30-year-old CEO of $11 billion Harvey earned the backing of OpenAI and Sam Altman. He says you have to 're-earn' your role every 6 months - Fortune

Winston Weinberg will tell you that nobody at Harvey, including himself, has a permanent job. The company's 30-year-old CEO has said publicly that everyone, himself included, has to re-earn their position every six months. It is a clean management philosophy. It also happens to be a useful frame for thinking about Harvey itself.

Harvey has raised nearly $1 billion in just over a year without having spent most of it. The company announced a $200 million round last week co-led by GIC and Sequoia, lifting its valuation to $11 billion. Since last February, Harvey has disclosed four funding rounds totaling about $960 million, per Business Insider reporting. Sequoia has now co-led three consecutive rounds in the same company — the kind of conviction that fund rarely extends twice.

The money is real. So is the question of what Harvey is building it on.

The short version: Harvey is not a model company. Its original thesis — that a purpose-trained vertical model would outperform frontier models at legal tasks — has already been tested and, by Harvey's own benchmarks, rejected. The company's BigLaw Bench showed Google, OpenAI, Anthropic, and xAI all surpassing its proprietary system on legal work. Harvey switched. What remains is workflow, enterprise relationships, and whatever moat lives in being the platform where legal work actually runs.

That is a defensible position. It is not an $11 billion thesis on a slide.

The Business Insider reporting — Melia Russell's piece is the most substantive account of Harvey's capital strategy to date — makes the company's rationale clear: frontier model advances, particularly in coding, have opened product paths that did not exist when the previous fundraising rounds closed. "The things that we wanted to build over the next three years, we can probably build in one now," Weinberg told Business Insider. The company says it has not touched the majority of rounds it has raised and is now stockpiling to be more aggressive. It hired its first chief product officer in February — Anique Drumwright, previously at Rippling — and brought on a former Gibson Dunn partner, Keith Enright, as chief strategy officer this week.

The sequencing is odd, though. If the opportunity is that clear, why has most of the previous capital sat untouched? Harvey's answer is that it did not know how fast the roadmap would accelerate. The honest read: it raised like a model company and is now trying to become a product company, and the $11 billion valuation was the price of getting into the room.

The competitive pressure is real from both directions. Legora, a Swedish legal AI startup, raised $550 million earlier this month at a $5.5 billion valuation — a fraction of Harvey's multiple on comparable or lower revenue, depending on whose numbers you trust. More significantly, the frontier model labs are circling. Anthropic released a legal-work plugin in February — a development that triggered a sharp sell-off in legal tech stocks, per reporting at the time. Weinberg has said he considers the model labs themselves a bigger competitive threat than any legal AI startup.

The implicit argument against the model labs is workflow depth. A general-purpose model trained on case law is not the same as a system built around how a law firm actually operates — the audit trails, the matter management, the institutional memory of what a partner likes in a due diligence package. Harvey says it runs more than 25,000 custom agents across M&A, due diligence, contract drafting, and document review for 100,000 lawyers at 1,300 organizations. The majority of the AmLaw 100 are on the platform, along with more than 500 in-house legal teams and 50 asset management firms.

Those numbers are substantial. They are also self-reported, and Harvey has a clear interest in them being as large as possible.

Harvey was founded in 2022 by Weinberg, a lawyer by training, and Gabriel Pereyra, a former AI researcher at Meta and Google DeepMind. Their first check came from an unlikely cold email: they messaged Sam Altman on July 4, 2022, got on a call that same morning, and landed their OpenAI Startup Fund investment shortly after. Andreessen Horowitz, Coatue, and Kleiner Perkins followed. Sequoia led the Series A.

The company's annualized revenue is "significantly north of $200 million," Weinberg said, up from $190 million at the end of 2025. That is real revenue for a four-year-old company. It is also a long way from the multiple a public software company at $11 billion would trade at — and a long way from proving that Harvey is the operating system for legal work rather than a well-funded feature running on someone else's model.

Weinberg's six-month re-earning cycle is good culture. For Harvey to justify $11 billion, it is going to need to show the product, not just the philosophy.