work.
The broader legal AI market is heating up alongside it. Legora raised $550 million at a $5.55 billion valuation in March, and Clio raised $500 million at a $5 billion valuation in November. A new venture fund backed by corporate general counsels and the law firm Wilson Sonsini launched this month to invest in early stage legal, regulatory, and compliance tech. Capital is flooding the vertical because the template works: deep domain data plus post training beats generic models at expert work.
For readers outside the legal world, Harvey is a preview of what comes for every knowledge profession. The winning pattern is ownership of a model shaped by your own data, your own documents, and your own judgment. The moat is moving from who has the biggest model to who has the richest expertise to teach one. Your profession's Harvey moment is coming, and the firms that build early will set the terms.
Sources
New to crypto? Read the crypto glossary, browse frequent questions, read our story, or explore the story archive.