By Patrick Ip
Authorized tech funding is booming, however the cash isn’t spreading evenly throughout the market.
Final 12 months, Crunchbase Information reported that authorized tech startup funding was using excessive as investor enthusiasm for AI reshaped authorized software program funding, citing a Goldman Sachs report estimating that 44% of authorized work may finally be automated. That focus has helped create one of many clearer success tales in authorized AI — and may additionally be obscuring an adjoining market that continues to be far much less developed.
Utilizing disclosed funding totals for a particular group of plaintiff-side authorized AI firms, the imbalance is difficult to overlook.
EvenUp has raised $370 million, Eve $164 million, Supio $85 million, and Darrow $63 million, for a mixed whole of roughly $682 million. Plaintiff-focused firms account for about 71% of disclosed capital for authorized AI, suggesting traders have discovered part of the sector the place adoption, workflow readability and venture-scale narratives already line up.
That investor curiosity will not be obscure. Plaintiff corporations are inclined to share extra standardized workflows round shopper consumption, case analysis, medical overview and demand era — all areas the place AI can automate repetitive work and enhance throughput. As these corporations have adopted software program, the class has develop into simpler to grasp, distribute and fund.
The underserved aspect: authorized protection
The protection aspect, in contrast, stays underdeveloped and will current the following large alternative.
Company authorized departments and the regulation corporations managing high-volume protection work nonetheless rely closely on fragmented methods, spreadsheets, email-based coordination and outdoors counsel processes that weren’t designed to provide portfolio-wide visibility. For firms dealing with a whole bunch or hundreds of energetic issues, litigation is usually nonetheless run extra as a companies perform than a software-enabled one.
That creates a large however harder-to-package alternative. Retailers, insurers, healthcare methods and monetary companies firms can every handle massive litigation portfolios, but many nonetheless lack a unified view of case danger, settlement patterns, authorized spend and outdoors counsel efficiency. The necessity will not be new. What has been much less clear is whether or not a venture-backable software program class might be constructed round it.
A part of the explanation defense-side authorized AI has lagged is structural. Workflows fluctuate broadly by business, matter sort and regulatory context, making the market much less standardized than plaintiff-side practices. Shopping for selections additionally are inclined to run by common counsels, authorized operations groups and outdoors counsel relationships, which may lengthen gross sales cycles and make the class look much less instantly viable to traders.
However a shift is underway. Final fall, Crunchbase Information reported that authorized tech funding reached file highs in 2025, reinforcing how shortly investor consideration has shifted towards AI-enabled authorized workflows. As plaintiff-side corporations get quicker at sourcing, valuing and prosecuting claims with software program, the operational strain on protection groups mounts. On the identical time, AI is making it extra possible to show messy litigation workflows into methods that may floor comparable issues, flag danger earlier and benchmark outcomes throughout portfolios.
From an investor perspective, that makes defense-side litigation AI look much less like a distinct segment and extra like an underbuilt phase of a broader authorized software program market. If plaintiff-side funding displays the place authorized AI has already develop into straightforward to fund, defense-side infrastructure might signify the place the following class nonetheless has room to kind.
Traders, take discover
For enterprise capitalists, that is the type of asymmetry value watching: a big enterprise market with measurable ache factors, bettering technical feasibility, and no entrenched class chief but. What traders ought to watch is whether or not startups within the class can pair proprietary end result information with repeatable enterprise adoption — the mix almost certainly to provide a sturdy class chief.
One rising method on the protection aspect is publicity and settlement benchmarking: utilizing historic decision information to estimate settlement ranges, authorized spend and case danger throughout related issues. In follow, that may imply evaluating claims by jurisdiction, plaintiff agency, declare sort or different working variables to assist in-house groups make quicker and extra constant selections.
If the class scales, one potential moat might come from proprietary end result information. Protection-side settlement particulars, matter economics and backbone patterns are sometimes tough to reconstruct from public data alone.
A platform that aggregates and normalizes these indicators throughout prospects may construct an information asset that turns into extra helpful with scale — a well-recognized dynamic in vertical software program, and a possible early sign for traders of sturdy benefit in defense-side authorized AI.
There’s nonetheless no clear, scaled, venture-backed winner constructed particularly round defense-side litigation intelligence. For startup and development traders, that makes the phase much less a settled market than an open query: whether or not one among authorized AI’s subsequent sturdy firms will emerge not from the workflows which have already attracted essentially the most capital, however from a big enterprise class whose software program stack continues to be taking form.
Patrick Ip is CEO and co-founder of Theo Ai, which builds AI-powered litigation intelligence for company protection groups and regulation corporations.
Associated Crunchbase question:
Associated studying:
Illustration: Dom Guzman
Keep updated with latest funding rounds, acquisitions, and extra with the
Crunchbase Every day.
