Nomura sees greater data-center permitting risk for 2028 projects than 2027 capacity

Much of the U.S. data-center capacity expected in 2027 is already under construction, limiting its exposure to potential post-election restrictions, Nomura said. Projects planned for 2028 face greater permitting and regulatory risk. In a November 4 note, the bank said tighter construction rules could become a more significant drag on AI investment from 2028.

Texas puts new permits in question

Democratic gubernatorial victories in the November 3 midterms raised the prospect of tougher scrutiny of data-center development. Texas governor-elect Gina Hinojosa has proposed a statewide moratorium on new data-center permits. It remains a proposal, not an enacted ban; projects already under construction would reportedly be exempt.

Borrowing costs sharpen the divide

The financing divide is already visible, according to James Sullivan, JPMorgan’s managing director and co-head of global fundamental research. Speaking to CNBC on November 5, he said borrowing spreads have widened significantly for neoclouds—specialist AI computing providers. Their cost of raising additional capital is substantially higher than that of the largest cloud companies, which have more diversified customers and revenue.

Sullivan linked the split in stock performance largely to investors valuing existing computing capacity more highly as potential supply constraints increase. On November 4, Microsoft and Amazon each gained more than 6%, while CoreWeave fell around 10%.

The financing pressure predates the election. CoreWeave, Nebius and Iren each lost at least 30% in October as tightening credit markets raised questions about funding AI infrastructure, CNBC reported.

An October 20 JPMorgan Asset Management note explains why companies may weather delays differently. Large cloud operators have broader businesses and more internally generated cash. Lower-rated neoclouds are more exposed to borrowing costs, energy prices and refinancing conditions. The asset manager said stronger operators could secure cheaper capacity or acquire assets if weaker competitors pull back.

A brake on spending has competing consequences

Some economists see a possible benefit in cooling construction. Bank of America economist Aditya Bhave said in a November 4 note that AI-related capital spending had grown at a 26% annualized rate, adjusted for inflation, since the beginning of 2025. He described that front-loaded spending wave as a symptom of overinvestment and a risk for a sharper downturn later.

Bhave argued that slowing the pace might benefit markets and the economy over the medium term, while stressing that the outcome remained uncertain. RSM US chief economist Joseph Brusuelas likewise argued that slower AI investment could make the boom more sustainable and redirect capital into other parts of the economy.