Blackstone executive says AI data center debt fears are overblown

In short: Blackstone’s AI leader Jas Khaira says worries that AI infrastructure borrowing will trigger a major debt crisis are overstated.

What happened

Jas Khaira, who leads Blackstone’s AI efforts, said the private equity firm is not panicking about rising debt tied to the buildout of AI infrastructure. He spoke at TechCrunch Disrupt, where he described Blackstone’s approach as funding the “compute” that AI needs.

“Compute” is the raw computer power used to run AI systems, often coming from data centers packed with specialized chips called GPUs (think of GPUs as high-powered engines for crunching lots of calculations quickly). Khaira said Blackstone has been making big bets across this space, including loans to data center operators and investments in AI companies.

The article points to CoreWeave as an early example. Khaira said CoreWeave came to Blackstone in 2023 for a $2.3 billion loan, using its Nvidia GPUs and customer contracts as support for the borrowing, similar to taking a loan with a valuable asset as backing. Since then, CoreWeave’s debt has grown sharply, and market concerns have increased as interest rates and lender premiums rise.

Khaira argued that much of the debt behind data centers is effectively supported by very large, financially strong customers, often called “hyperscalers” (companies like Microsoft and Meta that run huge cloud and internet services). He said this looks different from the kind of risky lending that helped fuel the 2008 financial crisis.

Why it matters

AI tools depend on expensive computing infrastructure, and someone has to pay for it upfront. If financing dries up, AI services could get more expensive or roll out more slowly. If financing continues but becomes costlier, expect more pressure on companies that rent out AI computing power and on the prices they charge.

Source: TechCrunch AI