The artificial intelligence infrastructure buildout will need more equity financing as lenders become increasingly exposed to the companies driving the buildout, Blackstone President Jon Gray said Tuesday.
Speaking at the Hong Kong Monetary Authority’s Global Financial Leaders’ Investment Summit, Gray said the scale of capital required for AI is creating concentration risks for debt investors, both across the sector and at individual companies.
Unlike shareholders, who can benefit from a company's growth, lenders have limited upside as their returns are largely confined to interest and principal repayments, he said.
Speaking at the same panel, Pimco CEO Emmanuel Roman warned investors that the scale of AI financing could prove to be "too much" for markets to absorb.
Roman said AI-related capital spending could reach around $10 trillion over the next five years, citing General Atlantic CEO Bill Ford's estimate. He added that about $550 billion in AI debt had been issued in various forms so far this year.
Even with a mix of equity and debt financing, the amount required would be “gigantic,” Roman said, comparing it with a global corporate bond market of roughly $20 trillion.
“At some point in time, it’s going to be too much,” he said.
Global data center demand will require about $8.5 trillion in capital spending over the next five years, Goldman Sachs wrote in a report Monday.
Gray said financing the buildout would require a combination of public markets, banks and private capital, with companies increasingly turning to structured transactions to limit the amount of debt they take on. Cash generated by hyperscalers would also fund part of the spending, he said.
“The creativity is needed because the scale of capital is enormous," Gray said.
Despite those financing challenges, Gray remained bullish on AI infrastructure, saying demand for computing power was “running well ahead of supply.”
Spending on AI models across Blackstone’s portfolio companies had risen 20-fold over the past year as businesses saw financial benefits from uses including coding, customer engagement and content creation, he said.
Meeting that demand remains difficult, however, as securing sites, permits, chips, power and capital becomes more challenging, according to Gray.
