States enact 109 AI laws by July 2026, trailing 2025's pace

Half a year into 2026, and states have not stopped writing AI law. According to a mid-year tally by Scott Babwah Brennen at TechPolicy.Press, 29 states have collectively enacted 109 AI laws and 28 data center laws as of July 1. That trails last year's pace, when 39 states had passed 121 AI laws by the same point, but it is still a lot of new law, and it is happening against a federal push to slow states down.

The category dominating the docket is companion chatbot regulation. Legislators introduced more than 100 bills on the topic and enacted 14 into law, most building on California and New York's 2025 frameworks that require operators to display warnings that the chatbot is artificial and address risks including sexual content involving minors and self-harm. Data center oversight is the other shift worth flagging: 28 new laws, and for the first time both Republican and Democratic-led states are going beyond tax incentives, into ratepayer protections, audits and disclosures, rollbacks of prior tax breaks, and outright development moratoria.

The cross-partisan piece is the unusual one. Brennen reports roughly 61% of Republican-controlled states and 69% of Democratic-controlled states have enacted AI legislation this year, and on companion chatbots specifically, six Democratic and eight Republican-controlled states passed similar laws. Consumer protection expanded on both sides too: six new restrictions on AI use by health insurers in states including Iowa and Washington, and limits on algorithmic dynamic pricing in Connecticut and Nebraska. Frontier model rules moved much slower. Illinois passed a bill requiring yearly third-party audits, and Connecticut enacted whistleblower protections, and that is most of what actually happened at the frontier layer this year.

The honest caveat is that the piece does not tell you why the pace slipped versus 2025. Some of it may be the Trump administration's December 2025 executive order, which created an AI Litigation Task Force and threatened to withhold federal broadband (BEAD) funding from states enacting what it called 'onerous' AI laws. Some of it may simply be that the obvious bills passed first. What Brennen does not settle is whether that federal pressure actually chilled bills that would otherwise have moved, or how the task force intends to test state laws in court.

If your product touches a chatbot, a health-insurance AI, a dynamic pricing engine, or a large data center, the practical read is that the compliance surface is now a 29-state patchwork that is still growing, and the split down the middle is not the traditional red-blue one you might expect.