Michael Burry Says Nvidia's Real Problem Is a $176 Billion Accounting Gap

Michael Burry, the investor famous for predicting the 2008 housing collapse, estimates that AI-related understatement of depreciation across the technology industry could reach $176 billion between 2026 and 2028. His warning lands on the same day Nvidia (NASDAQ:NVDA) announced that CEO Jensen Huang resigned this morning amid revelations about his personal relationships, and it frames a problem Burry argues will threaten the chipmaker long after the leadership controversy fades.

The context matters. The artificial intelligence boom has produced one of the greatest spending sprees in technology history, with hundreds of billions of dollars flowing into data centers, advanced processors, and power infrastructure. An uncomfortable question is now emerging among investors: will those investments generate enough revenue to justify their cost? With technology stocks already under pressure, the market is questioning whether the industry's spending ambitions have outpaced its earning power. The answer could determine whether the AI boom has years of growth ahead — or whether Wall Street has badly mispriced its most celebrated companies.

What Is the $176 Billion Problem Burry Identifies?

Burry has spent months warning that the AI boom could be approaching a reckoning, and his concern centers on how companies account for the billions they spend on AI infrastructure.

When Microsoft (NASDAQ:MSFT), Amazon (NASDAQ:AMZN), and other technology giants buy Nvidia processors, they do not immediately charge the entire purchase against earnings. Instead, they depreciate the equipment over its estimated useful life (the accounting period over which an asset's cost is expensed). The longer that period, the smaller the annual expense — and the higher reported profits appear.

Burry has argued that companies use depreciation schedules that do not adequately reflect how quickly AI chips become outdated. He puts the resulting understatement of depreciation at up to $176 billion across 2026 to 2028.

The arithmetic is straightforward. A $10 billion investment in processors, spread over six years, produces annual depreciation of approximately $1.67 billion. Shorten the useful life to three years, and that expense doubles to $3.33 billion. Nothing about the underlying business changed. Yet annual accounting profits would fall by another $1.67 billion.

There is a counterargument, and the source acknowledges it. Longer depreciation schedules are not automatically improper, because older chips can remain productive even after newer generations arrive. Burry's concern is that rapid technological advances in AI hardware could make those accounting assumptions increasingly optimistic over time.

Why Do Nvidia's Customers Matter More Than Its CEO?

The more immediate threat to Nvidia, according to Burry's thesis, is what happens when customers begin questioning their returns on AI spending.

Amazon's October 29 earnings release provides a revealing example. Amazon Web Services revenue increased 39% to $46 billion in the third quarter. Yet Amazon's companywide free cash flow (cash generated after capital expenditures) swung from a $14.8 billion inflow to a $42.4 billion outflow over the trailing 12 months, primarily because of increased AI infrastructure spending.

Microsoft presents a different picture. Its Azure revenue grew 48% in the September quarter, while the company generated $18.6 billion in free cash flow despite $45.2 billion in capital expenditures. Both companies are expanding their AI businesses. But their cash flows tell different stories.

The danger of betting against Nvidia and the AI boom, as the report notes, is that strong demand can persist longer than skeptics expect. Nvidia's fiscal second-quarter revenue surged 106% to $96.2 billion, while data center revenue climbed 117% to $89 billion. Its gross margin reached 75%. Those are not numbers suggesting customers have stopped buying.

But if Microsoft's, Amazon's, and other hyperscalers' (large cloud infrastructure operators) AI investments eventually fail to deliver adequate returns, today's spending boom could become tomorrow's capital expenditure cuts. Nvidia would feel those reductions directly.

What Does Huang's Resignation Change?

Huang's departure introduces real uncertainty of its own. Nvidia has named CFO Colette Kress interim CEO, and investors must now assess whether the leadership transition will disrupt the company's customer relationships and product roadmap.

The damage already inflicted on Nvidia shareholders demonstrates how quickly confidence can disappear. Nvidia fell 8.3% after hours on a report of a second Jensen Huang affair, and the stock has declined more than 30% from its peak this year, per the report's account of the week's losses.

Bur Burry's warning goes beyond one executive. In short, Nvidia's greatest long-term vulnerability is not necessarily losing Jensen Huang. It is losing the willingness of its biggest customers to keep spending at today's extraordinary pace.

What Should Investors Watch Next?

The report closes with a concrete framework. Investors should not confuse a 30%-plus stock decline with proof that Nvidia is suddenly cheap.

Before buying the dip, the analysis points to three indicators:

• Whether hyperscaler free cash flow improves
• Whether depreciation expenses accelerate
• Whether AI capital spending continues growing

Those numbers will ultimately tell shareholders more about Nvidia's future than the name on the CEO's office door.