Wall Street is staring at Nvidia again after a fresh fight broke out between the chip giant and two of the most feared short sellers alive. Over the weekend, Nvidia sent a seven-page memo to analysts denying that it uses vendor financing, a method where suppliers loan money to their own buyers. The Substack post that triggered the memo compared Nvidia to Enron and Lucent. Enron broke after hiding debt off its balance sheet during the internet boom. Lucent broke after loaning money to customers that later failed. The writer said Nvidia now fits that same pattern through what was described as circular financing . Jim Chanos, the investor known for calling the collapse of Enron, said the comparison between Nvidia and Lucent makes sense. When the dot-com crash hit and the payments stopped, Lucent wrote off billions. Chanos calls out Nvidia’s customer cash loop In an interview with Yahoo Finance, Jim said Nvidia is sending money into firms that are losing cash just so those firms can keep ordering its chips.“They’re putting money into money-losing companies in order for those companies to order their chips,” Jim said. Nvidia has made direct investments into OpenAI, xAI, CoreWeave, and Nebius, all of which depend heavily on Nvidia’s hardware and are still burning cash. Nvidia denied any link to accounting abuse. In its memo, the company said it does not look like historical fraud cases and said its business is sound and its books are clean. It also said its customers pay within 53 days of buying chips, not over several years like traditional vendor loan deals. The company said it does not use vendor financing to grow revenue and said its sales system is simple. That denial did not change the tone of the short sellers. Michael Burry , the investor who predicted the 2008 housing crash, went further in a post on X last week. Michael said Nvidia is one of several AI companies showing “suspicious revenue recognition” because of how deeply it invests in its own buyers. His warning landed just days after Nvidia’s memo hit inboxes. Debt loads and data center rush raise fresh risks Jim also warned that debt inside the AI sector is turning into another stress point. He said some of Nvidia’s biggest buyers now use off-balance-sheet debt to pay for chips. He named Meta and xAI as firms using those structures. Others like Anthropic are using standard loans. Jim said the mix of heavy borrowing and money-losing business models creates a weak base under the entire market. “Putting lots of credit and really arcane financial structures on top of these money-losing entities is the real Achilles heel to the AI tech market,” Jim said during the same interview. Both Jim and Michael Burry said the bigger danger may not be billing methods at all. They said the real risk is scale. The largest tech firms on Earth are spending billions on AI data centers before users show up. In his Substack called Cassandra Unchained , Michael wrote that the market now has “catastrophically overbuilt supply and nowhere near enough demand.” He said the world is filling with too many chips, too many servers, and too many empty buildings waiting for real users. Nvidia told investors the opposite in its latest earnings report. The company said demand for its chips is “off the charts” and said it is still a full generation ahead of rivals. It also responded to growing pressure from Google, whose rising chip efforts pushed Nvidia’s stock lower before a rebound on Wednesday. Jim said the speed of the build-out itself is the core danger. If demand falls short of what buyers expect in 2027 or 2028, he said orders could start disappearing. “If it turns out that we don’t quite need all the data center or chip capacity we thought we will in ’27 or ’28, you could see orders begin to be canceled,” Jim said. “That’s a big risk that not a lot of people are talking about.” Get up to $30,050 in trading rewards when you join Bybit today
