Why Tech Billionaires Couldn't Buy the 2026 Election | Hard Fork
The midterms showed tech money and tech-right politics do not buy elections: AI and data centers proved toxic, Musk's spending backfired, and only a few donors, like Hoffman and Brin, got what they paid for.
Watch on YouTube ↗Charlie's summary
- ▶ 2:01
AI is politically toxic because the pitch is bad and the people rolling it out seem untrustworthy
Ezra cites exit polls: 56% said AI and data centers were a major factor in their vote, 57% said AI would hurt society, and 76% opposed a data center nearby. Trump took the Saxs/Andreessen line of “let them rip,” which put Republicans on the wrong side of opinion. Ezra says the basic sell is that AI might kill everyone and probably take your job. Regulation could have built trust.
- ▶ 5:34
Revealed preferences don't prove people want AI: heavy usage hasn't bought public acceptance
Erin says many VCs argue people distrust big tech but love their iPhones, Facebook and Amazon, so ChatGPT's billion users mean they'll accept AI too. She thinks this election shows that reading is wrong, since voters were angry enough to reject AI outright. Ezra adds that many AI insiders privately say they would push the pause button but feel it can't be stopped. Politicians who don't buy that inevitability want to pace it and regulate it.
- ▶ 13:43
Data centers were a spiky wedge issue that crossed party lines where the buildout is heaviest
Data centers ranked low in issue polls but won on a yes/no question. Ezra and Erin say they mattered most in states building them: Texas, Ohio, Georgia and Arizona. Democrats hit Ramaswamy and Husted on the issue in Ohio, and it worked as a wedge that won over Republicans. Erin says investors fear 10 to 15 gigawatts of the roughly 45 gigawatts due next year could now be at risk, especially in Texas.
- ▶ 18:49
A data center slowdown might actually be good for markets by pacing capex to real usefulness
Ezra offers a contrarian take. Markets bounced Wednesday despite investor panic. He suspects a collective-rationality problem in which everyone overinvests in data centers for fear of being left behind. Externally paced capex could weed out fly-by-night players and push Anthropic and OpenAI toward making products economically useful instead of racing to superintelligence. What markets need is alignment between buildout speed and adoption.
Give your email to unlock the rest of the key points — for every episode.
Do more with Charlie
Free — sign in with your email, no password.