OpenAI vs Anthropic - The $125M Battle for Congress

"We are actively working on a broader super PAC strategy to ensure that pro-regulation candidates have the resources to win."
Jack Clark, Anthropic co-founder, on the company's political spending

Two of the most powerful AI labs in the world are no longer content to lobby Congress - they're buying it. Super PAC networks backed by OpenAI insiders and Anthropic are pouring over $125 million into the 2026 midterms, turning a technical debate about model safety into one of the most expensive proxy wars in the history of American campaign finance.

Neither side mentions AI in a single advertisement.

TL;DR

Two rival AI super PAC networks have amassed over $125 million for the 2026 midterms
OpenAI co-founder Greg Brockman and a16z back "Leading the Future," which opposes AI regulation
Anthropic donated $20 million to "Public First," which supports stricter AI safeguards
The central battleground is New York's 12th district, where a pro-regulation assembly member faces $1.5 million in attack ads funded by the pro-innovation side
None of the ads running in these races mention artificial intelligence

The Two Sides

Leading the Future - OpenAI, a16z, and the Deregulation Machine

Leading the Future had $39 million banked at the end of 2025 and is targeting a $100 million total war chest. Its funders read like a who's-who of Silicon Valley's pro-acceleration wing: OpenAI co-founder Greg Brockman and his wife Anna, venture capitalists Marc Andreessen and Ben Horowitz of a16z, Palantir co-founder Joe Lonsdale, veteran investor Ron Conway, and AI search company Perplexity.

The group operates through two connected entities - "Think Big" for Democrats and "American Mission" for Republicans - letting it work both sides of the aisle with a unified strategic agenda. That agenda is straightforward: oppose strict AI regulation, push for federal preemption of state-level AI laws, and elect candidates who won't put guardrails on AI development.

The strategy mirrors the broader industry campaign. As we reported in our coverage of Trump's executive order preempting state AI laws, the push for federal preemption is now a core priority for the largest AI labs.

Public First - Anthropic and the Pro-Regulation Counter-Network

On the other side, Public First is a nonprofit co-founded by two former members of Congress: Brad Carson, a Democrat from Oklahoma, and Chris Stewart, a Republican from Utah. The bipartisan framing is deliberate - it signals that AI regulation is not a partisan issue.

Anthropic has confirmed a $20 million donation to Public First. The group operates through "Jobs and Democracy PAC" for Democratic primaries and "Defending Our Values PAC" for Republican ones. Anthropic co-founder Jack Clark has signaled that the company views this as ongoing, not a one-time contribution.

Anthropic's bet here is consistent with its public identity as a safety-first lab. As we have covered in our analysis of Anthropic's safety commitments and research, the company has consistently argued that AI systems need more oversight, not less - even when that position creates friction with other labs and with its own commercial interests.

A ballot being cast into a voting box - the 2026 midterms have become a proxy war for AI regulation Neither AI super PAC mentions artificial intelligence in its campaign ads. Both are spending millions to shape who writes the rules.

The Battle Lines

New York's 12th District - Ground Zero

The most watched race is in Manhattan. New York Assembly member Alex Bores is running for the congressional seat being vacated by retiring Rep. Jerry Nadler. Bores is a former Palantir data scientist who quit the company in 2019 over its work with Immigration and Customs Enforcement. At the state level, he was the chief architect of the Responsible AI Safety and Education Act - the RAISE Act - which was signed into law in December 2025.

That record makes Bores exactly the kind of candidate that Leading the Future wants to remove. Think Big has already spent over $1.5 million in TV ads and other messaging in the district attacking him. Public First has responded with $450,000 supporting his campaign.

Bores was named to Time's 100 AI list in 2025. He's running explicitly on AI safety as a congressional priority. He's also running in a Democratic primary in a heavily Democratic district, which means the primary - not the general - is the real contest. That's where the super PAC money is landing.

North Carolina, Illinois, and Texas

The spending isn't limited to New York:

Race	Leading the Future	Public First	Net Spend
NY-12 (Bores vs. field)	$1.5M+ against Bores	$450K for Bores	LTF leads by ~$1M
NC (Foushee re-election)	Not disclosed	$1.6M for Foushee	Public First leads
Illinois (Jackson/Bean)	$1M+ each for candidates	Not disclosed	LTF uncontested
Texas (Gober)	Not disclosed	$500K for Gober	Public First uncontested
NC swing district (Buckhout)	Not disclosed	$500K for Buckhout	Public First uncontested

The pattern that emerges: Leading the Future is concentrating firepower on defeating the one candidate most closely identified with AI regulation. Public First is spreading its money across a wider slate, betting on volume - elect enough members who support guardrails and the committee math changes.

Hidden in Plain Sight

The most telling detail about this entire operation is what the ads do not say. Both networks have made a deliberate choice to run campaigns about immigration, healthcare, and opposition to Trump. Not AI. Not regulation. Not safety.

The reason is simple political arithmetic. In Democratic primaries, AI regulation isn't a top-of-mind voter issue. Voters respond to pocketbook questions. So the money flows in under the cover of bread-and-butter politics, shaping the makeup of the next Congress without ever triggering a public debate about why these particular candidates were targeted.

Stacks of hundred dollar bills - AI companies are redirecting billions in capital toward political influence The combined war chest from AI-backed super PACs exceeds $125 million - larger than many national party committee budgets.

Who Benefits, Who Pays

Stakeholder	Impact	Timeline
OpenAI / a16z portfolio	Reduced regulatory risk, federal preemption of state AI laws	2027 onwards if candidates win
Anthropic	Pro-safety Congress, potential mandatory red-teaming requirements	2027 onwards
Alex Bores and state-level regulators	Career-defining race, national profile win or lose	Primary election Q2 2026
American voters	No public debate about AI in races being decided on AI grounds	Ongoing
Small AI startups	Regulation shaped by incumbents who can absorb compliance costs	2027-2028

The incentive structure is worth examining. For large AI labs, political spending in the millions is a rounding error compared to the cost of complying with - or fighting - state-by-state AI regulations. California's Transparency in Frontier AI Act and Texas's Responsible Artificial Intelligence Governance Act are already on the books. A Congress that passes federal preemption legislation would wipe those out in one vote.

That's the business case for the $100 million war chest. It isn't philanthropy. It's regulatory arbitrage at scale.

What Happens Next

The first real test comes in Q2 2026 when several of these primaries conclude. If Leading the Future succeeds in defeating Bores, it sends an unmistakable signal to every state legislator in the country: sponsor an AI safety bill and the industry will end your political career.

If Public First prevails - if Bores wins and the pro-regulation slate holds - it confirms Anthropic's theory that safety-minded candidates can survive the AI money machine.

The Federal Trade Commission is separately due to issue a policy statement by March 11, 2026 on how federal law applies to state AI regulations - a deadline set by the Trump executive order we covered earlier. Depending on what that statement says, the political landscape could shift markedly before a single vote is cast.

What's certain is that AI companies have concluded that lobbying isn't enough. They're now in the business of selecting who sits in Congress. The $125 million figure will almost certainly grow before November. And none of the ads will mention AI.

The deeper question is what this spending means for the democratic process. Both sides frame their involvement as a matter of national importance - one says innovation requires freedom from patchwork state rules, the other says frontier AI is too consequential to leave unregulated. Both arguments have merit. But when the entities most affected by regulation are also the ones spending nine figures to elect the regulators, the line between stakeholder and principal has effectively disappeared.

Sources: