Anthropic and OpenAI take their beef to the midterm elections

Hello and welcome to Regulator, a newsletter for Verge subscribers about the car crashes piling up on a daily basis at the Washington-based intersection of technology and politics. If you’re not a subscriber, sign up for our fine editorial enterprise today, especially as we process the end of Musk v. Altman. And if you have any tips about impending or hidden Washington car crashes, send ’em over to tina.nguyen+tips@theverge.com. A quick note: Regulator will be on hiatus for the next two weeks while I take a much-needed vacation. Unfortunately, this means I’ll be missing the public release of Pope Leo XIV’s encyclical on humanity in the age of technology, which I have been hearing about for months, but I anticipate that the rest of the Verge staff will be all over it, so bookmark us! Heated rivalry, AI super PAC edition Here’s a weird sign of the AI super PACs becoming their own political behemoths: They’re now becoming their own political weaknesses. On Tuesday, New York Democrat congressional candidate Alex Bores, whose campaign leans heavily on promoting AI regulation, challenged Leading the Future — the $100 million pro-AI super PAC funded by Palantir’s Joe Lonsdale, Andreessen Horowitz, and OpenAI’s Greg Brockman — to an in-person, real-world debate. In a press release, the Bores campaign laid out their conditions: Leading the Future could pick the moderator, it could pick its own representative, but it has to commit to a debate before the June 23rd primary. The likelihood of this debate taking place is slim to none. (Leading the Future declined to comment about the debate challenge.) Still, it’s a rapid escalation in a phenomenon I’ve been tracking for months: AI industry super PACs gaining their own political reputations, reflecting the companies and founders who fund them, then using those reputations to fight each other. When Leading the Future was launched last year, it was fairly typical for a super PAC, in that it was backed by several wealthy individuals and companies with shared policy goals, operating on both the state and federal election level. (It was, of course, politics on steroids: The Supreme Court famously ruled in Citizens United that corporations had the right to free speech, leading to the creation of special campaign finance vehicles that allowed companies and wealthy donors to donate unlimited sums toward political advocacy groups.) But shortly afterwards, Meta announced that it was launching its own AI-focused super PACs — a sign that the company’s AI interests, political and otherwise, were not necessarily aligned with the entities funding Leading the Future. Over time, LTF came to be viewed as a vehicle not for the general AI industry, but for OpenAI specifically. (Several of LTF’s backers are investors in the frontier AI company.) That perception was solidified earlier this year, when Anthropic donated $20 million to Public First Action, a bipartisan super PAC network that’s backing Bores. Legally, super PACs are not allowed to coordinate with candidates on things such as ad buys and messaging. But while it’s normal for companies to use super PACs to back candidates against other candidates, it’s rather innovative, perhaps, for companies to use super PACs to attack their corporate rivals (and the candidate is, in some ways, incidental). Now, Public First is synonymous with Anthropic and “doomerism” (in LTF’s terms), and LTF, as Bores put it, is now known as “the Marc Andreessen-Greg Brockman-Joe Lonsdale-backed Leading the Future super PAC.” And the beauty of non-coordination campaign finance laws is that Bores, the coauthor of the New York state RAISE Act, can plausibly distance himself from whatever Anthropic-funded political shenanigans are going on on his behalf. (Corporate money is corporate money.) Dark money? More like dork money: We haven’t even gone into the shadier world of campaign finance vehicles, including one that might start firing on LTF in order to appease Trump. (Apparently, according to The New York Times, Leading the Future is too bipartisan to be trusted by Republicans.) In March, a pro-AI, political advocacy messaging nonprofit called Innovation Council Action revealed itself to the public, run by Donald Trump’s former adviser Taylor Budowich and already boasting a $100 million war chest. Crucially, it received the “blessing” of a recurring Regulator character, David Sacks, former White House special adviser on AI and crypto. ICA will be focused explicitly on promoting Trump’s AI agenda, and that means addressing a new issue inside the Republican Party: populist-leaning candidates unwilling to cave to whatever pro-industry positions Donald Trump has been convinced to repeat at any given time. (Who’s pushing this agenda? We currently do not know. ICA is known as a “dark money nonprofit,” which means that unlike super PACs, its donors do not legally have to be disclosed.) Time to regulate prediction markets now! The latest technology causing a meltdown in