Winston Weinberg understands what his junior colleagues in law firms need because he was once one of them. In 2022, the then 27-year-old was working at O'Melveny & Meyers when he began testing OpenAI models alongside his roommate, Gabe Pereyra, then a research scientist at Meta. After realizing that AI could fix inefficiencies in the legal field, the two co-founded Harvey, a startup built for lawyers.
Since then, Harvey's tools have become a staple of Big Law, used by more than half of the AMLaw 100 firms. Clients include powerhouses like Paul, Weiss, and the in-house legal teams at PwC and KKR. In three years, Harvey serves more than 1,000 customers in 6o countries. “The broader need for law enforcement AI is very high,” Weinberg told the Observer via email.
Silicon Valley has taken notice, too. After Weinberg emailed the OpenAI Startup Fund, he became Harvey's first institutional investor. The company was recently valued at $8 billion following a $160 million round in 2025—the year it raised capital three times—and closed with more than $190 million in annual recurring revenue.
Harvey has continued to grow through partnerships with LexisNexis, which integrates its legal content and citations, as well as collaborations with leading law schools at Stanford, NYU and UCLA to bring AI to legal education. Last month, even a full circle came, in cooperation with them Suits actor Gabriel Macht—the man behind Harvey Specter, the actor who inspired the startup's name—in what Weinberg called a “meaningful nod” to its origins.
But Harvey doesn't rest easy. It faces new competition from emerging players such as Sweden's Legora, which is valued at $5.5 billion, and established entrants such as Thomson Reuters' Counsel. Big AI developers like OpenAI, Anthropic and Google are also interesting as they expand into business applications.
The Spectator caught up with Weinberg to discuss what's next for Harvey as he expands globally, navigates competitive and staffing shifts, and charges ahead on his mission to integrate law and technology.
The following interview is planned for length and clarity.
How did you first get the idea for Harvey?
Gabe and I began experimenting with the original GPT-3 and quickly realized its potential for legal work. I was a junior partner at O'Melveny & Myers at the time, and after taking a simple test with real legal questions and a blind attorney review, it became clear that technology could fundamentally change the way lawyers work. We decided to go all in and cold email Sam Altman at OpenAI. That conversation led to our initial and early support for Harvey.
Is it true that Harvey takes its name Suits?
One of the things people forget about AI adoption is that getting people to change the way they work is very difficult. When Gabe and I started testing our product with users, we noticed that they spoke to us in a normal human voice, but when they went to tell, they were very strict and organized. Their information was worse than the clear instructions they would give to a young friend, so we deliberately tried to differentiate when people spoke a person's name versus a legitimate brand name, and the results were amazing.
The naming of Harvey has three parts: 1) the name of a person given the above behaviors that we have seen with lawyers, 2) inspiration from a famous lawyer who inspired many people to pursue law, and 3) it sounds like “Harvard,” people laugh, but in this case, there is a story as we think about building a brand that is reliable and trusted by legal groups around the world.
How did it get the attention of law firms in the first place, and what have been some of the biggest challenges in enabling the adoption of AI across the legal sector?
Initially, it was about finding futurists, such as David Wakeling at A&O Shearman, Claudia Junker at Deutsche Telekom and Bivek Sharma at PwC, and meeting people who knew AI would have an impact on law. And then it's about really listening to what's important to the world's top legal teams, which is why we spent a lot of time early on looking at security and governance as priorities. We also got a lot of “nos”—it's easy to look back now on the draw and think it was a walk in the park, but the truth is that we were lucky enough to meet some early changers and build their trust, and the industry as a whole has changed a lot in a very short period of time in terms of attitudes towards AI.
How many of your clients are law firms vs company official parties? Are you hoping to change that ratio going forward?
We have a strong mix of both—law firms are where we started, so they remain the majority of our clients, but the in-house business is growing rapidly. Only a few weeks ago, we announced the work with Eversheds Sutherland, Corrs, Stinston, Kromann, global expansion with DLA Piper, and renewal with Vinge on the law firm side, and HSBC on the in-house side.
Our Shared Spaces product enables collaboration between law firms and firms, and teams within firms, like never before. So our goal is not a ratio of firms and companies per se, but a ratio of how many legal teams are working together within Harvey on any given day—that's a metric I value the most.
What's next for Harvey? What are you most excited about in 2026?
So much—we just hired our first chief product officer, Anique Drumright, and our vice president of design, Christina Nguyen White, and I'm excited to see the people they're hiring and the impact they'll have on our customer experience. We also announced the opening of additional offices in Singapore, Dallas, Dublin, Paris and Bangalore in addition to our existing locations in Sydney, San Francisco, New York, London, and Toronto. We've held our first ever event, the Harvey Forum in London, and we'll be hosting our next Harvey Forum in New York soon, so there's a lot to look forward to in the near future.
What do you think about the impact of AI on white collar jobs, especially when it comes to small roles?
It is one of my main beliefs that AI will revolutionize the legal profession, but if done right, it will create more opportunities for training and the kind of legal careers that many people (myself included) envisioned when applying to law school. You're starting to see this—the best lawyers charge for high-level strategy, sales or breach work and are well compensated for it. I don't think that will change. What will change are all the red-line repetitive tasks, where AI can transform many existing workflows.
But your point is well taken that the industry as a whole needs to continue to invest in young talent, which is why we at Harvey are doing two things: 1) creating our own internship and early career program to lead the way in this, and 2) collaborating deeply with law schools in the US, UK, Spain, and Australia to help prepare them and the school's faculty to be part of our industry mission to be part of our legal industry mission. the solution.
How will you continue to differentiate Harvey from the competition? What will stop the big AI players, like OpenAI, Anthropic and Google, from dominating the legal field?
We have great respect for each of those organizations, and they are all valuable partners because Harvey's platform integrates their different models. We have said for many months, publicly and privately, that for a long time, model companies are the most competitive with direct AI That does not change our approach, but increases our urgency to focus on different areas, such as security, governance, collaboration, memory and integration, to ensure that we remain the solution of choice for law firms and for internal collaboration with their internal and in-house teams.
