The AI industry has decided that the 2026 midterms are too important to leave to voters. So they are buying them instead.

Here is the twist nobody saw coming: the two biggest AI companies in the world are spending hundreds of millions of dollars on the same election, on completely opposite sides.

Anthropic, the company behind Claude, dropped $20 million into a group called Public First Action that supports candidates who want to regulate AI. OpenAI co-founder Greg Brockman and his wife put $25 million into Leading the Future, which backs candidates who oppose new AI laws. Public First Action has raised around $50 million total. Innovation Council Action, tied to two of Trump's AI advisors, announced it would spend at least $100 million.

That is not a typo. The companies building the most powerful AI systems on Earth cannot agree on whether those systems should be regulated.

The Split That Defines Everything

Anthropic's position is straightforward: AI is risky, most Americans want regulation, so fund the candidates who will write the rules. OpenAI's position is equally blunt: AI is an arms race, regulation slows you down, and America cannot afford to fall behind China.

OpenAI published an economic blueprint last year comparing AI regulation to the rules that "stunted" the British car industry while America's grew. Their argument: light touch wins. Anthropic's counter: light touch is how you get a Facebook-scale disaster, except this time the technology can think.

Rep. Alexandria Ocasio-Cortez was characteristically direct about the spending. "Their money will end up being toxic anyway," she posted. "People are catching on."

Follow the Money

FEC filings show the money is already flowing. Committees linked to Leading the Future have spent more than $500,000 each supporting Republican House candidates in North Carolina and Texas. Public First Action has run ads thanking both Democrats and Republicans for their AI records, including Rep. Nikki Budzinski, Sen. Marsha Blackburn, and Rep. Josh Gottheimer.

University of Rochester professor David Primo nailed the dynamic: "Companies have always tried to shape regulations in their favor. What we are seeing now is that the big companies are not united."

The Take

This is the most revealing thing happening in AI right now. Not model benchmarks. Not fundraising rounds. The fact that the two leading AI companies have such fundamentally different views of their own technology that they are spending nine figures to elect different kinds of politicians.

Anthropic genuinely believes AI is dangerous enough to need guardrails. OpenAI genuinely believes guardrails will hand the future to China. They cannot both be right. But the midterms will decide which vision becomes American policy for the next decade.

The stakes, as Professor Primo put it, are enormous: "Once a regulatory system gets entrenched, it is really hard to change it." Whoever wins in November does not just win an election. They lock in the rules of the AI era.