Blackstone’s Jas Khaira Addresses AI Debt Concerns Amid Market Jitters
Blackstone is positioning itself at the forefront of the AI boom, as firm veteran Jas Khaira highlighted during a recent appearance at TechCrunch Disrupt. The private equity giant is engaging in various strategies, including substantial loans to data centers and investments in advanced AI models and applications.
Earlier this year, Blackstone launched Blackstone N1 (BXN1), a central hub for its AI initiatives based in San Francisco, led by Khaira, who moved from New York to be closer to the AI sector.
Unlike previous tech cycles, such as mobile or SaaS, which required moderate capital for significant scaling, the AI transition demands enormous investments. This capital requirement positions Blackstone, with $1.3 trillion in assets under management, as a pivotal player in this technological shift.
Khaira noted that Blackstone differentiates itself from traditional venture capitalists by focusing on compute financing and offering flexible capital structures, including both debt and equity. The firm has been actively investing in this space.
“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 stated.
BXN1 has been instrumental in several compute-related deals, including a partnership with Google for Crux AI, a neocloud that commercializes custom TPUs, to which Blackstone committed $5 billion. The firm also invested in Etched, a chip startup gaining investor attention. Neoclouds are cloud providers that lease AI GPU infrastructure.
Blackstone has developed a financing model that enables it to lend significant amounts to venture-backed neoclouds by assessing customer contracts and using GPUs as collateral. The first such deal occurred in 2023 with CoreWeave, which sought a $2.3 billion loan despite having only $30 million in revenue at the time.
CoreWeave leveraged its substantial Nvidia GPU assets for the loan, with Nvidia also being an investor. Blackstone became more comfortable with the loan due to CoreWeave’s contracts with reputable clients like Microsoft.
This transaction established a financing model that has allowed Blackstone and other credit institutions to provide billions in subsequent debt to CoreWeave and competing neoclouds.
Khaira explained that the nature of the customer plays a critical role in risk assessment. “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.
Building on its success with CoreWeave, Blackstone has secured additional neocloud financing deals, including co-leading a $10 billion credit facility for Australian GPU provider Firmus earlier this year.
Khaira acknowledged that recent weeks have seen rising concerns regarding the substantial debt burdens of neocloud operators. CoreWeave’s debt has reportedly surged to about $50 billion, prompting Altimeter Capital founder Brad Gerstner to express concerns about the adequacy of collateral.
Khaira attributed market jitters to several factors, including rising interest rates and inflation, which signal increasing borrowing costs. The influx of AI infrastructure debt offerings is also contributing to these concerns.
“We’re also seeing higher credit spreads,” Khaira explained, noting that lenders are demanding higher premiums on top of elevated base interest rates for financing AI projects. He pointed out that even established tech giants are facing increased borrowing costs.
These worries have impacted the stock market, with CoreWeave’s shares dropping over 20% recently, while Firmus postponed its planned IPO due to unfavorable market conditions.
“Ultimately, the market wasn’t there at that valuation, and so I think as a result, that deal didn’t get done,” Khaira remarked regarding Firmus’ canceled listing.
Despite these pressures, Khaira emphasized that AI companies continue to require substantial compute resources to drive their models, which fuels ongoing data center development.
He dismissed fears that the accumulation of AI infrastructure debt could trigger a widespread credit crisis, stating, “This doesn’t feel like something where you’ve got 2008 banks that are shaky credits.”
Khaira noted that most data centers are built for hyperscalers, and the debt financing them is backed by highly creditworthy companies.
In addition to funding data centers, Blackstone is facilitating AI adoption through a $1.5 billion joint venture called “Ode with Anthropic,” established with Anthropic and Hellman & Friedman. Ode provides engineers to assist companies in integrating AI into their operations, starting with Blackstone’s 280 portfolio companies.
When asked about tangible gains from AI within Blackstone-owned companies, Khaira acknowledged that while there is a return on investment, it is uneven. He described technology adoption as a process that may exhibit a “sawtooth-type pattern,” which can appear as an oversupply of compute before demand catches up.
Source: TechCrunch AI