Michael Burry, a monumental investor big short Predictions of a subprime mortgage collapse have sparked new alarm about semiconductor giant Nvidia ( NVDA ), which is at the center of the artificial intelligence (AI) boom.
In a recent Substack post published on Thursday, Burry argued that Nvidia’s latest financial disclosures revealed a build-up of supply commitments, similar to the pattern seen at Cisco Systems (CSCO) during the final stages of the dot-com bubble. His warning focused on a sharp increase in purchase obligations, a balance sheet detail that he believed indicated growing internal risks rather than temporary external pressures.
According to Nvidia’s latest financial report, its purchase obligations surged to $95.2 billion from $16.1 billion a year ago. If other supply-related obligations are included (such as inventory and additional purchase agreements), the total is closer to $117 billion. That number nearly matches the company’s annual operating cash flow.
During the company’s fiscal fourth-quarter earnings call on Wednesday, Chief Financial Officer Colette Kress acknowledged that inventory was up 8% from the previous quarter. She said Nvidia has “strategically secured inventory and capacity to meet demand in the coming quarters, later than usual.”
For Bury, this language marked a fundamental shift in strategy.
He believes Nvidia committed large amounts of supply before fully understanding long-term demand. The result: more capital tied up in inventory and contractual obligations with longer maturities.
“What is happening now is not temporary,” Burry wrote. “This is not an export shock. It’s not even an external shock. This is coming from within the business plan.” He described the move as a deliberate effort by Nvidia to further lock in future supply chain capacity, something Nvidia has never done.
Burry’s comparisons to Cisco date back to 2000 and 2001, when the networking equipment maker aggressively pursued supplier commitments and projected continued growth of 50 percent annually. When enterprise technology spending suddenly cooled in the wake of the dot-com bubble, Cisco was saddled with excess inventory and supply agreements it no longer needed.
The consequences are serious. Cisco ended up writing down billions of dollars in inventory and saw its stock price collapse along with the broader tech industry.
“This is not business as usual. This is risk,” Burry said of Nvidia’s current posture.
At the heart of his concern is that today’s extraordinary demands driven by artificial intelligence may not continue at the same pace indefinitely. If enterprise AI spending moderates or competitive dynamics change, Nvidia could find itself stuck with supply commitments made under more optimistic assumptions.
Margin buffer or mirage?
To be sure, Nvidia’s financial situation is very different from Cisco’s during the height of the tech bubble. Nvidia’s gross profit margin is now over 70%, well above Cisco’s levels two decades ago.
Burry acknowledged that higher profitability could provide some cushion during an economic downturn. However, he questioned whether those margins could be sustained if demand cooled.
He said current profitability is amplified by exceptional demand conditions and Nvidia’s ability to command premiums amid limited GPU supply. If the balance of supply and demand changes, pricing power, and hence profits, could be quickly compressed.
“As demand changes, this margin could recover quickly,” he wrote.
Nvidia shares fell nearly 4% in early trading on Thursday after the earnings report was released, although the stock remains one of the decisive winners in the artificial intelligence era. The losses continued on Friday, with NVDA currently down nearly 3% so far this week. Surges in demand for AI chips following the launch of ChatGPT and other AI software have led to explosive gains from 2023 to 2025, but the stock is down about 3% so far this year.
The differing views underscore a broader debate unfolding on Wall Street: whether Nvidia’s dominance in artificial intelligence infrastructure represents the early stages of a long-term expansion or a late-stage boom vulnerable to classic cycles of overinvestment and overcapacity.
For Bury, the risk lies not in external shocks but in internal confidence. To him, this confidence may look very familiar.
While the picture painted may look relatively bleak, Nvidia is still up 50% over the past 12 months. Many analysts continue to express confidence in the stock, and major hyperscalers pouring billions into AI will only continue to pour more money. In addition, Nvidia CEO Jensen Huang continued to accept media interviews and expressed confidence in the future of artificial intelligence and Nvidia.
45 of the 50 analysts tracked by Barchart currently rate NVDA as a Strong Buy. Only 1 has a “Strong Sell” rating, 1 has a “Neutral” rating, and the rest have given it a “Buy” rating. The average price target for Nvidia is an impressive $255, which leaves Nvidia with room for 41% upside over the next 12 months.
On the date of publication, Caleb Naysmith did not hold (either directly or indirectly) any securities mentioned in this article. All information and data in this article are for reference only. This article was originally published on Barchart.com
