Investment Review №357. A Change of Leadership

The most consequential U.S. election result for the AI sector came not in Congress, but in Texas. For the first time since 1990, the state elected a Democratic governor, with Gina Hinojosa defeating Greg Abbott after making opposition to data-center development a central campaign issue. Texas accounts for a quarter of all U.S. data-center capacity planned through 2030. Projects totaling approximately 13.1 GW could now face delays, posing a substantial risk.

Hinojosa has pledged to impose a moratorium on new project approvals until rules are established for water use, power supply and the allocation of grid infrastructure costs. According to an ABC News poll, 69% of Texans oppose data-center construction in their communities. Even Republican Greg Abbott had previously suspended permit approvals. The issue goes beyond partisan politics: local residents are increasingly unwilling to subsidize corporate AI projects through higher electricity bills and taxpayer support.

Billions of dollars in planned investment by some of the S&P 500's biggest companies are at stake. Alphabet (GOOGL) plans to invest $40bn in cloud and AI infrastructure in Texas through the end of 2027. Meta (META) is building a 1 GW data center in El Paso at a cost of more than $10bn, while Microsoft (MSFT) is planning a 2 GW facility in Pecos. Oracle (ORCL) and its partners are developing the Stargate project near Abilene, while Amazon (AMZN) is building an AWS facility in Wharton County. Projects already approved or under construction are unlikely to face cutbacks, but future contracts remain vulnerable. Texas is also not alone: similar challenges are emerging in other states, including Ohio and Georgia.

The risks could ripple across the entire AI infrastructure supply chain. For Oracle (ORCL), Meta (META), Alphabet (GOOGL), Microsoft (MSFT) and Amazon (AMZN), the threat is delayed capacity deployment and longer investment payback periods. For Nvidia (NVDA), Broadcom (AVGO), Micron (MU) and Arista Networks (ANET), it means potential delays in orders for chips, memory and networking equipment. Vertiv (VRT), Eaton (ETN) and Quanta Services (PWR) could face postponed equipment purchases. Utilities CenterPoint Energy (CNP) and American Electric Power (AEP), along with power producers Vistra (VST) and NRG Energy (NRG), could lose some of their anticipated demand growth. The market has yet to fully price in these risks. The scale of the threat will become clearer as the implications of the election results emerge. Investors already reacted sharply to the initial results, with Nasdaq-100 futures falling roughly 2% overnight on November 4.

Does this mark the beginning of the end for the AI race? Not at all. The market made that clear by fully recovering its losses before the opening bell.

Voters’ concerns are understandable, and the election results showed just how close the race was. For corporations, however, this does not mean widespread cancellations of construction contracts. Instead, it is likely to mean fewer subsidies and a shift toward more remote areas or other states, where infrastructure is less developed and opposition to new projects may be weaker. This would put some pressure on margins, as companies would have to shoulder part of the infrastructure costs themselves. Several existing projects already show that companies are willing to absorb those costs, supporting this view.

There are potential beneficiaries, however, including Equinix (EQIX) and Digital Realty (DLR). A shortage of existing grid-connected capacity could drive higher lease rates at operating facilities. Technology accounts for 37.4% of the S&P 500’s market capitalization, even before including Alphabet, Meta and Amazon, which are classified in other sectors. For now, 2027 earnings forecasts remain unchanged, although that could shift over the next 6–9 months. Meanwhile, technical indicators point to a partial near-term rebound in beaten-down stocks. We do not expect a broad market correction before year-end, as sustained long-term demand continues to underpin the multiyear bullish investment case for AI infrastructure.