Nvidia Corp (NASDAQ:NVDA, XETRA:NVD) is trying to kill a dangerous meme before it really takes off: the idea that its record-breaking AI chip boom is being juiced by creative financing.
Over the weekend, the company sent a seven-page memo to analysts insisting it is not using “vendor financing” (the classic trick where a supplier effectively funds its own customers so they can buy more kit).
That accusation came from a niche Substack post, but it hit a nerve because it echoes the playbook of Enron and early-2000s Lucent.
Nvidia’s defence is straightforward: yes, it invests in some customers (think OpenAI, xAI, or specialist cloud players) but it says those stakes are small compared with overall sales, and crucially, customers pay for chips in about 53 days, not over years like in traditional vendor-finance schemes.
Jim Chanos, the short seller who called Enron, is unconvinced.
He argues Nvidia is backing loss-making buyers who then order more GPUs, and that rising use of debt and off–balance sheet structures around AI data centres looks uncomfortably familiar.
Michael Burry, of “Big Short” fame, has gone broader, calling out “suspicious revenue recognition” and warning of an AI build-out that looks wildly ahead of real-world demand.
Strip away the accounting jargon and the real question is simple: are hyperscalers and start-ups over-ordering chips for AI products that do not yet exist at scale?
Nvidia says demand is “off the charts” and that it is a generation ahead of rivals. The bears say this is what every bubble sounds like on the way up.