DooDooLamb News
The $27 million Al proxy war over Alex Bores ends in a draw
Brief published June 26, 2026 ยท Original source published June 24, 2026
Original reporting by Tina Nguyen at theverge.com.
Automated brief. Verify important details at the original source.
Two AI Companies Spent $27 Million Fighting Each Other in a Congressional Race and Both Lost
The AI industry, not content with disrupting every other sector of the economy, has now successfully disrupted the concept of winning. Last night, in New York's 12th Congressional District, Anthropic and OpenAI burned through a combined $27 million trying to get their preferred outcomes in a Democratic primary, and both companies walked away with absolutely nothing to show for it except a smaller bank account and a valuable lesson they will definitely not learn.
What Actually Happened Here
Alex Bores is a New York state Assemblyman who had the specific misfortune of becoming a human chess piece in a proxy war between two of Silicon Valley's most well-funded theological disputes about artificial intelligence. He narrowly lost the NY-12 Democratic primary. Not because he was a bad candidate, but because the political physics of being publicly championed by a pro-AI super PAC in a Democratic primary in New York City in 2025 are roughly as favorable as announcing you were personally mentored by a crypto exchange.
Here is how the money broke down, as best as the public record shows: Anthropic's affiliated interests backed Bores, betting that a tech-friendly congressman would be useful in Washington as AI regulation starts to take shape. OpenAI's side worked against that outcome, apparently deciding that an Anthropic-aligned legislator was worse than no AI-friendly legislator at all. Twenty-seven million dollars. A primary race. In New York. Bores lost by a margin thin enough that you could attribute it to a rainy Tuesday and a bad subway delay, which means both sides can technically claim the other side failed.
This is what the industry calls a "draw," and what everyone else calls "lighting money on fire in a parking lot and arguing about whose fire was more strategic."
The Deeper Absurdity Here Deserves a Moment
To understand why this is genuinely funny rather than just expensive and sad, you have to understand what Anthropic and OpenAI actually are. They are two companies that both believe, with varying degrees of sincerity depending on the earnings call, that they are building one of the most transformative and potentially dangerous technologies in human history. Their entire public-facing brand is built on careful, responsible development of AI, on moving thoughtfully, on taking the risks seriously. Anthropic's founding mythology is literally that its founders left OpenAI because they thought OpenAI was being insufficiently careful.
These are the companies that just spent $27 million on a Congressional primary in the style of competing pharmaceutical lobbies fighting over a committee chairmanship. Claude and ChatGPT are out here writing poetry and summarizing legal documents while their parents are in the backyard throwing lawn furniture at each other over a state assemblyman.
"We believe in AI that benefits humanity," said every corporate communications team simultaneously, while the super PAC check cleared.
The Political Calculation That Made Sense on a Whiteboard
To be fair, and fairness is important even when it's inconvenient for the joke, the underlying logic here is not insane. Washington is actively trying to figure out what to do about AI, and the people who write those regulations will matter enormously. Bores was genuinely tech-literate, having worked in the space before his political career, which in Congress puts him in a cohort so small you could fit them all in a medium-sized conference room. AI companies watching the legislative calendar are not wrong to think that having allies in Congress beats not having them.
The problem is the execution, which had all the subtlety of a product launch email sent to your entire contact list. When a pro-AI super PAC runs ads for your candidate in a New York Democratic primary, you have not given that candidate a gift. You have handed his opponents a flyer that says "this person is backed by the tech industry" and then been surprised when New York Democratic primary voters, who have extremely strong opinions about the tech industry, treat it accordingly. Bores' vote total reportedly went up after the attacks. The super PAC ads against him may have helped him. The whole operation had the strategic coherence of a RAG pipeline pointed at the wrong database, returning confident answers that were completely disconnected from the actual question being asked.
What This Means for Everyone Who Isn't a Super PAC
For developers and founders watching this, the practical read is this: the AI industry's political strategy is still in the "move fast and see what breaks" phase, which is a phase that works considerably better for software than for electoral politics. Twenty-seven million dollars that did not elect anyone is twenty-seven million dollars that also did not fund safety research, did not improve inference costs, and did not do anything for the actual humans using these products.
The honest version of what happens next is that both companies will continue spending on politics, because the regulatory stakes are real and the money is not. They'll get better at it, eventually, the way every industry eventually learns to launder its influence more elegantly. For now, the AI policy apparatus in Washington remains genuinely unsettled, nobody won a seat over it, and the next round of congressional races is already being quietly scouted.
The draw is temporary. The $27 million is gone forever.
The industry that promises to solve intelligence as a technical problem just discovered that political intelligence follows different rules, operates on different latency, and cannot be improved by throwing more compute at it. Both Anthropic and OpenAI built systems that learn from their mistakes at superhuman speed. It would be a real shame if neither company applied that capability to the lesson sitting right in front of them.