CNBC’s Jim Cramer said Monday that the latest chapter in the artificial intelligence boom is bringing back memories of the excesses that fueled the dot-com bubble.
“I lived until 2000,” the “Mad Money” host said. “I don’t need a sequel.”
The Wall Street Journal reported on Sunday. Nvidia was discussing a $250 billion backstop for OpenAI to help finance a planned 10 gigawatt artificial intelligence data center campus in Ohio. CNBC confirmed the report on Monday, but NVIDIA declined to comment. The proposed guarantee would support the project’s lease and construction debt, not the Nvidia chips deployed within the facility. Nvidia stock fell more than 4% on Monday, and many semiconductor stocks also fell.
This discussion is the latest example of how AI lending is becoming increasingly circular. Nvidia has invested in several companies that are also major customers for its chips, including a $30 billion investment in OpenAI in March and a $10 billion investment in Anthropic last year. The chipmaker also backs several neocloud providers that rent Nvidia-powered computing power to customers. Nvidia says these investments provide attractive long-term returns while supporting the growth of the AI ecosystem.
Kramer said the circular nature of the arrangement is reminiscent of the late 1990s, when telecom equipment makers helped customers finance large purchases to fuel growth. Although these deals initially boosted sales, many collapsed as cash-strapped buyers were unable to pay, he recalled, resulting in heavy losses for both suppliers and investors.
“What we learned in 2000 is don’t lend to companies that buy your products,” Kramer said.
Cramer emphasized that he still views Nvidia as a very strong company and does not predict a repeat of the dot-com crash. Rather, he said, history has shown that investors can quickly lose confidence when suppliers become too reliant on customers who rely on large spending for continued access to capital.
“If the buyer, in this case OpenAI, can actually afford to pay for these chips, maybe because the chips are made public…then Nvidia is in great shape,” Cramer said. “If the buyer can’t pay, it’s a different story.”
OpenAI secretly filed for an initial public offering in June, but has not announced a timeline for its debut. The company, valued at more than $800 billion by private investors in March, is rushing to expand the computing infrastructure needed to power its artificial intelligence models, while competing with rivals such as: alphabet and meta.
Kramer said the risks extend far beyond Nvidia as more companies rely on continued investment in AI infrastructure.
“There are so many companies that rely on their data centers for revenue,” he said. “If the market decides it doesn’t want to fund data centers anymore, and the companies themselves don’t have the money or the salaries aren’t paid, we’ll be back in 2000.”
Cramer said NVIDIA has the financial resources to support a project of this size, but argued that a strong balance sheet is not always enough to protect the company from the effects of overextension by its customers.
“Even with all the money in the world, NVIDIA shouldn’t be making these guarantees. It’s just history, that’s all, just history,” he said.
