The headline going around is that OpenAI employees gave roughly $296,000 to Democratic groups in the second quarter and nothing to Republicans.
That figure is accurate — about $71,000 individual and $225,000 in PAC contributions, per FEC filings. It is also close to meaningless as a statement about the AI industry's politics, and the same filings contain the reason.
The numbers the headline leaves out
Google staff: over $608,000, about 72% to Democrats, 1.8% to Republicans.
Meta staff: over $245,000, about 65% to Democrats, 22% to Republicans.
Anthropic staff: roughly $1,616,000, about 69% to Democrats and 12.9% to Republicans — a Republican share roughly nineteen times Google's and four times Meta's. Its employee PAC reportedly raised over $275,000 in Q2 on a 55/45 Republican-to-Democrat split.
And OpenAI's own president and co-founder, Greg Brockman, gave a reported $25 million to a pro-Trump super PAC in 2025 with his wife.
So: the company whose employees gave zero to Republicans last quarter is the company whose president wrote the largest pro-Trump check in the sector. “Republicans got nothing” describes one firm, one quarter, one category of giving.
The fight that's actually happening
The largest single OpenAI-linked contribution was $200,000 to DreamNYC, a super PAC backing Alex Bores — a New York assemblyman known for pushing AI regulation, in a House run he lost.
More than $245,000 from current and former OpenAI employees went to Guardrails Alliance, a super PAC created to counter Leading the Future — a committee Brockman helps fund.
Read that twice. Rank-and-file employees are financing organized political opposition to their own president's political project.
Cornell's Sarah Kreps put it plainly: this is not the tech industry versus its outside critics. It is a fight inside the industry about what responsible AI policy looks like.
That is a far more interesting story than a partisan tally, and it's the one the tally obscures.
Why I think the framing matters
If you sort this by party, you conclude that AI is a Democratic industry and regulation is a Democratic project. Then you decide your position by tribe.
Sort it by the actual dispute and it looks different: a substantial number of people who build these systems, who have seen the internals, are spending their own money to make the rules stricter — against the wishes of executives at the same companies. The Republican giving concentrates at the lab with the loudest public safety posture. The largest single check in the sector went to a pro-Trump PAC from a man whose employees are funding the opposition.
Nothing about that maps onto left and right.
What I'd tell a reader to watch
Not the party column. The direction of travel between employees and executives at the same firm.
When the people closest to a technology start funding constraints on it, over the objection of the people selling it, that's a signal. It may be self-interest, it may be genuine alarm, and honest people can read it either way. It is not partisanship.
The industry with the most at stake in AI regulation is divided against itself. That's the story.
Add your perspective
Right: a workforce donating this lopsidedly builds products that reflect it, and users are right to be skeptical of the output. Left: employees pushing for guardrails against their own leadership is the most encouraging thing in this data. Center if you think employee giving tells you little about a company and the executive checks tell you more.
