In the crowded Democratic primary for New York's 12th Congressional District, the AI industry has already spent more than $2.4 million to defeat state Assemblymember Alex Bores, the candidate who authored one of the nation's first major AI safety bills and opposed the industry’s lobbying campaign to water it down.
The heavy spending is part of a civil war taking place inside the artificial intelligence industry that has already impacted more than a dozen primaries ahead of the 2026 midterms. One faction, the deregulatory side, is backed by donors including OpenAI President Greg Brockman and his wife Anna Brockman, who have each given $12.5 million, as well as venture capital firm a16z founders Marc Andreessen and Ben Horowitz, who each gave $12.5 million. Other key donors include Palantir co-founder Joe Lonsdale, investor Ron Conway, and the company Perplexity AI.
Leading the Future, its Democratic arm Think Big, and its Republican arm American Mission, reported raising a bit more than $50 million last year in its FEC filings, though the network claimed it raised more than $125 million last year and entered 2026 with about $70 million in cash on hand. The money is being spent to elect lawmakers who support a single national AI framework that would block states from enacting stricter regulations, which they argue is necessary to ensure the United States remains competitive with China.
While Think Big is spending to defeat candidates backing stricter AI rules, the group is also supporting a select group of Democratic incumbents. Its recent endorsements include business-friendly moderates like Reps. Josh Gottheimer (N.J.) and Sam Liccardo (Calif.), but also New York Rep. Yvette Clarke, who has proposed legislation to impose new regulations on high-impact AI systems in areas like housing, employment, and credit.
The other side of the industry's internal debate is also supportive of AI development but argues that some safety protections are needed, and that states should be allowed to enforce their own, stronger rules. That faction is led by Anthropic, which donated $20 million to Public First Action in February. Public First Action has raised at least $50 million overall and operates through two super PAC arms—the Democratic Jobs and Democracy PAC and the Republican Defending Our Values PAC. Founded by two former members of Congress, Reps. Brad Carson (D-Okla.) and Chris Stewart (R-Utah), the nonprofit Public First Action says it wants safety standards and transparency rules, and wants to ensure that states can set their own standards rather than having regulations preempted at the federal level.
While Anthropic has positioned itself as generally supportive of AI regulation, it has pushed back when legislation proposed upfront constraints on its practices. In 2024, the company intervened in California’s AI safety bill, SB 1047, arguing that its “pre-harm enforcement” approach would be counterproductive and instead lobbied the legislature for a looser, “outcome-based” liability regime that would give companies more flexibility in how they manage risks while limiting prescriptive oversight of model development. The bill was vetoed by Gov. Gavin Newsom, who said it was overly stringent, then later signed the much weaker SB 53, which mainly requires AI developers to publish safety frameworks and issue transparency reports.
