Legal Tech Has the Capital. The Ownership Rules Are Moving the Other Way.

Two things are true about the legal-services market this year, and they point in opposite directions. More capital has entered the sector than in any comparable period, most of it backing businesses that deliver legal work rather than tools for lawyers. At the same time, states have begun writing statutes that restrict exactly the arrangements those businesses depend on.

The capital first. Legal technology took in roughly $2.34 billion in the first quarter of 2026 alone. Legora raised $550 million at a $5.55 billion valuation on 10 March; Harvey raised $200 million at $11 billion on 25 March. The more interesting movement is one tier down, where the money is going into firms rather than software. A public index of AI-native law firms listed 27 entries in March 2026 and roughly 40 by late April. Eudia operates an AI-augmented firm under Arizona's alternative business structure regime, backed by institutional venture capital. In England, the Solicitors Regulation Authority authorised Garfield AI, an AI-driven practice handling small-claims debt recovery. The thesis is consistent across them: if the model can do the work, the margin belongs to whoever owns the firm, not whoever licenses the tool.

That thesis runs directly into the rule that a non-lawyer may not share in a legal fee, and several states have just restated it in statute. Colorado's Legal Practice Integrity and Fee-sharing Prohibition Act, HB26-1421, was signed on 4 June 2026 and takes effect on 12 August 2026, with a sunset in September 2029. It bars entities that economically participate in, hold themselves out as providing, or share in the fees from legal services, and it prohibits management-services organisations from taking compensation tied to law-firm profits or calculated as a percentage of profits, recoveries or settlements. Flat-fee and hourly MSO compensation is permitted, as are in-house arrangements, non-profit legal services, ordinary salaries and traditional litigation funding. Two features make it unusually sharp: it reaches legal services arising in whole or in part in Colorado and expressly addresses out-of-state structuring, and it creates private enforcement — clients may sue to recover fees, and competing law firms deriving ten per cent or more of their revenue from Colorado may seek injunctions, disgorgement and the voiding of contracts.

Illinois took the regulate-rather-than-ban route. HB 5487, passed by the General Assembly on 31 May 2026, defines alternative business structures and management-services organisations broadly and then constrains their conduct: no interference with a lawyer's professional judgment, no non-lawyer control over client records, personnel or competency standards, no compensation tied directly or indirectly to legal fees, revenues or profits, no post-termination non-competes, and no restrictions on discussing service quality or ethics. Administrative, technical and back-office services — billing, marketing, facilities, technology — remain permitted, which is to say that selling software to law firms is untouched. Enforcement is private, with statutory damages of $10,000 per violation and no requirement to prove harm.

The permissive jurisdictions have not closed. Arizona has allowed non-lawyer ownership through its ABS programme since 2021, updated its applicant criteria in March 2026, and remains the venue of choice: it approved KPMG Law US, the first Big Four-owned law firm in the United States, in early 2025. Utah's sandbox and the District of Columbia's rule continue to permit forms of non-lawyer participation, and Washington is piloting a programme. California, Illinois and New Mexico have all examined the question. The result is a patchwork in which the same corporate structure is lawful in one state, actionable in another, and the deciding factor is often where the client's matter arises rather than where the company is incorporated.

For founders and investors this reorders the diligence. The exposure is in the revenue model, not the code. Per-seat licences, flat fees and hourly services are broadly safe; percentage-of-recovery pricing, MSO economics tied to firm profit, and anything that reads as a share of the matter are the flagged patterns, and a nationwide product inherits the strictest state it touches. Multi-state operations should expect to route matters and price differently by jurisdiction, and should assume a competitor — not only a regulator — may be the party that sues. There is a second question buyers now ask, which is reliability: the most widely used public tracker of fabricated AI citations passed 1,500 court decisions worldwide by mid-2026, more than a thousand of them in the United States, and sanctions have escalated from four-figure fines to five-figure penalties per lawyer with referrals and suspensions attached. Verification records and citation audit trails have become a procurement requirement, which is a good sign for the category and an expensive one for anyone who shipped without them.

Key Takeaways

• Legal technology raised roughly $2.34 billion in Q1 2026, and the fastest-moving segment is AI-native law firms rather than tools sold to firms.
• Colorado's HB26-1421 took effect on 12 August 2026: it bars fee-sharing and profit-linked MSO compensation for matters arising in whole or in part in Colorado, and lets clients and competing firms enforce it directly.
• Illinois HB 5487 regulates rather than bans, but adds $10,000 in statutory damages per violation with no proof of harm required.
• Arizona, Utah, DC and a Washington pilot remain open; the practical constraint is the state where the matter arises, not the state of incorporation.
• Diligence should focus on the revenue model — percentage-of-recovery and profit-linked MSO economics are the exposure — and on verification records, as sanctions for fabricated citations continue to escalate.