Exporting

Legal AI Startup Harvey Hits $15.6 Billion Value With $550 Million Round

Legal tech startup Harvey is now valued at $15.6 billion after raising $550 million in a new funding round aimed at powering efforts to build its own AI models.

The fundraise was co-led by Lightspeed Venture Partners and Diffusion, a new firm cofounded by longtime Harvey backer and former Coatue Management investor Kris Fredrickson. Sapphire Ventures and Whale Rock Capital Management, as well as a number of Harvey’s existing investors, also participated in the round.

With over $1.5 billion raised to date, Harvey is the highest valued and best funded legal AI startup. The company, which sells to Big Law firms like Latham & Watkins and legal teams inside large companies like Microsoft Corp., has passed $400 million in annual recurring revenue.

It’s now working with more than 3,000 customers, up from 1,300 in March, when it hit an $11 billion valuation with a $200 million raise.

The startup has also acquired AI agent security platform Guardrails AI, which allows developers to stress-test AI agent behavior. It’s Harvey’s fourth acquisition this year and a reflection of how the startup is increasingly relying on long-horizon agents, which can work autonomously on complex tasks for hours. Terms of the deal were not disclosed.

Harvey plans to use its latest cash infusion to support its work to build its own custom models, including to hire more specialized engineers. In August, the startup released its first fine-tuned model for legal tasks, called Tenet, which it built on a Chinese open-weight model, Moonshot AI’s Kimi K3. Open-weight AI models like Kimi K3, which allow users to download and customize the technology in a process known as post-training, are becoming increasingly capable and often cost less to use than closed models from companies like Anthropic PBC and OpenAI.

Harvey has been working on directing user tasks to the most affordable AI models available, as long as they perform comparably to more expensive options. It’s also collaborating closely with law firms to fine-tune models to their needs using these firms’ historical data.

“All software companies need to turn into AI companies, full stop. And post-training models is going to be part of that process,” Harvey Chief Executive Officer Winston Weinberg said. “This is going to become a muscle that you need to have to compete as a software company.”

Some of the biggest companies in AI have moved to compete directly in legal tech. Anthropic has released several legal plug-ins for its Claude model, while OpenAI has partnered with multiple law firms to customize ChatGPT for their attorneys.

Skeptics have also argued that law firms aren’t incentivized to adopt AI tools that make them more productive, since they typically charge their clients by the hour.

Lightspeed partner Sebastian Duesterhoeft, who has joined Harvey’s board as an observer, sees its biggest opportunity with in-house legal teams, which he says are more incentivized to maximize their productivity with Harvey’s tools. “In-house law ultimately is the direct path to go after all of legal services spend,” he said, adding that his firm thinks legal services may have the second-largest addressable market in AI, after coding.

Following its Guardrails AI bet, Weinberg said Harvey plans to make more acquisitions, focused on bringing on quality teams rather than seeking out specific products. “Because it’s so much easier to build software right now, more than ever, teams matter,” he said. “No matter the size of the acquisition, I think of it as an acquihire.”

(Disclaimer: Bloomberg Beta, the venture capital arm of Bloomberg News parent Bloomberg LP, is an investor in Guardrails AI.)

Written by: @Bloomberg

Bloomberg.com