AI infrastructure needs spark financing concerns, executives say

Pimco CEO Emmanuel Roman cautioned investors that the sheer volume of capital required for AI development could overwhelm financial markets. Roman cited an estimate from General Atlantic CEO Bill Ford, suggesting AI-related capital expenditures could total approximately $10 trillion over the next five years. He noted that roughly $550 billion in AI debt has already been issued this year.

Roman expressed that even a blended approach of equity and debt would necessitate a “gigantic” amount of funding, dwarfing the roughly $20 trillion global corporate bond market. He warned that the market’s capacity could be exceeded at some point.

Goldman Sachs projected that global data center demand will necessitate around $8.5 trillion in capital investment over the next five years.

Gray suggested a combination of public markets, banks, and private capital will be essential to finance this buildout, with companies increasingly utilizing structured transactions to manage debt levels. He also highlighted that revenue from major cloud providers, or hyperscalers, will contribute to funding these investments. “The creativity is needed because the scale of capital is enormous,” Gray stated.

Despite these financing hurdles, Gray remains optimistic about the future of AI infrastructure, citing that demand for computing power is significantly outpacing supply. He reported that AI model spending across Blackstone’s portfolio companies has surged twentyfold in the last year, driven by financial gains in areas like coding, customer interaction, and content creation.

Securing necessary resources – including locations, permits, semiconductors, power, and capital – presents ongoing challenges, Gray added.