We just raised another $550 million to bring AI to the legal industry. I still believe there will be more work than lawyers

Legora has closed a $550 million Series E at an $8.8 billion valuation, and CEO and co-founder Max Junestrand used the announcement to make a bigger claim: legal AI is creating more work than it is eliminating. That’s a bold pitch in a market crowded with well-funded rivals, but it also fits the way Legora says it has built the business from the start.

The company’s origin story is very specific. Its first big customer was Mannheimer Swartling, one of the largest firms in the Nordics, and Legora set up its first office in one of that firm’s conference rooms. Junestrand says the company made a rule early on: every new customer had to make Legora stronger. Three years in, he says, the company has more than 2,400 customers and the rule still holds.

The product strategy is just as deliberate. Legora did not train its own general legal model. Instead, it built on frontier models from others, arguing that the real work is in the layers around them: security, privacy, parsing and handling large document sets. The company routes tasks to the model that handles them best, and only post-trains for narrow jobs where the frontier models fall short. That approach, Junestrand says, has improved both performance and cost as better models have appeared every few months.

Pricing follows the same logic. Legora was among the first legal AI companies to use consumption-based pricing, so customers pay more or less depending on how much work the product does. The company says it only grows when it is earning its place inside a firm. It also has more than 150 legal engineers, including former lawyers, and now has more lawyers than software engineers. Daily active users spend more than two hours a day on the platform.

The bigger bet is on how firms actually work. Legora says legal AI is only part of the answer; the rest is fitting into the messy reality of law firms, where coordination is still done by hand across too many tools. Its agentic operating system is meant to sit between the models and the firm, letting agents handle whole tasks like diligence or discovery while lawyers step in for judgment. The company says annual recurring revenue now exceeds $200 million, and that it would make the same choices again.