Neoclouds build data centres around Nvidia processors and rent computing capacity to customers.
Nvidia has disclosed a commitment to buy up to $36 billion in cloud services from such providers.
The report examines the financial exposure: if Nvidia could not meet its commitments, debt-funded infrastructure firms and their lenders could be affected.
It does not report that this has happened.
The company has pledged to invest hundreds of billions of dollars in other ventures. CNBC has described it as the “central bank of artificial intelligence,” as its financial relationships connect it to infrastructure companies, lenders and customers. The Fed and other financial institutions are watching closely to see if a sudden shift in expectations for AI could cause broader disruption.
How Nvidia connects to AI infrastructure
A crucial link are the so-called neoclouds, smaller companies that build data centers with Nvidia processors and rent their computing power to customers. Nvidia supports them financially, while their services create demand for its products. In July, it announced that it had committed to buying up to $36 billion worth of cloud computing services from such providers.
So provider growth and demand for Nvidia's equipment reinforce each other. If the company were unable to meet its commitments, providers that have borrowed to build their infrastructure could face difficulties, with potential consequences for the banks that have lent to them.
The commitments aren't limited to cloud services purchases. Nvidia has guaranteed up to $105 billion for a large data center project in Ohio, tied to OpenAI and SB Energy. It has also committed to covering up to $125 billion of a $500 billion credit program.
↳
READ ALSO
HACKS.GR / RELATED
AI Unlocked
SpaceX plans two supercomputer complexes worth more than $74 billion in Texas and Arizona
↗
Signs of stress have been recorded in the financial data of companies in the sector. From October 1 to November 5, the cost of insurance on CoreWeave's debt increased by 94%. At the same time, the price of an IREN convertible bond fell by 23% in two days. These changes are not in themselves an indication that a company has defaulted on its obligations.
The magnitude of the risk
Stein van Nuyenburg, a professor at Columbia Business School, estimates that Nvidia’s total liabilities are close to $400 billion. The amount is small compared to the company’s market value, which remains close to $4 trillion. In his estimation, Nvidia would have to come very close to extinction before all of its guarantees would be at risk.
Analysts and former regulators say a 2007-2008-style crisis is unlikely. Sheila Burr, former head of the U.S. Federal Deposit Insurance Corporation, told CNBC that Nvidia is not a bank that can collapse from a sudden loss of confidence. She based her assessment on strong demand for its products and the company's significant equity, which acts as a financial buffer.
Expectations for Nvidia’s performance remain high: analysts predict that its revenue will more than double in the current fiscal year and grow at about the same rate over the next two years. However, the International Monetary Fund warns of the growing debt in the industry. Fund official Jason Wu points out that markets and the real economy are increasingly dependent on the same driving force. If the value of processors declines, investors may question whether the companies that borrowed to buy them will be able to repay their debts.
As early as July, Fed officials were considering the possibility of a broader decline in asset prices if the industry were to suffer. Over time, the development of their own AI chips by AMD, Broadcom, Intel, Google, Microsoft and OpenAI could reduce the market’s reliance on Nvidia. For now, however, central banks are concerned that a sharp correction in the AI market could affect businesses and lenders beyond the tech sector.
>Stay connected with Hacks.gr
Set us as a preferred source on Google and find us on social media.
Preferred source
↗
end of article
Top comments (0)