- Yardeni Research argues Thursday's sharp selloff was overdone, viewing it as a buying opportunity.
- The firm contends that political resistance, financing constraints, and infrastructure bottlenecks will slow—not end—the AI investment boom.
- A more disciplined construction cycle could reduce overinvestment and make the expansion more durable, according to Yardeni.
A Contrarian Call
Yardeni Research is urging investors to look past Thursday's steep market decline, arguing that the selloff was an overreaction to growing political opposition to AI infrastructure. The independent research firm, led by economist Ed Yardeni, believes the pullback presents a compelling entry point for long-term investors.
The firm's central thesis is that obstacles such as permitting delays, higher costs, and tighter financing will act as "speed bumps" that prevent the AI boom from overheating. In a research note, Yardeni argued that a more disciplined construction cycle could actually extend the longevity of the AI investment wave by reducing the risk of overinvestment.
Market Carnage
The numbers from Thursday's session were brutal. The S&P 500 tumbled 4.36% to 7,353.73, while the Nasdaq Composite plunged 5.52% to 24,315.00, closing more than 10% below its October 6 record and confirming a correction. The Dow Jones Industrial Average lost 3.44%. Semiconductor stocks were hit particularly hard, with the sector shedding more than 12%.
The catalyst was a reassessment of the midterm elections. After an initial relief rally on Wednesday, investors zeroed in on whether Democratic victories in Congress and key data-center states would lead to tougher permitting, higher costs, and slower construction. The prospect of regulatory headwinds for AI infrastructure sent shockwaves through the market.
The Texas Factor
One significant late-breaking development softened the blow. According to Yardeni's November 5 research note, Texas Governor-elect Gina Hinojosa clarified after the market close that already-permitted data centers could finish construction. Yardeni interprets this as protecting projects covering 2027 and likely much of 2028, though the firm cautions that this is its assessment, not a blanket regulatory guarantee.
Texas is a crucial battleground, accounting for nearly 30% of planned U.S. data-center additions in 2028, according to industry estimates cited by Yardeni. Hinojosa campaigned on pausing new approvals pending safeguards, but her post-close clarification distinguished existing permits from future projects. Her proposed legislative agenda includes requiring newly permitted facilities to pay taxes, protect consumers from higher electricity prices, preserve water supplies, use union labor, and provide greater community participation.
Quantifying the Risk
The often-cited "45%" figure describes exposure, not cancellations. Yardeni attributes to Goldman Sachs (NYSE:GS) an estimate that states facing possible restrictions account for roughly 45% of planned U.S. data-center capacity. This does not mean that 45% has been prohibited or will disappear, but rather that a significant portion of future capacity is at risk of delay or modification.
Yardeni's bullish interpretation stands in contrast to more cautious voices. Citi (NYSE:C) argues that data centers may still be built, but their locations, construction methods, and timing could change—a less favorable outcome for semiconductor and infrastructure suppliers. Capital Economics' Joe Maher said the political shock reinforces his concern that the AI equity boom is in its "final innings."
Winners and Losers
The distinction between companies paying for AI infrastructure and those selling it is key. Large cloud platforms like Amazon (NASDAQ:AMZN), Microsoft (NASDAQ:MSFT), Alphabet (NASDAQ:GOOGL), and Meta (NASDAQ:META) could benefit from lower capital expenditures and improved near-term free cash flow, though insufficient capacity could constrain future revenue. Chipmakers and equipment vendors, on the other hand, face delayed orders and weaker growth expectations. Nvidia (NASDAQ:NVDA) fell more than 9% on Thursday, while AMD (NASDAQ:AMD) declined roughly 17%. Smaller AI-cloud operators like CoreWeave (NASDAQ:CRWV), IREN (NASDAQ:IREN), and Applied Digital (NASDAQ:APLD) each lost more than 10%.
Power and infrastructure suppliers also suffered, with GE Vernova (NYSE:GEV) and Constellation Energy (NASDAQ:CEG) falling 7.5% and 2.5%, respectively. The broader economic concern is that AI-related investment has become an important source of activity, and project delays could ripple through construction, equipment purchases, and local development.
Political and Public Backdrop
The political context is complex. Democratic control of Congress raises expectations of greater scrutiny, but it does not itself impose a nationwide construction moratorium. State and local decisions are likely to exert substantial influence over where projects proceed. Yardeni cites an NBC (NASDAQ:CMCSA) exit poll finding that 76% of voters opposed new facilities in their communities—a reported polling result, not evidence that 76% oppose AI generally.
Residents and ratepayers are concerned about electricity bills, grid reliability, and water availability. Workers and local governments face a tradeoff between investment and construction opportunities on one side, and infrastructure burdens and tax concessions on the other. The Texas proposals seek to change those terms rather than simply reject all development.
Global Ripple Effects
The selloff had immediate international consequences. South Korea's Kospi dropped almost 9% at the open on Friday, triggering a trading halt, while European chip-related shares also suffered. U.S. construction expectations matter to suppliers and investors well beyond America's borders.
Bond markets reflected growth concerns, with the 10-year Treasury yield falling to approximately 4.84% on Thursday. Nevertheless, Federal Reserve officials were still discussing possible further rate increases, leaving investors exposed to both slower AI spending and restrictive financing conditions.
The Road Ahead
Yardeni maintains a year-end S&P 500 target of 7,900 and expects computing demand to keep growing despite slower construction. The firm's argument predates Thursday's plunge; in late October, Yardeni described three constraints on the AI boom—financial, physical, and political—and argued that they could extend its life by preventing excessive investment.
The key analytical question is not simply whether AI survives tighter regulation, but which companies can turn continuing AI demand into profitable growth under a slower, more expensive, and more politically negotiated infrastructure buildout. For now, Yardeni is betting that the selloff has created an opportunity for those willing to look past the near-term noise.
Correction: An earlier version of this article misstated the percentage decline in semiconductor stocks. It has been corrected to reflect a decline of more than 12%.