EvenUp passes $150M in revenue as legal AI braces for a bubble test

EvenUp has surpassed $150 million in recurring revenue under contract and now serves more than 2,000 personal injury law firms — a number its CEO, Rami Karabibar, offers as evidence that narrow, deeply embedded legal AI survives when the broader market turns. His theory is simple: the startups that last are the ones woven furthest into a customer's daily workflow, not the ones with the widest addressable market.

The claim lands at a nervous moment. Investors have been plowing money into legal tech, valuations are soaring, and "AI-native law firms" keep appearing, each pitching itself as different from Crosby. With tech stocks selling off and AI doomsayers declaring a bubble burst, the question of which legal AI businesses are built to last has stopped being academic.

How did EvenUp get to $150 million?

The company started as something of a one-trick pony, drafting demand letters for plaintiff-side lawyers. Over the past 18 months it expanded into software that manages cases from start to finish. The product line broadened; the target market did not. EvenUp has never strayed from personal injury, and Karabibar thinks that focus keeps the company bubble-proof.

The contrast is deliberate. Harvey and Legora build applications that serve everyone from Big Law to multinational corporations. Horizontal breadth wins large contracts but invites more competition; niche depth builds switching costs within a lucrative, underserved vertical. Both strategies raise money today — only one may prove resilient if funding tightens and customers start cutting tools that sit at the edges of their workflow rather than its core.

What else moved this week?

• Harvey held early IPO talks with lawyers at Latham & Watkins, the clearest signal yet that a legal AI exit window is being tested.
• Lerer Hippeau is backing Blackletter, a new law firm started by former executives at Cohere and Blue J, with a $6 million raise — a "neofirm" bet on restructured legal services rather than software alone.
• Two college students in Singapore are helping law firms figure out technology, profiled by Business Insider's Aditi Bharade.

Can Big Law learn founder thinking?

Herbert Smith Freehills Kramer shipped 31 firm members to Stanford for a three-day retreat before the start of the school year, complete with shared lodging, a morning bus to campus, and an invitation from chief AI officer Ilona Logvinova to pack hoodies — only a few obliged. She wanted the firm's top brass out of their offices and thinking less like lawyers and more like founders.

Day one covered the state of AI, with sessions from legal-tech founders, OpenAI representatives, Stanford researchers, and general counsels on what clients increasingly expect from outside counsel. On day two, the executives used AI developed at Stanford, including a tool that turned the expertise rattling around in their heads into instructions autonomous agents could potentially follow. The exercises targeted what Logvinova calls the "imagination gap": the distance between knowing what the tech can do and actually applying it.

Stanford got something too. The program doubled as a dry run for Legal Engineering Academy, a bootcamp launching next summer. Adam Sterling, an associate dean at Stanford Law, said the university wanted to see how lawyers engaged with the technology in practice, not just in theory, and that working with practitioners helps faculty understand the profession its students are entering.

What does this mean for the road ahead?

The experiment suggests preparing the next generation of lawyers may start with retraining the current one. The scene convenes in person next week at the LegalTech Fund Summit in Scottsdale, Arizona, now in its fifth year, where founders, funders, and power brokers will test whether niche depth or horizontal scale is the safer bet.