US AI investment slowdown may ease financing pressure, benefit hyperscalers: PIMCO
New Delhi, November 17
A potential slowdown in artificial intelligence infrastructure development in the United States following the midterm elections could ease near-term financing pressures and benefit large technology companies, while creating challenges for highly leveraged infrastructure providers, according to a report by PIMCO.
The report highlighted that increased regulatory restrictions, higher costs and longer development timelines could slow the pace of AI infrastructure investments without significantly reducing long-term demand for computing capacity.
According to data compiled from Cleanview and cited by PIMCO, approximately 45 per cent of planned US data centre capacity is located in states where tighter restrictions could emerge, with Texas alone accounting for nearly one-quarter.
However, the report clarified that exposure to potential restrictions does not necessarily mean projects would be cancelled, as companies could delay, relocate or redesign developments.
"For now, however, we expect a longer, and less linear, AI buildout rather than a materially smaller one," the report stated.
PIMCO noted that diversified hyperscalers, supported by strong balance sheets and multiple revenue streams, would be better positioned to manage delays in infrastructure development.
"Slower deployment extends the scarcity of existing compute, supporting utilisation and pricing while reducing the risk of eventual overcapacity," the report said.
The report further highlighted that slower infrastructure deployment could reduce the immediate supply of new debt issuance, potentially supporting credit market conditions.
However, companies with higher leverage could face challenges as existing financial commitments, including equipment purchases, leases and interest expenses, continue despite delays in revenue generation.
"For leveraged business models, the risk is that the scarcity premium may be more than offset by a valuation discount for recontracting risk and a longer, less certain monetisation timeline," it noted.
The report also highlighted the substantial scale of anticipated AI investments.
Before the elections, consensus estimates indicated that capital expenditure by major hyperscalers, including Amazon, Alphabet, Meta, Microsoft and Oracle, could reach approximately USD 1.2 trillion annually in both 2027 and 2028.
These investment projections could now be revised downward or pushed further into the future.
"The U.S. election outcome may slow the AI buildout, but it's unlikely to stop it," PIMCO stated.
The report emphasised that investors would increasingly need to assess companies' financial strength, their ability to withstand project delays and the sustainability of returns from existing computing infrastructure.
— ANI


Leave a comment