States have enacted 84 new AI laws across 27 jurisdictions in the first half of 2026, according to a Transparency Coalition mid-year report released July 21, already past 2025’s full-year total of 73, with California, Massachusetts, Pennsylvania, and Michigan still in session. NYU’s Center on Technology Policy, using a broader methodology, counts 109 AI laws plus 28 data center laws by July 1, and puts the total number of states that have enacted AI legislation this year at 29.

The scoreboard matters because the Trump administration spent the last eight months trying to stop it. A December executive order pushed federal preemption but carved out child safety and data center infrastructure, which happen to be the two categories driving nearly all of the statehouse volume. Companion chatbot regulation alone accounts for more than 100 introduced bills and 14 enacted so far, most of them mandating disclosures, self-harm protocols, and parental controls.

Connecticut is the template. In June, Gov. Ned Lamont signed Public Act 26-15, a 74-page omnibus bundling chatbot safeguards, employment disclosures, an AI regulatory sandbox, and workforce training into one vehicle.

Federal pressure is real but uneven. DOJ’s new AI litigation task force intervened April 24 on xAI’s side in its suit to invalidate Colorado’s AI Act. Sen. Marsha Blackburn released a 17-title federal framework, and the Senate Judiciary Committee unanimously advanced preemption-adjacent legislation. MultiState reports that White House intervention killed AI bills in Florida and Utah, and that BEAD funding threats moved Virginia lawmakers.

The 2026 arc now depends on California’s 30 pending measures. Clear them and the year passes 2025’s 159-law mark; stall them and preemption catches up.

Sources

Sources