Jim Cramer AI bubble warning centers on Nvidia and OpenAI financing

Jim Cramer’s AI bubble warning put a fresh spotlight on Nvidia Monday after reports that the chipmaker is in talks to help back OpenAI’s next wave of data center spending. For everyday investors, the issue is less about one deal and more about whether the AI boom is becoming dependent on suppliers financing the customers that buy from them.

Cramer, the host of CNBC’s “Mad Money,” said Monday that the latest financing discussions in artificial intelligence reminded him of the late-1990s tech buildout that preceded the dot-com collapse. He said he lived through 2000 and did not want to see a repeat.

The Wall Street Journal reported Sunday that Nvidia was discussing a $250 billion backstop for OpenAI tied to a planned 10-gigawatt AI data center campus in Ohio. CNBC confirmed the talks Monday. Nvidia declined to comment, according to CNBC.

The proposed guarantee would help support lease and construction debt for the project, CNBC reported. It would not directly finance the Nvidia chips used inside the facility. Nvidia shares fell more than 4% Monday, and the move weighed on a range of semiconductor stocks, according to CNBC.

What is circular AI financing?

Circular financing is when money moves between companies that also depend on each other commercially. In this case, the concern is that a supplier such as Nvidia could help finance an AI company’s infrastructure expansion while that same AI company is a major buyer of Nvidia chips.

That structure can make revenue growth look powerful while capital is easy to find. The risk, as Cramer described it, is that the loop can break if customers cannot keep paying for the equipment or facilities that underpin their expansion.

Cramer compared the setup with telecom equipment financing before the dot-com crash. He said suppliers in that period helped customers fund large purchases, which lifted sales for a time, but losses followed when buyers with weak cash positions could not meet their obligations.

“What we learned in 2000 is that you don’t lend to companies who buy your goods,” Cramer said on CNBC.

Nvidia has already put money into several AI companies that buy or use its chips. CNBC reported that Nvidia invested $30 billion in OpenAI in March and $10 billion in Anthropic last year. The company has also backed neocloud providers, which rent out computing power built on Nvidia hardware. Nvidia has said those investments help expand the AI ecosystem and can produce long-term returns, according to CNBC.

Cramer said he still considers Nvidia a strong company and was not calling for another dot-com crash. His warning centered on market confidence: investors can become uneasy when a supplier’s growth depends on customers whose spending requires constant access to outside capital.

OpenAI confidentially filed for an initial public offering in June, CNBC reported, though it has not announced when it might list shares. Private investors valued the company at more than $800 billion in March, according to CNBC, as it builds more computing capacity for AI models and competes with Alphabet and Meta.

Cramer said Nvidia could be in a strong position if OpenAI can afford the chips and infrastructure it needs, including through a potential public offering. If OpenAI or other AI infrastructure buyers cannot pay, he said, the situation would look different.

He also said the risk reaches beyond Nvidia because many companies now expect data center demand to support earnings. If capital markets stop funding new data centers and the companies involved cannot pay from their own cash, Cramer said, investors could face a version of the pressure seen in 2000.

This story draws on original reporting from CNBC.