Ed Zitron 分析美国中期选举后数据中心限制对 AI 泡沫的冲击

Ed Zitron 长文 / Where’s Your Ed At· Ed Zitron·2026-11-07 00:44· 47 分钟前AI 评分71

Premium: Fault Lines

AI 导读

Ed Zitron 撰文分析美国中期选举结果对 AI 数据中心扩张的冲击，认为选民反对 exposing 了 AI 泡沫的政治脆弱点。

正文 · 原文

I’ve generally avoided politics here and the few times that I have done so I’ve received a few pissy emails, but Tuesday night was a wonderful, abundant font of hope in what has been an unspeakably grim two years. 

We have felt land-locked in a suffocating era defined by an oafish, half-hearted fascism that lashes out in random directions, where the cruelty is the point, creating a sensation that nothing good can ever happen again — something amplified by the fact that cynical malcontents have turned everything into a casino or strip-mining operation.

And we’ve watched as the political and social infrastructure has caved to their every demand, the media has enabled them, and the existing systems have empowered them. It’s felt for most of my life that we’d never escape the weight of the eras of Reagan, Clinton and Obama, every time people have been sold further entrenchment in systemic thinking.

Nothing will ever be perfect, and I do not think every wish gets fulfilled. Tuesday night proved that things can change, for the better, and that seemingly-impossible movements can be kneecapped and nearly-entirely collapse in the space of months.

And yes, the Blue Wave could very much lead to the bubble bursting, per Bloomberg’s aptly-named “Big Tech Soars on Bet Dems Will Save AI Spenders From Themselves”:

“The market is acting like this represents the end of the AI buildout,” said Robert Pavlik, senior portfolio manager at Dakota Wealth Management. “If that means the hyperscalers aren’t spending ridiculous amounts of money to build out compute, that means they could become cash-flow positive again.”

Exit polls showed that voters strongly oppose new data centers in their communities, Keith Parker, UBS’s head of global macro equity strategy and US equity strategy research, wrote in a note to clients on Wednesday. “The blue wave will have implications for equities, with the biggest impact likely coming from an increase in data center regulation at the state level near-term,” he wrote.

A great deal of the “inevitability” of the AI bubble comes from the assumption that the rich and powerful always get what they want, and that anything they say — even the most deranged and unrealistic things in the world — will simply happen. Things like building 190GW of data centers, or, shit, 7.1GW of data centers just for Sam Altman and OpenAI only make sense when you’re under the assumption that everything in the world is at a fixed point derived from the whims and proclivities of untethered capitalism. 

While nothing has truly shifted yet, we are no longer in a political (or social) climate that rewards backing everything that the tech industry does. Texas is now a thoroughly blue state outside of the state senate, where Republicans still hold a thin majority and are already desperately trying to undermine the powers of the lieutenant-governor before Vikki Goodwin takes over in January, which spells ill for the 57.7GW of planned IT capacity in the state, more than half of which is being built by first-time developers per Bloomberg. Thirty-three separate ballot measures to restrict data centers were passed by voters or were leading as of writing this sentence per Stateline, with 21 municipalities in Ohio alone (one of the fastest-growing data center buildout geographies in the world) voting to restrict them. According to Bloomberg, “...[the] gubernatorial races from Georgia to Ohio and Pennsylvania could leave half of proposed AI data center additions across the US vulnerable to stricter local scrutiny and pushback.”

Per The Information:

In 12 states, the winning or leading candidate for governor has called for a pause on data center development, including Alaska, Georgia, Kansas, Maine, Michigan, New Hampshire, New Mexico, New York, Ohio, Oregon, Texas and Wisconsin. And in even more states, winners have also pledged other restrictions on data centers including ratepayer protections, environmental limits and the elimination of tax incentives for building the facilities.

And of the 24 incoming governors who did not call for a pause, 22 back at least one limit on data centers. Ratepayer protections, which would make sure households don’t shoulder the costs of increased power demand or energy infrastructure buildouts, were nearly universal: 34 of 36 winners support them. And 27 back restrictions relating to water usage or the environment.

It turns out that people hate data centers, are enraged with the tech industry, are tired of Elon Musk, and don’t want to feel like they are under the thumb of billionaires. The AI industry accidentally created the single-easiest bipartisan talking point by force-feeding rural communities giant, expensive monoliths explicitly built to take their jobs, and now face outright opposition in effectively every state other than the Dakotas, with little or no response other than “uhh, having data centers makes America stronger.”

The hubris of the AI bubble is the belief that it could continue to grow unfettered without seeking the consent of those who have to live with its consequences. The AI bubble threatens to take the jobs of working people, to degrade their working conditions, to increase the cost of living, and to force them to live next to noisy, fume-polluting monstrosities — and those most responsible for the existence of that bubble believe that those most keenly affected by the consequences of the technology and the infrastructure must either meekly acquiesce to their will, or will ultimately come to accept the better wisdom of Altman, Musk, Amodei, et al. 

Those responsible for the profligate splurge in AI-related spending didn’t anticipate that people — voters — would and could say no. And they absolutely fucking did, and in doing so, exposed a vulnerability in the AI bubble that I (and many others) have failed to cover with the due attention it deserved.

The AI bubble is only made possible in a climate of unrestrained capitalism. NVIDIA’s growth is entirely dependent on an unimpeded data center buildout, and without it, there’s little reason to keep buying tens of billions of dollars of GPUs every quarter. Without more capacity, Anthropic, OpenAI and other AI labs are unable to sell their stories based on infinite growth. 

And once that narrative sets in, everything begins to implode.

So, as promised, today’s newsletter is about some of the markets, industries and companies most-exposed to the AI data center capex bubble bursting, the data points to watch closely for signs of collapse, and the things that might cause things to finally fall apart.

来源：Ed Zitron 长文 / Where’s Your Ed At · wheresyoured.at

事件后续 · 1

查看事件全部

• 同事件美国中期选举蓝潮冲击 AI 行业，44 项数据中心暂停提案中 43 项通过The Verge：AI · 23 小时前