A Harvey listing would be the first time public investors put a price on legal AI, and I think the price they pay will be lower than the one private rounds have set. Private money has valued legal software on growth alone. Public markets ask a harder question: how much each customer pays and how much more they will pay next year. On that measure Harvey is strong, but its valuation still assumes far more than the business shows today.
Business Insider reported on November 5 that Harvey has held early talks with lawyers at Latham & Watkins about going public, citing two people familiar with the matter. The talks are preliminary and no listing is imminent. Harvey is expected to watch how the next group of tech offerings performs, with Anthropic possibly listing before Thanksgiving, before it decides. Harvey and Latham did not comment.
The numbers behind the talks are large for a four-year-old company. Harvey was recently valued at $15.5 billion and said in September it had passed $400 million in annual recurring revenue. This fall it raised $550 million in a round led by Diffusion and Lightspeed, with Goldman Sachs Alternatives, J.P. Morgan's growth equity team, Coatue and T. Rowe Price accounts joining. It has more than 3,500 customers, including 80% of the top 100 US law firms and 20% of the Fortune 500, and it has hired a revenue chief who helped take Rubrik public.
Here is my arithmetic. Divide $15.5 billion by $400 million and Harvey is valued at about 39 times its recurring revenue. When I wrote about Fal a few days ago, inference was priced at about 20 times revenue, and Legora's October round worked out to roughly 44 times. So Harvey sits much closer to the application premium than to the infrastructure discount, and an IPO would test whether that premium holds outside a private room. The second number matters more to me: $400 million spread across 3,500 customers is about $114,000 a year per customer. For a product used by most of the largest law firms in the country, that is a modest bill.
That modest bill is the bull case and the risk at the same time. If the top firms are paying an average of six figures a year, there is plenty of room to expand inside accounts Harvey already has. But public investors will want proof that firms are moving from pilots to firm-wide rollouts, the way Latham did, and not just adding seats. Net revenue retention, not logo count, will decide where this stock trades. A company that has already signed 80% of its core market cannot grow by finding new top firms. It grows only by charging the ones it has more.
The other lesson is where the money is not going. Harvey sells to large firms and corporate legal departments, the best-funded buyers in law. The consumer end, where people face a parking ticket or a fine and cannot afford an hour of a lawyer's time, gets almost none of this capital. At iQuash we check tickets, fines and notices for defects and then connect people with licensed lawyers, so the software does the screening and the lawyer handles the cases worth fighting. Much lower revenue per user, but a far larger number of users, and a market that the large-firm tools were not built to serve.
Through ELP Ventures I look for founders who price against that gap. A Harvey IPO, if it happens, will reset the comparable for every legal AI round that follows. If public investors pay something near 39 times, late-stage valuations hold. If they pay closer to the 20 times that infrastructure gets, the next wave of legal startups will need to show revenue per customer and retention, not just a list of famous firms. I would rather invest after that reset than before it.