Michael Burry, the investor famous for predicting the 2008 financial crisis, has raised fresh questions about Nvidia’s revenue reporting, particularly its sales linked to China and Singapore. Burry believes a change in the chipmaker’s method of reporting geographic revenue may have made China’s earlier sales decline appear less pronounced before US export restrictions on advanced AI chips took effect.
In a recent post on his Substack publication, Cassandra Unchained, Burry described the effect of the reporting change as a possible form of “mild manipulation.” His comments have renewed scrutiny of Nvidia’s exposure to China, the transparency of its international sales and the risks surrounding the global AI chip market. However, Burry acknowledged that restating earlier financial periods is standard practice when a company changes its reporting methodology. His criticism concerns the timing and effect of the change, not an established finding of accounting fraud.
Why Michael Burry Is Questioning Nvidia’s China Sales
Burry’s concerns centre on the geographical distribution of Nvidia’s revenue.
For years, Nvidia reported geographic revenue based on the billing location of its customers. This meant that sales were attributed to the country where a customer was invoiced, even when the chips were ultimately shipped to and installed in another country.
Burry pointed out that Nvidia’s reported revenue share from China had declined in the years before US restrictions on advanced AI chip exports became more restrictive. At the same time, Singapore’s share of reported revenue increased significantly.
He questioned whether the geographic figures accurately reflected where Nvidia’s AI chips were eventually being used.
His argument is that the reporting method may have obscured the actual destination of some products because billing locations do not necessarily reveal where customers deploy the hardware.
Nvidia Changed Its Geographic Revenue Reporting Method
A major part of Burry’s criticism involves a reporting change Nvidia introduced in late 2025.
The company moved from reporting geographic revenue according to customer billing locations to using the headquarters locations of its customers. Nvidia also restated prior-period figures to make them comparable under the revised methodology.
In its regulatory filings, Nvidia explained that billing locations often reflected centralised invoicing arrangements rather than the destinations where products were shipped. The company said using customer headquarters provided a better representation of its geographic revenue profile.
The distinction is significant because a customer’s headquarters, billing address and the location where its AI infrastructure operates can all be different.
For example, an international company may centralise its invoicing in Singapore while operating data centres in several other countries. Under the previous method, revenue could be attributed to Singapore even if the hardware was shipped elsewhere.
Burry argues that the revised methodology made China’s earlier revenue decline appear less pronounced in the restated historical figures. He also questioned why Singapore disappeared as a separately reported geographic category.
What Nvidia’s Earlier Financial Figures Show
Nvidia’s FY25 annual report provides an illustration of why the geographic reporting issue has attracted attention.
Under the earlier billing-location method, Singapore accounted for $23.68 billion, or approximately 18% of Nvidia’s FY25 revenue. China, including Hong Kong, accounted for $17.11 billion.
However, Nvidia specifically explained that Singapore-based customers used the country to centralise invoicing and that shipments to Singapore represented less than 2% of total revenue in FY25.
| Geographic market | FY25 revenue under billing-location method |
|---|---|
| United States | $61.26 billion |
| Singapore | $23.68 billion |
| Taiwan | $20.57 billion |
| China, including Hong Kong | $17.11 billion |
| Other | $7.88 billion |
| Total | $130.50 billion |
These figures demonstrate why billing location should not automatically be interpreted as the final destination of Nvidia’s products.
Under the revised reporting method, Nvidia’s geographic revenue categories changed. Its November 2025 filing used customer headquarters locations and no longer listed Singapore as a separate category.
The company said the revised approach offered a better picture of its geographic revenue distribution. Burry, however, believes the change made it more difficult to understand the actual geographical exposure of its chip sales.
Why Singapore and China Matter to Nvidia’s AI Business
China is an important market for the global semiconductor industry, but US export restrictions have significantly complicated Nvidia’s ability to sell its most advanced AI chips to Chinese customers.
Washington has introduced restrictions on exports of certain high-performance semiconductors to China, citing national security concerns. These controls have limited the products Nvidia can legally sell into the market and created uncertainty around future demand.
Singapore has also attracted attention because of its role in international technology supply chains.
The Wall Street Journal recently reported that Chinese companies were accessing Nvidia AI chips through data centres in Southeast Asia, raising questions about whether overseas infrastructure could be used to circumvent US restrictions.
This issue makes the distinction between a customer’s headquarters, billing location and the final deployment location particularly important.
Regulators and investors want to understand whether restricted chips are reaching prohibited destinations through intermediaries. Geographic revenue disclosures alone may not answer that question, but they can provide useful information about the structure of Nvidia’s international business.
What Does Michael Burry Expect for Nvidia Stock?
Burry’s latest criticism adds to his broader concerns about Nvidia and the valuation of companies benefiting from the artificial intelligence boom.
He has repeatedly questioned the sustainability of AI-related investment and the financial arrangements supporting massive data-centre spending. His concerns include the possibility that major technology companies are committing too much capital to AI infrastructure before the long-term returns become clear.
His position on Nvidia is more complicated than a simple prediction of an immediate stock collapse.
In an earlier November 2026 post, Burry suggested that Nvidia shares could become an attractive short-term trading opportunity following their selloff. However, he also said he was not interested in holding the stock as a long-term investment from those levels.
He has warned that Nvidia’s competitive advantage could weaken over the next five years as competitors develop alternative AI chips and customers explore ways to reduce dependence on its technology.
His latest comments about geographic revenue reporting reinforce his concerns about the company’s exposure to China and the transparency of its international business.
Burry’s views should not be treated as a formal investment recommendation or a guarantee of future share-price performance. His comments reflect his personal investment analysis.
Nvidia Faces Competition and AI Spending Risks
Beyond the China revenue issue, Nvidia faces several challenges that could influence its long-term stock performance.
Growing competition: Semiconductor companies and large technology businesses are developing alternative AI accelerators. Google’s Tensor Processing Units, AMD’s AI chips and custom processors developed by major cloud providers could challenge Nvidia’s dominance in selected workloads.
Customer concentration: Nvidia depends heavily on a relatively small group of major technology customers for a substantial portion of its data-centre business. Any reduction in their AI infrastructure spending could affect demand.
AI infrastructure spending: Technology companies are investing enormous amounts in data centres, networking equipment and advanced processors. Investors are increasingly questioning how quickly these investments will generate sufficient returns.
Export restrictions: US restrictions on advanced semiconductor exports to China could limit Nvidia’s access to an important market. Changes in export policy could also create uncertainty around product launches and sales opportunities.
Valuation concerns: Nvidia’s stock performance is closely tied to expectations for future AI demand. If revenue growth or profit margins fall short of market expectations, the shares could face pressure even if the company remains profitable.
The Bigger Picture
Burry’s criticism highlights a broader challenge facing global technology companies: how to provide investors with a clear picture of revenue generated across complicated international supply chains. For semiconductor businesses, the country where a customer is headquartered may not be the country where a chip is ultimately installed or used.
Nvidia has explained that its revised reporting method better reflects the geographic profile of its customers. However, the change has raised questions about how investors should interpret its China exposure, particularly amid tighter US export controls and concerns about indirect access to restricted AI hardware. The disagreement does not establish that Nvidia manipulated its accounts, but it demonstrates why transparent disclosures matter in an industry with significant geopolitical and regulatory risks.
Looking Ahead
Investors will be watching Nvidia’s upcoming financial results and management commentary for additional information about demand, international sales and the impact of export restrictions. The company’s ability to maintain its competitive position, sustain data-centre growth and provide greater clarity around its geographic revenue will be important factors influencing market sentiment. Any additional regulatory scrutiny of overseas AI chip shipments could also affect expectations for its China-related business.
For Burry, the central question is whether Nvidia can maintain its competitive advantage as AI infrastructure spending expands and the semiconductor market becomes more competitive. His latest warning reflects continued caution rather than a confirmed prediction of an immediate collapse in the stock. Investors will ultimately need to evaluate Nvidia’s financial disclosures, business performance, valuation and long-term demand rather than relying solely on one prominent investor’s comments.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.



