The midterm election tears open the "abyss" of AI trading, facing a moment of reckoning for trillion-dollar capital expenditure narratives

According to Zhitong Finance APP, for months, investors have been concerned about a string of risks behind the "cash cow" of artificial intelligence (AI) capital expenditure. This week, the market finally hit the panic sell-off button.

The U.S. midterm elections concluded on Tuesday, with the Democrats securing majority in both the House and Senate, as well as achieving major victories in several state legislatures and governor seats. The market immediately began pricing in the worst-case scenario for AI trades. Many winning candidates campaigned on limiting AI development, raising concerns that data center construction in many regions may come to a standstill.

This shift in expectations prompted traders to rush for the exits starting Thursday, leading to sell-offs in all assets tied to the hundreds of billions expected in AI capital expenditures. The Philadelphia Semiconductor Index plunged 12%, while the tech-heavy Nasdaq 100 Index dropped 6.4%. Both indices marked their largest single-day declines since the shock of the COVID-19 pandemic in March 2020. Nvidia, the benchmark in AI, led the losses with a drop of over 9%, erasing more than $440 billion in market value.

Although subsequent reports stated that Gina Hinojosa, the governor-elect of Texas, has made it clear that all currently approved data centers in Texas will be allowed to complete construction. Hinojosa plans to legislate that future newly approved data centers, scheduled for delivery between 2028 and 2029, must pay taxes, commit to not raising residential electricity prices, not affecting residential water usage, use union labor for construction, and ensure "Texans have a seat at the table" in decision-making.

This news prompted a pre-market rebound in Nasdaq futures, but skepticism in the market does not appear to have dissipated as a result.

"The market is questioning growth assumptions for data center supply chain companies, and capital is pulling out en masse," said Peter Tuz, President of Chase Investment Counsel, whose fund holds shares in chipmakers Nvidia (NVDA.US) and AMD (AMD.US), as well as major AI spenders like Alphabet Inc. (GOOGL.US) and Amazon (AMZN.US). "From the election results, it's easy to infer increased regulatory uncertainty, which in turn drags down these companies' growth prospects."

Since the beginning of this year, multiple pressures have continued to build, stoking market anxiety over AI capital spending: rising interest rates, growing debt levels for data center capacity expansion, and increasingly vocal public opposition to data center construction. AI capital expenditure has been the core engine driving the latest S&P 500 rally, and for a time served as a buffer against macro risks such as high oil prices, persistent inflation, and tensions in Iran.

Of course, considering the typically slow pace of government action, how efforts to curb data center development and regulate AI will evolve remains to be seen. But this wave of sell-offs has clearly shown that the bull market's core pillar has already been deeply gripped by panic.

"Right now the economy seems to be running on a single engine, and this engine is not only slowing down, but the election results may slow it further," said Barry Knapp, Managing Partner of Ironsides Macroeconomics. "I wouldn't buy anything right now."

According to estimates by strategists at UBS Group, nearly one-third of the computing power scheduled to go online in the U.S. by 2028 is now at risk. Most of this risk comes from data centers planned for construction in Texas—where, in the gubernatorial race, Hinojosa, who pledged to halt new data center approvals, defeated the incumbent Republican Governor Greg Abbott, becoming the state's first Democratic governor since 1990.

These concerns are prompting a reassessment of already pressured valuations in the AI sector. Nvidia, for example, has been the biggest beneficiary of AI investment. Its current price-to-earnings ratio is 13 times its expected profit for the next 12 months, down from over 32 times a year ago. Broadcom (AVGO.US), the second largest U.S. chip company by market value, has a projected price-to-earnings ratio of 15 times, down from a peak of 42 times in December last year.

"Previously the market took it for granted that data center expansion was a long-term trend that could last for years," said Knapp. "If capacity building can only be sustained for another few years, and the chip industry’s long-term growth slows, then what kind of valuation should we assign?"

This latest sell-off has been particularly brutal in the riskiest corners of AI trading. AI computing services provider CoreWeave (CRWV.US) plunged 20% this week, more than 75% down from its June 2025 peak and just a few dollars above its March 2025 IPO price. Debt-laden new cloud service provider Nebius (NBIS.US) fell 15% on Thursday, dropping 58% from its June high.

To make matters worse, the uncertainty surrounding AI regulation and data center construction oversight has made investors reluctant to bottom-fish at this stage.

"It's hard to judge whether the current decline has gone too far, because the regulatory outlook remains unresolved," said Luz of Chase Investment Counsel. "It looks oversold, but we'll wait until the situation becomes clearer before considering buying."

A benefit of the slowed pace of data center construction is that tech giants like Amazon, Alphabet, Microsoft (MSFT.US), and Meta (META.US) will be forced to cut spending. Judging by the share price reactions following recent earnings reports, investors are no longer willing to tolerate capital expenditure consistently outpacing operating cash flow.

Just this Wednesday, the combined market value of these four giants, plus Apple (AAPL.US) after its stock surged, increased by more than $800 billion; but on Thursday, the broad market fell, erasing a portion of earlier gains.

"If they cut spending, maybe they won't need to borrow as much, and maybe cash flow will improve," said Luz. "But at the same time, growth may slow. There are just too many questions now."