youtube.com web signal

Vergecast: tech billionaires' 2026 midterm bets flopped

TL;DR

  • Hosts Nilay Patel and David Pierce argue the midterm results reset the political backdrop for AI regulation and data center expansion.
  • Elon Musk spent close to $500 million on the cycle, including $216 million routed through America PAC, with his backed candidates losing.
  • Patel says AI products aren't good enough to justify the data center costs voters are being asked to absorb in their communities.

On The Vergecast, hosts Nilay Patel and David Pierce argue that the 2026 midterms broke in a specific direction for the AI industry: against it. Voters delivered Democratic wins up and down the ballot, and dozens of local ballot measures restricting data center construction passed, even where industry-aligned PACs outspent the organizers opposing them.

The money trail is the striking piece. Elon Musk spent nearly $500 million across the cycle, including roughly $216 million routed through America PAC, according to reporting in the SF Standard. Marc Andreessen and Ben Horowitz were part of the same push. Their candidates lost anyway. Ken Paxton lost a Texas Senate race to Democrat James Talarico despite Musk spending heavily against him; Darline Graham lost her South Carolina Senate bid to Annie Andrews. In Ohio, Sherrod Brown held on while a crypto-aligned PAC burned about $55 million trying to unseat him.

The ballot-measure column is where the political economy of AI itself got a direct vote. At least 35 local measures banning or limiting data center construction passed nationwide, and exit polling cited by the SF Standard put the share of voters calling AI or data centers "a crucial or major factor" in their congressional vote at 56%. California's Prop 40, a 5% wealth tax on billionaires, drew opposition spending of around $340 million from a coalition including Sergey Brin, Eric Schmidt, and John Doerr; it was trailing on election night.

Patel's reading on the podcast is that the industry's problem is product, not politics. "The products aren't good enough to sell people on the costs," he says. "You cannot tell people that you have to listen to the whine of a data center in their backyard for AI overviews and some of the most hilarious deep fake slop in history." The fix he frames is the inverse of marketing: the industry "will not get out of it until they make better products… that just using these products makes your life better."

Two researchers we follow posted the Vergecast link within hours. Sen. Jon Ossoff's line, quoted in the SF Standard reporting, captured the mood on the Hill: Silicon Valley leaders "dig bunkers for themselves and warn that the artificial intelligence they're creating risks human extinction."

Shared on Bluesky by 2 AI experts