Michael Burry, the investor famed for predicting the 2008 financial crisis, has doubled down on his bet against Nvidia, signaling a strong conviction that the AI sector is in a bubble.
His argument goes deeper than simply calling the stock overvalued. Burry's thesis is built on what he sees as fundamental weaknesses in the quality of Nvidia's earnings and the sustainability of its growth. He points to three core issues.
First are Nvidia's massive purchase obligations, which total $95.2 billion. This is an unusually large fixed commitment for a company whose products become outdated every 12 to 18 months. When this figure was disclosed in an official filing, the stock fell over 5%, suggesting investors share this concern about the company being locked into immense, potentially risky, spending.
Second, Burry highlights aggressive accounting practices by hyperscalers—Nvidia's biggest customers like Google and Microsoft. These companies have extended the official 'useful life' of their servers. This accounting change reduces their current depreciation expenses, which in turn makes their profits appear higher. This inflated profitability may create a misleading picture of their ability to sustain huge purchases of Nvidia's chips.
Third is the risk of 'circular financing.' This concern was amplified by a proposed (and now reportedly stalled) $100 billion deal where Nvidia would invest in OpenAI, which would then use the capital to buy Nvidia's chips. Such arrangements can create artificial demand, making growth appear more organic than it truly is.
Adding to these concerns are external pressures like US-China trade restrictions, which already forced Nvidia to take a $4.5 billion charge, and rising competition from rivals like AMD. Of course, the demand for AI chips is undeniably strong right now, exemplified by Google's parent company Alphabet planning to spend up to $185 billion on AI infrastructure in 2026. This creates the central tension: is this spending a sign of a genuine technological revolution, or the final, frantic phase of a bubble? Burry is betting on the latter, and he expects we'll have our answer by the end of the year.
- Put Contracts: A financial instrument giving the owner the right, but not the obligation, to sell a stock at a specified price. It's a common way to bet that a stock's price will fall.
- Hyperscaler: A term for a massive cloud services provider, such as Amazon Web Services (AWS), Google Cloud, and Microsoft Azure. They are the largest buyers of AI chips.
- Circular Financing: A practice where a company funds its own customer, who then uses the money to buy the company's products. This can create a false impression of market demand.