Nvidia has been accused of engaging in "circular financing," where it invests in promising AI companies, which then use that capital to purchase Nvidia's graphics processing units (GPUs). Prominent short sellers Jim Chanos and Michael Burry have likened this practice to vendor financing schemes that contributed to the downfall of companies like Lucent during the dot-com era. Chanos specifically stated that Nvidia is "putting money into money-losing companies in order for those companies to order their chips," and Burry highlighted "suspicious revenue recognition" linked to equity investments in customers.
Nvidia has vehemently denied these claims, issuing a detailed seven-page memorandum to analysts. The company asserts that its investments in entities like CoreWeave, Nebius, OpenAI, and xAI are strategic minority stakes, not disguised financing arrangements. Nvidia emphasized that its average days sales outstanding remain at 53 days and that it does not offer multi-year repayment terms, distinguishing its practices from historical accounting scandals. The company maintains that its financial reporting is transparent and its business is economically sound, highlighting "off the charts" demand for its latest-generation chips.
Beyond direct vendor financing concerns, critics argue that the AI sector is experiencing significant over-investment. Burry described the current build-out as "catastrophically overbuilt supply," drawing parallels to the fiber-optic glut of the early 2000s, suggesting that billions are being spent on AI infrastructure before proven end-user demand has scaled. Chanos also pointed to the growing role of off-balance-sheet and structured debt in the AI ecosystem, noting that customers like Meta and xAI are using complex financial structures to fund massive GPU purchases, which he considers the "real Achilles heel" of the sector. Nvidia, however, counters that it has sustained order visibility and none of the impaired receivables or write-downs that plagued past corporate failures.
Despite Nvidia's defense, the ongoing debate has intensified scrutiny on the sustainability of AI capital-expenditure cycles and the potential for financial engineering to inflate reported growth in the semiconductor and cloud sectors. While Nvidia claims its practices are dissimilar to past frauds, the vocal skepticism from celebrated short sellers ensures that questions about revenue quality and potential infrastructure overbuild will remain central to the AI investment narrative. Nvidia is also pivoting to become more of a financier, with an "AI Compute Partnership" that aims to provide hardware and absorb risk for smaller cloud operators, backed by a $500 billion order backlog and strong analyst support, including a consensus "strong buy" rating and a $309 price target.
