The administration is pulling every lever to create a single, lenient federal standard for artificial intelligence, but state capitols and Capitol Hill have other ideas.
There is a standoff happening in American artificial intelligence policy, and the federal government is currently losing. The Trump administration has spent months building a multi-front campaign designed to strip states of their ability to regulate AI, deploying a Department of Justice litigation task force, Commerce Department reviews of supposedly burdensome laws, and a formal legislative framework. The goal is simple: clear the path for a single, minimally restrictive national standard that keeps American companies ahead of global competitors.
The states, however, are accelerating in the exact opposite direction. According to tracking data referenced by The Next Web, lawmakers have introduced roughly 1,208 AI-related bills across the country this year alone. Rather than waiting for Washington to establish a unified rulebook, state legislatures are aggressively writing their own, creating a sprawling web of compliance requirements that major technology companies and early-stage startups alike must navigate.
California remains the most aggressive standard-setter, with Governor Gavin Newsom making it clear that the state will independently dictate how its massive concentration of AI startups manage risk. The state continues to advance legislation targeting algorithmic discrimination in hiring and the proliferation of deepfakes, essentially ignoring the administration's push for deregulation. Other states are following suit with their own targeted approaches. Minnesota and Wisconsin are focusing on consumer data protections and safeguards against AI harms to children, while Utah has become a notable testing ground for AI transparency requirements.
What makes this conflict so fascinating is the legal scaffolding supporting the White House approach. Executive Order 14365, signed in late 2025, attempts to use interpretations of the Dormant Commerce Clause to argue that a patchwork of state laws unfairly burdens interstate commerce. Legal scholars from institutions like the Brennan Center for Justice have pushed back hard on this premise, noting that the President simply lacks the constitutional authority to nullify state statutes without explicit congressional approval. The order, as one analysis put it, is likely more bark than bite, setting the stage for years of federal-state litigation that will keep the actual rules of the road in limbo.
Congress could theoretically resolve this tension by passing the administration's preferred framework into law, thereby legally preempting state jurisdictions. But the legislative route has hit a brick wall. During National Defense Authorization Act negotiations in late 2025, Republican leadership quietly dropped a Trump-backed preemption provision following bipartisan pushback. Fiscal conservatives and Democrats alike proved deeply reluctant to strip states of their traditional police powers, and that coalition has not shifted. The result is a legislative landscape where the administration's AI framework is effectively stalled, lacking the votes to move forward.
What This Means for the Market
For the tech industry, this regulatory vacuum is a mixed blessing. Major players like Google, Meta, and OpenAI have lobbied heavily for a single federal standard, arguing that navigating dozens of different state requirements drives up costs and slows deployment. They have a legitimate point about operational friction. A startup building an AI hiring tool currently faces vastly different compliance obligations in Texas versus California, a dynamic that inherently favors large legal teams over lean engineering operations.
At the same time, the absence of a binding federal standard means there is still no overarching compliance baseline. Companies are being forced to adopt what is effectively the strictest state standard as their default operating procedure, a phenomenon familiar to anyone who watched California's CCPA shape national data privacy practices. Until either the courts rule definitively on the limits of executive preemption or Congress manages to pass a comprehensive bill, the real power over AI regulation in America will continue to reside in state capitols. Any business deploying machine learning tools in 2026 needs to be tracking Sacramento and St. Paul just as closely as Washington.
