TL;DR
- Campaign Financing: AI-lab linked lobby groups Leading the Future and Public First are funding candidate advertising to influence who writes AI rules in Congress.
- Independent Expenditures: Selected federal election committees spent $57.48 million outside candidates’ campaigns, according to AI news publication Transformer’s September 30 tracker snapshot.
- Policy Education: AI developer Anthropic says its $40 million for Public First Action, a Public First network nonprofit, cannot be used to influence candidate elections.
- Safety Priorities: The fight involves enforceable model safeguards and whether national standards should replace state AI rules.
AI-focused super PAC networks are spending tens of millions of dollars on the 2026 U.S. midterms to influence who writes rules for artificial intelligence. Leading the Future and Public First are buying political reach across party lines, backing candidates who fit their competing policy priorities rather than one party’s ticket.
On September 30, Transformer’s campaign-finance tracker displayed $57.48 million in independent expenditures by the AI-focused federal committees it follows. That measures money spent supporting or opposing candidates, including advertising and other voter communications. Fundraising and payments between committees measure earlier stages of the same financing process.
How Funding Becomes Candidate Advertising
Super PACs are political committees that can raise unlimited contributions from individuals, companies and other permitted donors. They finance candidate advocacy outside the candidate’s campaign. An independent expenditure pays for communications explicitly supporting or opposing a candidate, without coordination with that candidate or a political party committee.
Leading the Future is both a broader political operation and the name of a federal fundraising committee. Its backers include OpenAI president and co-founder Greg Brockman, his wife Anna, and venture investors associated with Andreessen Horowitz, a firm that invests in OpenAI, the developer of ChatGPT.
Tech Influence Watch, independent researcher Molly White’s campaign-finance project, lists $75.1 million in direct contributions to the federal committee during the 2025–2026 election cycle. Its individual donor entries attribute $50 million to Marc Andreessen and Ben Horowitz combined, and $25 million to Greg and Anna Brockman combined. OpenAI has said it does not finance or direct Leading the Future, and that employees can donate personally.
During April–June 2026, Leading the Future paid $10 million each to Think Big and American Mission. Think Big promotes Democratic candidates; American Mission supports Republicans. Leading the Future reported no independent expenditures of its own during that quarter.
Think Big’s September 25 filing records $901,161.12 in independent expenditures, with payments to Lantern Production Consultants and Manhattan Strategies. The fundraising committee supplies affiliated committees, which pay outside vendors for electoral communications.
Anthropic’s Donations Have a Different Purpose
Anthropic, OpenAI’s main competitor, supports Public First Action, a nonprofit focused on AI policy and public education. The organization is distinct from the electoral committees associated with the Public First network.
In July, Anthropic said its two $20 million donations brought its support to $40 million, restricted to public education and policy work. It said both donations cannot be used to influence the election of any federal, state or local candidate.
The wider network also finances election activity. Public First Action had raised $100 million by August 20, its spokesman told Reuters, with about half of its funds going toward elections through two aligned PACs. He declined to identify the other donors.
Jobs and Democracy, a Public First-aligned electoral committee supporting Democrats, reported a $3 million payment from Public First Action Inc. on April 15. A separate super PAC, Public First, paid it another $1.3 million between May 22 and June 3. Defending Our Values is the network’s Republican-facing electoral committee.
The Dispute Is About Which Safeguards Apply
Leading the Future seeks to elect candidates who advance AI development. Its launch statement framed the priority as preserving U.S. leadership and resisting policies it argues would delay innovation or give China an advantage. The network favors a national approach to AI standards and safeguards.
Public First Action supports state and local safeguards when adequate federal protections are absent. Anthropic’s February donation statement makes the condition explicit: it opposes federal rules displacing state AI laws unless Congress enacts stronger safeguards.
Anthropic’s July proposal calls for independent evaluations of the most capable AI models, enforceable safety duties and government authority to slow or block deployments posing a serious risk of catastrophic harm. The company argues that governments need time to build the capacity to manage risks to critical systems while retaining AI’s benefits.
In July, OpenAI advocated common state safety rules that could develop into national standards, including incident reporting and independent audits, alongside federal evaluation of advanced models.
The cross-party strategy has a precedent in cryptocurrency. Reuters describes AI interests adapting the Fairshake network’s approach of backing policy allies regardless of party. Rick Claypool, research director at corporate-accountability group Public Citizen, told Reuters such spending leaves less room for voters’ household concerns. Supporters cited in the report argued that emerging industries deserve the political voice established sectors already have.
New York Primary Shows the Rivalry and Its Limits
New York Assemblyman Alex Bores became a target after co-sponsoring the Responsible AI Safety and Education Act, or RAISE Act, which requires safety-incident reporting and information about developers’ safeguards. In the Democratic primary for Manhattan’s 12th Congressional District, Think Big opposed him while Jobs and Democracy backed him.
Think Big’s anti-Bores advertising argued that state AI laws would create conflicting rules and impede innovation. Rival groups countered with pro-Bores ads, mailers and texts. The disagreement over regulation became a contest over which candidate would represent the district.
Alex Bores is wrong for AI & wrong for Congress ⬇️ pic.twitter.com/6MCsKBz8Gf
— Think Big PAC (@thinkbig_pac) December 10, 2025
Bores lost the June 23 primary to fellow assemblyman Micah Lasher, who also co-sponsored the state AI law. Bores accused the wealthy backers opposing him of trying to make other politicians afraid to challenge them.
White told NPR that a single junior member of Congress might have limited influence over eventual legislation, while PAC spending could send a message to other candidates considering stricter AI rules.
