Michael Burry, renowned for his bets against the housing market prior to the 2008 crash, has turned his focus to Nvidia. He highlights a major risk linked to the company's customer concentration, which has escalated alarmingly. Currently, three customers represent a staggering 64% of Nvidia's receivables, a figure that had only been 56% a quarter earlier. Burry characterized this concentration as extremely high and has backed his assertion with substantial investments in put options against Nvidia.
#What Are the Risks Associated with Nvidia's Customer Concentration?
This concentration of revenue is primarily attributed to hyperscale cloud service providers such as Microsoft, Google, Amazon, and Meta. In recent reports, these clients have driven more than 50% of Nvidia's data center revenue during specific periods. Nvidia’s recent SEC filing revealed that two undisclosed customers comprised 39% of the company's total revenue, making the situation even more precarious.
An interesting trend has emerged with Microsoft. While Microsoft’s share of Nvidia’s receivables has increased, its revenue contribution to Nvidia has decreased for the first time in over three years. This disconnect implies that although Microsoft owes Nvidia a larger proportion of revenue, it is purchasing fewer products.
#Why is Burry Drawing Parallels with Cisco Systems?
Burry's concerns echo the fate of Cisco Systems during the dot-com era. Initially, Cisco thrived by supplying essential infrastructure during the internet boom. However, as the tech bubble burst, Cisco experienced a significant downturn when their major revenue sources diminished, resulting in a peak-to-trough decline of about 80%. Cisco’s stock has never regained its former heights.
Currently, Nvidia stands as the dominant player in the AI boom, with its GPUs integral for training and inference processes that hyperscale cloud providers need. Burry's thesis revolves around determining whether the ongoing surge in AI spending is a lasting trend or merely a short-lived cycle. The crux of his argument lies in whether the phase of training these AI systems is temporary instead of a new norm.
#How Could Hyperscalers Impact Future Demand for Nvidia Products?
Burry previously flagged concerning accounting practices among hyperscale cloud providers, emphasizing that these firms potentially underreport depreciation on Nvidia GPUs. They extend GPU lifespans to 5-6 years, thereby reducing their need to purchase new equipment. This method may suggest that while current expenditures appear sustainable, they could lead to diminished demand in the future.
#What Should Investors Monitor Moving Forward?
The competitive landscape is shifting seemingly overnight. While Nvidia currently holds a strong position in the AI accelerator market, rivals such as AMD, Intel, and various in-house solutions from major players like Google, Amazon, and Microsoft threaten its dominance. Burry warns of a potential sharp decline in Nvidia's stock if spending from hyperscalers begins to slow down. Investors should keep a close eye on the percentage of Nvidia's revenue that is concentrated among a few clients, specifically whether that 64% mark continues to grow or begins to decline. Understanding these dynamics could be pivotal for decision-making in the investment landscape.