Legal AI startup Harvey hits $100M ARR and $5B valuation just 3 years after launch

Three years ago, two friends—one a lawyer, the other an AI researcher—started tinkering with OpenAI’s GPT-3. That side project just turned into a business doing $100 million in annual recurring revenue.

On Monday, legal AI tech startup Harvey announced it had crossed the $100 million ARR mark, becoming one of the fastest-growing AI companies of its kind. The milestone comes just two months after Harvey hit a $5 billion valuation off the back of a $300 million Series E round led by top-tier investors.

“As a business, we generate more than $100M ARR. 42% of AmLaw 100 firms trust Harvey for their AI needs, and our team now includes 350 employees,” Harvey said in a blog post.

Harvey builds AI tools for legal professionals—think in-house counsel and law firm associates buried in contracts, research, and document review. Its natural language platform plugs directly into lawyers’ existing workflows, helping them tackle legal research, drafting, and diligence tasks without bouncing between a dozen legacy tools.

The company’s growth hasn’t been fueled by hype. It’s been driven by usage.

“Most of our accounts grow pretty massively,” co-founder and CEO Winston Weinberg told CNBC. “You’ll sell to a Comcast or to a law firm, and they’ll buy a couple hundred seats, and then they expand that usage pretty quickly.”

Harvey Emerges as One of the Fastest-Growing AI Startups in Legal Tech

Harvey now serves over 500 customers, including Comcast, the parent company of CNBC. Weekly active users have quadrupled over the past year.

“Three years in, Harvey is helping more than 500 customers in 54 countries transform their business with AI. Our Weekly Active Users (WAU) grew 4x in the past year and 6x the year prior. Monthly queries grew 5.5x in the past year. And active files stored in Harvey grew 36x in the past year from 268K to 9.75M,” the company said.

Founded in 2022 by Weinberg and Gabe Pereyra, Harvey is part legal startup, part AI experiment gone right. Weinberg, a former associate at O’Melveny, partnered with Pereyra, a former DeepMind and Meta researcher. The idea came after the duo started playing with GPT-3 and saw how it could dramatically speed up legal workflows.

The company’s name, “Harvey,” is a nod to the sharp-witted lawyer character from the TV show Suits, Weinberg said.

Harvey has raised more than $800 million to date, backed by Sequoia Capital, Kleiner Perkins, and the OpenAI Startup Fund. It also landed a spot on the 2025 CNBC Disruptor 50 list—yet another signal that investors see this space as more than a niche.

With revenue flowing and a massive valuation to live up to, Harvey is now entering its next phase: scaling. Weinberg said the team has learned that speed is everything in the AI race.

“With gen AI, and how fast everything’s moving, you just have to learn how to scale really, really fast,” he said. “I’d say, like every six months I go through a new scaling experience.”

To help with that, Harvey has brought in some heavyweight talent. Former Twitter engineering director Siva Gurumurthy recently joined as CTO, and Stripe veteran John Haddock came aboard as chief business officer.

Weinberg said having strong leadership in place makes all the difference during high-growth phases. “We’re starting to get to the point where we have really good leadership in place,” he said. “That just changes your ability to scale to such a massive degree.”

With global expansion on the horizon and enterprise adoption growing, Harvey looks less like a startup and more like a legal infrastructure play built for the AI age.