Harvey raises $550M at $15.5B valuation, months after $11B round

Harvey, the startup applying AI to legal work, has raised $550 million at a $15.5 billion valuation, the company announced on Wednesday. According to TechCrunch, the round was co-led by Diffusion and Lightspeed Venture Partners, and it lands only months after the company was last valued at $11 billion.

A third up-round in under a year

The pace of Harvey's fundraising is unusual even by the standards of the current AI boom. TechCrunch reports that the new financing follows a $200 million round at an $11 billion valuation announced in March, which itself arrived a few months after a December round valuing the company at $8 billion. Taken together, Harvey has now raised more than $1.55 billion in total and has nearly doubled its valuation in roughly nine months.

Harvey did not attach a name to the round. PitchBook estimates cited by TechCrunch put the company at a minimum of eight priced rounds since 2023, five of them since 2025. Because some of those earlier rounds were structured as extensions, the latest raise could reasonably be viewed as a Series F.

Betting on open-weight models

The raise comes shortly after a notable technical shift at the company. A couple of weeks earlier, Harvey introduced Harvey Tenet, its first in-house model. According to TechCrunch, Tenet was built on the open-weight Kimi K3 model and then post-trained on legal data with help from inference provider Fireworks, the same company that assisted code-editing startup Cursor in training its own model.

Harvey is also pushing the approach beyond its own product: the company encourages customers to adopt and post-train open-weight models of their own. As TechCrunch observes, this makes Harvey an illustration of how a whole profession can weave AI deeply into its workflows while sidestepping dependence on proprietary frontier labs such as OpenAI and Anthropic.

Why it matters

The round is another data point that investor appetite for AI startups has not cooled. A company going from an $8 billion valuation to $15.5 billion inside nine months, across multiple priced rounds, indicates that competition among investors for credible AI plays remains intense, and that a domain specialist can now command valuations once reserved for foundation-model developers.

The open-weight strategy is arguably the more consequential signal. By building on Kimi K3 and helping customers do the same, Harvey reduces its reliance on a small number of frontier labs, gains more control over cost, customisation and sensitive legal data, and offers a template that other vertical AI vendors may follow. If the bet pays off, the legal industry could become proof that serious professional AI adoption does not require proprietary frontier models at its core.