US Midterms Deliver a Reality Check for AI and Big Data Centers

Last week’s midterms did not topple Donald Trump, but they should serve as a wake-up call for the tech elite who lined up behind him at his inauguration two years ago.

Clearly, the cost-of-living squeeze tied to Trump’s war in Iran was a major driver of the Democratic wave. But AI was also on the ballot—in many places, literally.

Dozens of measures calling for limits on datacentre expansion were passed, and checking the unchecked growth of power-hungry mega-sheds became a rallying cry.

At a campaign event, the Texas senator-elect James Talarico recalled meeting a mother who said the lights from a neighbouring datacentre were so bright that her son believed it was daytime in the middle of the night.

Whether the tech sell-off resumes this week remains to be seen, but the anti-AI sentiment from last week’s results appears to have dented at least one narrative sustaining the market boom.

Every bubble needs a story, and for AI, one chapter involves carpetting the US in debt-financed datacentres that will become necessary—and viable—as demand for the technology increases.

That now looks far less feasible. It turns out that hundreds of humming warehouses, while also warning workers they may be superseded by superhuman bots, are not great public relations.

The public was meant to set aside its qualms for the greater good that AI would bring to the economy. But the positive benefits people may derive from interacting with chatbots—the shopfront of AI for most of us—coexist with very obvious downsides.

Just to name a few: the flood of online slop; those unsettling companion glasses; and tech firms’ insistence on shoehorning the technology unbidden into every app, platform and device—the latest example of what tech writer Cory Doctorow calls “enshittification”.

At the same time, voters watched the leaders of large tech firms become obscenely rich while average consumers continued to be hammered by inflation. So perhaps it should not be surprising that the US electorate decided: enough.

Ultimately, all businesses in a democracy—even those embraced by the US president—have to operate within the bounds set by society. And the tech sector has done a poor job of winning over the public.

Elon Musk’s approach in Texas, where he poured hundreds of millions of dollars into trying to buy the electorate, has now become a selling point for Democrats.

It remains to be seen whether objections to datacentres are the start of a broader shift in opinion that will eventually tame tech companies, reminiscent of when Theodore Roosevelt swept into the White House at the turn of the 20th century to challenge the oil and railway monopolies.

However, at the very least, the opposition seems likely to put the brakes on the rapid pace of investment in building out AI’s physical infrastructure.

And this shift appears to have accelerated a market rethink that was already under way. As Joe Maher of Capital Economics put it last week: “Sentiment towards the AI trade was already looking fragile and the US equity market was showing many late-cycle signs. As such, this [the midterms] could plausibly become the catalyst that leads to a more lasting unwind of the AI equity boom.”

Two things can both be true: AI is a transformative technology that will bring welcome productivity benefits; but they will not be large enough to justify current market valuations.

As the economist Dario Perkins, of TS Lombard, put it in a brilliant recent note, written in the past tense as if describing the aftermath of a future AI crash: “What people who worked in tech and finance did not realise was that people outside the ecosystem—you know, regular people—just were not that interested in AI.

“Yes, it made web search easier (sort of); but contrary to the assumption in Silicon Valley, regular people were never going to use AI to organise their lives or ‘run workflow’. Tech insiders were living in a bubble—in at least two senses of that phrase.”

Perkins described the industry in late 2026 as essentially “$200bn in current revenues supporting more than $1tn in future commitments from just two companies [Anthropic and OpenAI].”

A reckoning was already in the offing, even without the intervention of US voters. The tech-heavy Nasdaq peaked a month ago, and a growing number of voices were sounding the alarm.

As just one example, the International Monetary Fund pointed last month to “growing evidence” that the successful rollout of AI “may not require investment on the massive scale being undertaken by ‘hyperscaler’ firms”, and that “potential economic gains may not be as large as expected”.

The IMF highlighted the growing capabilities of lower-cost Chinese models and pointed to the circular financing that has become rife within the AI sector.

More recent reports have also highlighted a string of corporate AI users either limiting staff use of costly products such as Anthropic’s Claude, or turning to much cheaper models hosted on their own servers.

The datacentre backlash appears to have brought new focus to some of these pre-existing concerns. Whether the stock sell-off persists or not, opposition has thrown down the gauntlet for the AI lobby to prove what the technology can do for societies—including those well beyond the US—that are reluctant to accept the very obvious costs.

Source link