Harvey, a San Francisco AI startup focused on the legal industry, has raised $300 million in a funding round led by Sequoia that values the startup at $3 billion — double the amount investors valued it at in July.
The Series D funding round builds on the momentum of the three-year old startup, which was founded by a former lawyer and a Google AI researcher, and reflects investors’ enthusiasm for AI tools that have succeeded in making inroads with business customers.
In an exclusive interview with Coins2Day, Harvey CEO Winston Weinberg said the startup has surpassed $50 million in annual recurring revenue and estimated it would surpass $100 million ARR in about eight months. In 2024, the company said it expanded from 40 customers to 235 customers in 42 countries, including the majority of top 10 U.S. Law firms.
Harvey uses large language models to help law firms and corporate legal teams streamline research, analyze contracts, and draft documents. “Legal work is incredibly complex and requires so much context, that a simple chatbot doesn’t work,” said Weinberg, who described his startup as developing a tool that offers suggestions and clarifications along the way, and allows lawyers to provide feedback about what they want.
Other investors in the company’s latest funding round include Coatue, Kleiner Perkins, OpenAI Startup Fund, GV, Conviction, entrepreneur Elad Gil, and REV, the venture capital arm of RELX Group which owns LexisNexis.
The market for AI apps in the legal industry has become crowded, with competitors including Casetext (which was acquired by Thomson Reuters), Andreessen-backed Eve, Spellbook, Robin AI, and Lex Machina (owned by Lexis Nexis). However, Weinberg argues that Harvey’s focus on hiring lawyers as domain experts and integrating them with Harvey engineers, has helped them get the most out of the AI models they use (which besides OpenAI include Anthropic’s Claude and models from Mistral and Xai).
“We have tons of lawyers on staff designing and evaluating the product, and they’re all from large law firms or from large in-house teams,” he said. “They’re literally telling our engineers, like, we need to do that section and that section and that section and that section, they’re explaining the process of how to actually create different work products.”
A lawyer and a scientist team up
Weinberg’s own inspiration for Harvey goes back to his experience a first-year associate at Los Angeles law firm O’Melveny in 2022. He teamed up with Gabe Pereyra, a former research scientist at Meta and DeepMind, to test OpenAI’s GPT-3, which was available to the public. The duo emailed OpenAI and shared a new chatbot they had developed using GPT-3 to answer legal questions pulled from Reddit. They also sent it to OpenAI’s general counsel at the time, because they knew a lawyer could understand the chatbot’s outputs and how good they were.
That led to a July 4 meeting with the C-Suite at OpenAI, who quickly made their first OpenAI Startup Fund investment in Harvey. The following April, Harvey raised $21 million in a Series A led by Sequoia Capital.
Pat Grady, partner at Sequoia Capital, told Coins2Day that targeting the legal market with generative AI “kind of felt like a bulls-eye.” For one thing, the industry is big — a $400 billion market in the U.S. Alone. In addition, it felt like a good fit for the technology.
“A lot of what happens in the legal world is text in, text out, and it’s reasoning over that text,” he said.
Read more from Coins2Day
However, after OpenAI’s ChatGPT launched in November 2022, many critics dismissed Harvey-like startups as mere “GPT-wrappers”: That is, those that did not develop their own AI models but used OpenAI’s GPT-3, and then GPT-4 and beyond, to power their applications, would not last.
“We have been delighted to let that narrative persist,” said Sequoia’s Grady. “We’ve been loading up on application companies.”
Grady cited the Weinberg—Pereyra duo as an important reason for Sequoia’s bullishness in the startup. “In the case of Winston, the superpowers that make you effective as a founder are not exactly the same as the superpowers that make you effective as a CEO,” he said. “We’ve really seen a transformation in the way that he operates, and in the way that he thinks about building teams, and the way that he thinks about this integrated system that is the company.” Seeing Weinberg grow into his CEO role, he added, gave Sequoia the confidence that “the machine is in place that will keep this company going for many years to come.”
