U.S. midterm elections could widen AI data center divide

The U.S. midterm elections could widen the AI ​​gap even further. Tighter data center regulations could give big tech companies access to production capacity, squeezing companies trying to catch up with debt financing.

Democratic victories in key gubernatorial races raise the possibility of increased scrutiny of data center projects and raise investor concerns about how quickly new capacity can be built and financed.

In Texas, Democratic Gov.-elect Gina Hinojosa is proposing a statewide moratorium on permits for new data centers, although projects already under construction would reportedly be exempt.

As the market digests Wednesday’s results. microsoft and Amazon AI infrastructure providers each soared more than 6%. coreweave It fell by about 10%. Semiconductor makers Micron and Nvidia fell 7% and 3.17%, respectively.

James Sullivan, managing director and co-head of global fundamental research at JPMorgan, told CNBC’s “Squawk Box Asia” on Thursday that differences in borrowing costs played a key role in the split.

Sullivan said neocloud companies’ debt spreads have widened significantly, and their cost of additional capital is significantly higher than hyperscalers, which benefit from more diversified customers and revenue streams.

“The fact that we’re seeing this polarization in stock performance is largely driven by the view that existing capacity is becoming much more valuable due to increased supply constraints,” he said.

If interim financial results make it easier for new projects to tolerate delays, that view may be strengthened.

Citing the company’s post-interim analysis, the Business Times reported that BMI warned that a sharp slowdown in data center construction and AI investment could cause overall investment to fall more than expected.

Potential impacts are not uniform across the construction pipeline. Much of the data center capacity expected to come online in 2027 is already under construction, so the short-term impact of the new restrictions will be limited.

Investment bank Nomura said in a note on Wednesday’s interim results that projects planned for 2028 are at greater risk of permitting delays and regulatory uncertainty.

The company said tighter regulations on data center construction could become a bigger headwind for AI investment from 2028 onwards.

Shares of Nvidia-backed neocloud CoreWeave, Nebius and Iren fell at least 30% in October as tight credit markets raised questions about how AI infrastructure projects would be financed. Coreweave’s five-year credit default swap spread has soared by more than 1,100 basis points, reflecting rising costs for investors seeking protection against potential defaults.

The difference may be how well companies can absorb delays and actually fund the ability to build. JPMorgan Asset Management said in an Oct. 20 research note that hyperscalers have broader cloud operations and more internally generated cash flow, while lower-rated neoclouds have greater exposure to borrowing costs, energy prices and refinancing terms.

The asset management company said hyperscalers with strong financial strength may also secure cheaper production capacity or acquire assets if weaker operators scale back.