Senate Republicans' 10-Year State AI Ban Failed in 2025; Midterms Make Revival Unlikely

The last time Senate Republicans tried to write a 10-year moratorium on state AI regulations into a budget bill, the chamber voted 99–1 to strip it out. That was July 1, 2025. Sixteen months later, after a midterm election that flipped both chambers to Democratic control, the political calculus has only moved further against a federal preemption of state AI law.

The 2025 reconciliation fight

In the spring of 2025, House Republicans attached a provision to their budget reconciliation bill that would have barred states from enforcing most AI regulations for a decade, tying compliance to federal broadband funding under the BEAD program. The measure was framed as a way to prevent a "patchwork" of state rules, but it drew opposition from 17 Republican governors, 40 state attorneys general, and groups ranging from the Heritage Foundation to the Center for American Progress, according to the Senate Commerce Committee.

Sen. Maria Cantwell (D-Wash.), then the Commerce Committee's ranking member, co-sponsored an amendment with Sen. Marsha Blackburn (R-Tenn.) to remove the moratorium. "The Senate came together tonight to say that we can't just run over good state consumer protection laws," Cantwell said after the 99–1 vote. "States can fight robocalls, deepfakes and provide safe autonomous vehicle laws."

The provision's defeat was a rare moment of bipartisan consensus on AI policy. It also crystallized the political forces that would shape the 2026 midterms. A pro-industry super PAC, Leading the Future, launched in August 2025 with roughly $100 million to oppose regulation. In response, a coalition of AI safety advocates, consumer groups, and effective altruism-aligned donors built Public First, a competing super PAC network that ultimately drew roughly $41 million from Anthropic and its CEO Dario Amodei, POLITICO reported.

Midterms rewrite the map

On November 3, 2026, Democrats won control of the Senate (approximately 54 seats) and the House (projected up to 244 seats), according to TechTimes and PPC Land. The election was fought on inflation, the Iran war, and President Trump's approval rating (38% in CBS exit polling). But AI and data-center issues surfaced as a significant secondary theme: NBC exit polls found 56% of voters called AI and data centers a "crucial or major factor" in their vote, POLITICO reported.

The results ended the unified Republican control that had made the 2025 preemption attempt possible. With a Democratic House and a Senate still subject to the 60-vote filibuster, any new budget bill containing a state AI ban would need Democratic support that does not exist. Sen. Cantwell, now in line to chair the Commerce Committee, has consistently opposed preemption without strong federal protections in return, TechTimes reported. Sen. Ed Markey (D-Mass.) and Sen. Elizabeth Warren (D-Mass.) have pledged to block any defense bill that includes a moratorium, according to a Markey press release.

State laws survive and expand

With federal preemption off the table, the center of gravity remains in state capitals. Colorado's rewritten AI Act (SB 26-189) is set to take effect January 1, 2027, though a federal court has already ordered the attorney general not to enforce the original law or any amending legislation until rulemaking concludes and the court rules, PPC Land reported. California's Delete Act gave residents a one-stop data-broker deletion tool that processed 242,000 sign-ups in its first eight weeks. New York's RAISE Act on frontier AI safety and its ban on human-posing chatbots for minors take effect in 2027. Other states have enacted or are advancing their own rules on algorithmic discrimination, data-center siting, and synthetic media disclosure.

The industry's argument that a patchwork undermines innovation is unchanged, but the election has made the patchwork "a cost of doing business rather than a temporary condition awaiting a federal fix," as PPC Land put it. Compliance teams that planned for a single 2027 federal standard now face 20-plus state regimes.

What comes next

Democrats have signaled three legislative priorities for 2027: consumer protections against data-center-driven electricity costs, independent testing of advanced AI models, and liability rules for AI-related harm, according to Shattered.io. Rep. Sara Jacobs (D-Calif.) and Rep. Don Beyer (D-Va.) are drafting a framework with a government shutdown power for models that fail safety tests. Sen. Cantwell proposes enforceable standards, pre-release audits, and mandatory incident reporting, TechTimes and Shattered.io reported. A bipartisan FRONTIER Act (Rep. Lori Trahan, D-Mass., and Rep. Jay Obernolte, R-Calif.) addresses catastrophic risk and a national framework, though its preemption scope is unresolved, Shattered.io reported.

Oversight arrives first. With committee gavels and subpoena power, Democrats plan hearings on AI safety, data-center impacts, and industry practices before any bill moves, PPC Land and Shattered.io reported. The House Democratic Commission on AI is preparing a roadmap for Minority Leader Hakeem Jeffries, Shattered.io reported.

The preemption question remains open — but not in a budget bill

The outgoing Congress returns for a lame-duck session, but its legislative agenda centers on the SECURE Data Act (a federal privacy bill with preemption), while AI preemption efforts continue through executive branch actions, PPC Land reported. With Democrats holding congressional majorities starting January 3, 2027, the procedural path for a Republican-only preemption measure has closed.

The 2025 vote showed that even a Republican Senate would not accept a blanket ban on state AI enforcement. The 2026 election ensured that the next Congress will be led by lawmakers who view state regulation as a floor, not a ceiling.