FT: How a former junior lawyer created a $5bn AI legal start-up

Key topics:Harvey grows to 500+ clients with $100mn annual recurring revenueLegal AI start-up valued at $5bn, backed by Sequoia, OpenAI, and Kleiner PerkinsUses LLMs to speed legal research, document review, and compliance tasks.Sign up for your early morning brew of the BizNews Insider to keep you up to speed with the content that matters. The newsletter will land in your inbox at 5:30am weekdays. Register here.Support South Africa’s bastion of independent journalism, offering balanced insights on investments, business, and the political economy, by joining BizNews Premium. Register here.If you prefer WhatsApp for updates, sign up for the BizNews channel here..By Suzi Ring in London.When Winston Weinberg, co-founder of legal artificial intelligence company Harvey, turned 30 last year, he marked the occasion with three work calls and a walk on the beach.While many young people would resent the lack of celebration, he said founders should be “willing to outcare and outwork other people”, adding: “In the AI space, we are on a compressed timeline, so what seemed fast execution-wise two years ago is simply not good enough now.”Harvey was founded in 2022 as the brainchild of Weinberg, once a junior lawyer, and his friend and former DeepMind research scientist Gabe Pereyra. It has grown from one client to more than 500 over the past three years, spanning some of the world’s largest law firms and companies, with annual recurring revenue of $100mn.Harvey was valued at $5bn after its last funding round in June, co-led by US venture capital firm Kleiner Perkins and investment firm Coatue, with funds also coming from OpenAI and Sequoia.Like many industries, AI has become a buzzword in the legal sector as lawyers and technology specialists spy a chance to eliminate some of the process-driven aspects of the job. However, sensitivities around data coupled with lawyers’ risk aversion has slowed meaningful changes to the workforce — or client bills..Read more:.OpenAI launches ‘Strawberry’ AI model for advanced reasoning.Harvey is one of the few operators that has punched through after law firm A&O Shearman became its first customer. The business uses large language models (LLMs) to speed up tasks including document review and legal research with the aim of freeing up lawyers to do more sophisticated work. Alongside law firms, in-house legal and compliance teams at the likes of private equity group KKR and asset manager Bridgewater Associates have also signed up.Running Harvey means Weinberg lives out of hotel rooms for half the year as the company, which also counts Google Ventures among its investors, tries to remain the largest legal AI start-up.“I don’t think the billable hour is going to, like, go away in the next two or three years,” he said over coffee at Claridge’s, where he is staying along with celebrities including actor Tom Hanks spotted in reception.Weinberg added: “I think what’s going to happen is a lot of tasks are going to turn into a fixed fee and then I think the billable hour for high-end partners — that’s going to stay the same.”For its backers, Harvey, named after Harvey Specter, the smooth-talking corporate lawyer in the American TV drama Suits, and Harvard University, offers a way to tap into the billion-dollar revenues of the world’s largest law firms, where top partners can take home pay packets in the double-digit millions.Pat Grady, a partner at Harvey investor Sequoia, said it was “amazing . . . how quickly they [Harvey] became the trusted partner” in the industry.“If you’re a partner at a law firm and you A/B test Harvey versus a third-year associate, you get better work out of Harvey faster,” said Grady. “It’s largely making lawyers a lot more productive.”Not everyone is a fan, however. A quick scan of Reddit messaging boards throws up criticisms from “Harvey is a Silicon Valley bubble that mistook practising law for just parsing PDFs”, to accusations that the hype around the company is akin to the “emperor’s new clothes”.Most of the criticism centres on the idea that Harvey offers little more than ChatGPT, which underpins its technology.A spokesperson for Harvey said: “What differentiates Harvey in the market is that we are purpose built for legal, and specifically, our multi-model approach, capacity for bulk uploading documents, and commitment to enterprise security are three of the core reasons customers choose us over consumer solutions or other tools.”The company said the tool costs “a few hundred dollars per person per month on average”.Harvey is also not alone in the race to become the AI solution for law.On my way to meet Weinberg, an ad for Luminance — another AI legal company founded in 2015 and backed by the late UK tech entrepreneur Mike Lynch — flashes on screens alongside the escalator from the London underground platform.Other competitors include Sweden’s Legora, the “AI workspace for lawyers”, and Germany’s Noxtua, developed with law firm CMS.Still, a combination of its top-flight investors and early entry to the market means Harvey’s valuation is more than seven times that of Legora, its nearest competitor.The company may list in the future, said Weinberg, but he is keen to remain independent in the near term.Harvey this year entered into a partnership with LexisNexis, one of the biggest repositories of legal data, with the aim of developing tools such as motion to dismiss and motion for summary judgment arguments for litigators.The company is also opening an office in Australia this month and Toronto in October, and plans a presence in Germany and India alongside its London, San Francisco and New York bases.Weinberg seems unfazed by his rapid acceleration from working in a shared Airbnb to staying in luxury hotels.“At the end of the year when things are a little bit more quiet, I always have a kind of, like, ‘what is going on?’ moment,” he said. “I like to say that we need to move really, really fast — but actually don’t break things.”.© 2025 The Financial Times Ltd.

FT: How a former junior lawyer created a $5bn AI legal start-up

Key topics:

Related Stories