OpenAI and Anthropic Pour $185 Million Into 2026 Midterms, Backing Rival AI Regulation Super PACs

The 2026 midterm elections have become the most expensive congressional cycle in history, and artificial intelligence companies are a principal reason why. Two rival super PAC networks—one orbiting OpenAI, the other Anthropic—have poured more than $185 million into congressional races, transforming a set of nominally local contests into a national referendum on who gets to write the rules governing advanced AI. The spectacle is unusual even by Silicon Valley standards: two of the most prominent AI safety-focused companies in the world are now locked in a political spending war over what AI safety regulation should actually look like.

Two Camps, One Enormous Pot of Money

The OpenAI-aligned super PAC, Leading the Future, was bankrolled by Andreessen Horowitz and OpenAI co-founder Greg Brockman and has raised more than $75 million. It has already deployed over $20 million across races in Texas, Georgia, Illinois, and Montana, with a stated mission to “oppose policies that stifle innovation, enable China to gain global AI superiority, or make it harder to bring AI’s benefits into the world.” The group advocates for light-touch federal standards that would preempt the patchwork of state-level AI laws emerging across the country.

On the other side stands Public First, backed by a $20 million pledge from Anthropic and affiliated with related PACs including Jobs and Democracy and Defending Our Values. According to OpenSecrets, those organizations have spent $16.6 million so far on congressional races in North Carolina, Texas, and Utah. Public First’s argument is the inverse of its rival: it opposes federal preemption of state AI laws unless and until Washington produces rules with sufficient safety guardrails.

Beyond those two networks, the broader AI and tech industry has collectively directed $83 million to federal campaigns in 2025 alone and spent $50.9 million on direct lobbying—with OpenAI, Meta, Alphabet, and Nvidia accounting for the lion’s share of that figure.

The $15 Million Race That Explains Everything

The proxy battle crystallized most visibly in New York City’s 12th Congressional District, where a single congressional primary has attracted an astonishing $15 million in outside AI-industry spending, making it one of the most expensive House primaries in American history.

At the center of the race is Alex Bores, a Manhattan assemblyman who co-sponsored the New York RAISE Act, which would have required AI companies to disclose safety incidents, maintain risk inventories, and implement whistleblower protections. Leading the Future and its allies trained fire on Bores; Public First and affiliated groups rushed to defend him.

The specific provisions at stake—algorithmic transparency, AI safety incident disclosure, data-center siting regulations—are technically dry but politically explosive. Each represents a line in the sand between an industry that wants to self-govern under flexible federal standards and a contingent that believes states should retain the ability to impose stricter rules on companies operating within their borders.

Tech critic Molly White observed that the spending is “less about any single candidate and more about” communicating policy preferences to every other politician watching. The message to Congress: cross AI companies at your electoral peril.

The Ideological Rift

The spending war reveals a genuine philosophical fault line between OpenAI and Anthropic that goes beyond competitive posturing. OpenAI has publicly argued that a fragmented state-by-state regulatory landscape is unworkable—that companies cannot practically comply with 50 different AI safety regimes while also moving fast enough to stay ahead of China. It points to the Great American AI Act, currently advancing in the Senate, as the kind of unified federal framework that should supersede state efforts.

Anthropic’s position is more nuanced. The company has historically supported strong safety regulation and has been a leading advocate for mandatory pre-deployment evaluations of frontier models. But Anthropic believes federal preemption should not arrive before adequate federal standards are in place—and that states serve as an important laboratory for policy experimentation in the interim. In practice, this means Anthropic opposes a federal law that blocks states from acting before Washington has agreed on meaningful safeguards.

The irony is not lost on observers: a company whose founding mission centered on AI safety is now spending tens of millions to defeat the federal legislation championed by companies that also claim to take safety seriously. The disagreement is real, but the spending has a secondary effect of normalizing the idea that AI companies should be active participants in shaping their own regulatory environment.

What’s Actually Being Decided

The bills and regulatory decisions at stake in the 2026 cycle include:

Federal AI preemption: Whether a single national AI framework would nullify state laws like California’s Transparency in Frontier AI Act, Colorado’s SB 26-189, and New York’s RAISE Act—all of which impose safety and disclosure obligations that go beyond current federal proposals.

Safety incident disclosure: Whether AI companies would be required to report capability-related safety incidents to regulators within a defined window, similar to cybersecurity breach notification requirements.

Data-center siting: Whether federal permitting processes for AI infrastructure projects should override local environmental and zoning review.

Algorithmic transparency: Whether companies deploying AI in high-stakes decisions—hiring, lending, healthcare—must publish auditable accounts of model behavior.

The outcome of the midterms will determine which of these proposals advance and which die. A Congress shaped by Leading the Future’s spending would likely pass a preemptive federal framework with limited safety requirements. One shaped by Public First would be more receptive to giving state regulators a longer runway.

The Bigger Picture

The scale of AI political spending in 2026 marks a qualitative shift from the industry’s prior posture. As recently as 2024, AI companies participated in Washington primarily through traditional lobbying. The pivot to eight-figure super PAC spending—directly targeting candidates based on their regulatory positions—reflects both the enormous financial stakes of regulatory outcomes and the speed with which the industry has adopted the political playbook of predecessors like financial services and pharmaceuticals.

What makes this moment unusual is that OpenAI and Anthropic have both, at different times, publicly called for robust AI regulation. Both signed voluntary commitments at the White House in 2023. Both have published extensive safety frameworks. Yet here they are, on opposite sides of a $185 million fight over what “safety regulation” should actually require—a fight that will shape the legal environment for AI development for the next decade.

Whether that money produces better AI policy, or simply more compliant politicians, remains the central question of the 2026 AI regulatory moment.