Half of the nation’s upcoming data center capacity is in four states that flipped control to Democrats.

The Democratic Party’s takeover of Washington and key state capitals — coupled with an abrupt financing chill — threatens the artificial intelligence boom that has been driving U.S. economic growth and stock values.

Investors worry that Democrats, emboldened by their landslide win on Tuesday, will slow the nationwide buildout of AI data centers.

The shift in political power in states such as Texas, Ohio and Michigan comes as Wall Street’s willingness to finance the industry’s trillion-dollar AI plans was already waning.

Bond investors on Friday demanded yields of more than 16 percent to lend money to CoreWeave, a data center company that paid around 10 percent for borrowed money earlier this year. Several data center companies — including one backed by chipmaker Nvidia — abandoned plans for initial stock offerings this month as investor sentiment cooled.

Wall Street’s rethinking accelerated this month as it became clear that growth forecasts for top AI companies were overly optimistic. In advance of its planned stock market debut, Anthropic last week said its third-quarter revenue growth dropped to 47 percent from 144 percent in the previous quarter.

Some analysts expect that number to fall by half in the final three months of the year. Growing investor doubts about AI emerged on Thursday when a tech stock plunge erased roughly $2.5 trillion in wealth.

“Some of the cracks in the AI thesis are starting to emerge,” said Ara Kharazian, lead economist for Ramp, a spend-management company. “The market is trying to parse out the winners and losers here.”If the AI industry’s investment math no longer works, the U.S. economy could take a hit. Sagging stock prices would threaten consumer spending on goods and services, which accounts for nearly 70 percent of the economy. Delays in AI infrastructure investment could knock out one of the props beneath strong recent growth in output and hiring.

On Friday, President Donald Trump weighed in, saying he was “100%” in favor of data centers and would “work with the many States in the U.S. that want to become RICH by having them.”AI development for months has been shadowed by warring narratives.

Top executives at leading U.S. labs such as Anthropic and OpenAI describe the rise of machine intelligence as a seminal moment in human history that will reshape work, leisure, governance and health.

Yet amid the accolades and self-promotion, some analysts warned that the hype had inflated an investment bubble. Shares of Nvidia, the leading AI chipmaker, rose by more than 430 percent between October 2023 and October of this year before shedding nearly one-quarter of their value over the past month. Chipmaker Intel has lost half its market value in the past month.

Since the debut of ChatGPT four years ago, AI has advanced from an economic curiosity to the center of the $33 trillion U.S. economy. In recent years, the “Magnificent 7” AI stocks — Amazon, Alphabet, Apple, Microsoft, Meta, Tesla and Nvidia — have been among Wall Street’s best performers.

Together, these seven stocks account for roughly one-third of the broader S&P 500 index’s market value.

Rising stock values paid economic dividends: Many Americans used investment gains to support their spending, according to the research from JPMorgan Chase Institute. Between 2019 and 2025, the share of consumers tapping their portfolios for cash doubled. Investment gains now support roughly 7 percent of household spending.

That’s important because amid muted after-inflation income growth, consumers have been struggling to maintain their living standards. The personal savings rate is about half what it was in January 2024, according to the Commerce Department.

Unusually large tax refunds this year, thanks to the president’s signature tax legislation, helped boost spending. But as taxpayers adjusted their paycheck withholding to take account of the new law, refunds should return to typical levels.

Seeking to juice their market gains, some Americans have borrowed money from brokerage firms to invest, a risky practice that can magnify gains or losses.

As of September, Americans had borrowed about $118 billion from their brokerages, up from $27 billion one year ago, according to Ian Shepherdson, chief economist with Pantheon Macroeconomics.

Investors can continue borrowing against their securities portfolio only if stocks keep rising.

“If you want to keep borrowing to keep spending more, you need stock prices to keep rising. Just leveling off isn’t enough. So that’s a very clear and present danger,” said Shepherdson.

It would probably take a serious stock market decline, well beyond the S&P 500’s more than 5 percent loss over the past month, to ignite broader economic problems, according to Joe Brusuelas, chief economist of RSM, a Chicago-based consultancy.

For now, that is a worst-case worry. The economy grew between July and September at a rapid 3.4 percent clip, and employers created 165,000 new jobs in October, the Labor Department said Friday. Stock prices rebounded Friday, easing investor fears of an imminent crisis.

Still, many observers see the AI industry’s political problems and financial arrangements as reasons to worry.

“The boom is in its final innings,” Capital Economics in London told clients.

The Democrats who seized control of Capitol Hill and won gubernatorial races in Texas, Ohio and Georgia — states with major data center projects — are far more skeptical of the industry than the laissez-faire Republicans who preceded them.

Trump has pursued a largely hands-off approach to AI development. He has eased federal permitting for data centers and warned against “regulation that’s going to stifle a very important industry.”The prospect of that changing smacked Elon Musk’s AI investments this week. After Democrat Gina Hinojosa won the Texas governor’s race, the stock of Musk’s Space X, which includes his Grok AI business, shed almost 13 percent of its value before recovering a bit on Friday.

Investors fear that Democrats’ response to public anger over proliferating data centers will derail profitable investments.

Half of the new data center capacity expected over the next year is located in four states that flipped to Democratic Party control, said hedge fund manager Dan Niles of Niles Investment Management in Fort Lauderdale, Florida. Excessive regulation could push the “hyperscalers” to take their investments elsewhere, he said.

“The biggest companies in the world can’t afford to just sort of sit there. They will put plants where they make the most sense. And there are plenty of places that would take them,” he said. The industry’s financial outlook is drawing intensified scrutiny.

Alphabet, Microsoft, Amazon and other hyperscalers are in a “whatever-it-takes” posture on the investment required to compete. They plan to deploy a total of almost $1 trillion this year and perhaps more next year, seeing the global AI business as a revolutionary opportunity that can’t be missed.

These giants are likely to survive any AI winnowing. Along with operating data centers, they each have other products that bring in revenue to buttress them against costly AI setbacks, said Will Sommer, vice president of Gartner, a research firm in London.

But smaller AI providers called “neoclouds” may not be so lucky. Lacking other products, hundreds of these companies — with names like CoreWeave and Nebius — effectively represent concentrated bets on demand for additional data-crunching capability. There will not be enough revenue to go around, Sommer said.

By 2030, 80 percent of neoclouds will fail, be acquired or merge with a larger company, he said. At the current pace of investment, data center owners will need to write off $3.5 trillion in assets.