Private equity giant Blackstone is positioning itself at the center of the AI boom, and one of its senior executives says the growing pile of debt behind that boom is not the threat some investors fear. Jas Khaira, a firm veteran who runs Blackstone’s AI push, made the case last week on stage at TechCrunch Disrupt.
Blackstone is writing checks through several strategies. It lends large sums to data center operators, invests in frontier AI models and applications, and backs companies that supply the computing power AI requires. Earlier this year the firm launched Blackstone N1, or BXN1, a San Francisco-based hub for those efforts. Khaira relocated from New York to run it.
The scale of the spending is the reason Blackstone’s balance sheet matters. Earlier technology waves, such as mobile and software-as-a-service, let companies grow to significant size on modest capital. AI does not work that way. Building the infrastructure takes staggering amounts of money, and Khaira said that is where Blackstone differs from traditional venture capitalists. The firm offers flexible capital structures, ranging from debt to equity, and focuses on financing compute.
“We’re the largest owner of compute outside of the hyperscalers today, whether it’s investments in neo clouds, investments in data centers, and being able to use that compute strategically to support companies,” Khaira said.
BXN1 played a central role in several compute deals. One was Crux AI, a neocloud launched in partnership with Google to commercialize its custom TPUs. Blackstone committed $5 billion to that company as an investor. The firm also backed Etched, a four-year-old chip startup whose rapid development has drawn investor attention. Neoclouds are cloud providers that rent out GPU infrastructure for AI work.
Blackstone pioneered a way to lend heavily to venture-backed neoclouds. Lenders look at committed customer contracts and at the GPUs the company pledges as collateral. The model dates to 2023, when CoreWeave approached Blackstone for a $2.3 billion loan. At that point, according to Khaira, the company had “30 million of revenue and minimal profitability.” CoreWeave was borrowing against its large collection of Nvidia GPUs, and Nvidia is also an investor. Blackstone grew more comfortable because CoreWeave had contracts with blue-chip customers such as Microsoft.
That deal became the template for how Blackstone and other credit firms financed billions in follow-on debt for CoreWeave and its rivals. Blackstone has since struck other neocloud deals, including co-leading a $10 billion credit facility with Coatue for Australian GPU provider Firmus earlier this year.
The customer, Khaira said, is the key variable. “One of the primary things that we’re looking to is the nature of the customer. And if the customer is an investment-grade customer — let’s say it’s Microsoft or let’s say it’s Meta — the risk gets priced relative to that customer,” he said.
Why Are Investors Worried About AI Infrastructure Debt?
Khaira did not dismiss the concerns. He acknowledged that in recent weeks, the debt piling up at these companies has raised red flags. CoreWeave’s debt has ballooned to about $50 billion. Altimeter Capital founder Brad Gerstner told CNBC that the collateral CoreWeave is posting, which is essentially only GPUs, “is not good enough anymore. They used that in the past to get cheap debt.”
Khaira said the credit market’s unease comes from several sources at once. Interest rates ticked up in September, and hotter-than-expected inflation suggests borrowing costs could keep rising. A flood of AI infrastructure debt offerings has not helped.
“We’re also seeing higher credit spreads,” Khaira said. Lenders are charging bigger premiums on top of already higher base rates to finance the buildout. “The cost of capital to price in debt for a Meta or a Microsoft has widened out,” he added, noting that even technology giants with massive balance sheets and solid earnings are paying more to borrow.
The pressure has reached the stock market. CoreWeave’s shares fell more than 20% in recent days. Firmus scrapped its planned initial public offering on the Australian Securities Exchange, which had a target valuation of roughly $30 billion.
“Ultimately, the market wasn’t there at that valuation, and so I think as a result, that deal didn’t get done,” Khaira said of the canceled listing. “The market is thinking what is the right price for neoclouds, what is the right price for these businesses in this moment when interest rates are moving up.”
Even so, demand for computing power remains intense. AI companies cannot get enough compute to run their models, and that appetite keeps driving data center construction.
Could AI Data Center Debt Trigger a Broader Credit Crisis?
Khaira also pushed back on the more dire predictions that the accumulating AI infrastructure debt could set off a broad credit crisis or crash the economy. “This doesn’t feel like something where you’ve got 2008 banks that are shaky credits,” he said. Because most data centers are built for hyperscalers, he argued, the debt financing them is ultimately backed “by the most creditworthy companies in the world.”
Blackstone is also working on the demand side. The firm formed a $1.5 billion joint venture called Ode with Anthropic and Hellman & Friedman earlier this year. Ode provides forward-deployed engineers who help companies adopt AI and fold it into their operations. Blackstone’s 280 portfolio companies are the starting point.
Is AI Delivering Real Returns for Blackstone's Companies?
Asked whether Blackstone-owned companies are seeing real returns from AI, Khaira was candid. “There is definitely ROI, but it is uneven.” He described adoption as something that could move in a “sawtooth-type pattern.” To skeptics, he acknowledged, that pattern can look like an oversupply of compute before demand catches up.
Dieser Artikel wurde mit Unterstützung von KI verfasst.
News Factory APP - agentische News für besseres SEO & AEO.