Discover top fintech news and events!
Subscribe to FinTech Weekly's newsletter
Read by executives at JP Morgan, Coinbase, Blackrock, Klarna and more
An Investment Fight That Became a Story About AI Itself
The debate between Michael Burry and Nvidia did not begin as a public quarrel. It grew slowly, through scattered remarks and early options filings, then erupted into a direct exchange that spilled into financial newsrooms, investor forums, and the broader conversation around the AI boom.
There are few moments when a single critic challenges a company worth several trillion dollars, and fewer still when that company responds with an internal memo addressing him directly. That unusual exchange reveals something deeper than a disagreement between an investor and a corporation. It exposes the tension between an era of soaring enthusiasm for artificial intelligence and the colder arithmetic underneath.
Many readers know Burry for his calls during the housing crisis. His recent focus is different. He is not warning about consumer credit or mortgage bonds. He is dissecting the economics behind the AI race, questioning whether companies racing to deploy infrastructure are building on assumptions that may not last.
Nvidia stands in the center of that debate because its chips have become the core of the current AI build-out. That prominence turns every concern about spending, depreciation, or incentives into a question about Nvidia’s future, even when the company itself is not the direct target.
How a Valuation Dispute Became a Public Clash
The seeds of this confrontation appeared long before the headlines. In 2023 and 2024, Burry cautioned investors about excessive optimism in technology. Those warnings did not single out Nvidia by name. They drifted across markets as commentary about valuations, interest rates, and speculation. The public still associated him with macro doom, not with AI economics.
The tone began to shift by late 2024. Regulatory filings abroad showed that Scion Asset Management had purchased large put options tied to Nvidia and Palantir. The position was quiet, unusual in size, and early. He was not yet arguing with Nvidia, but he was placing a bet that the AI story would meet resistance.
The moment the debate became unavoidable arrived with Scion’s Q3 2025 filing. The document revealed more than a cautious hedge. It disclosed more than a billion dollars in notional puts across Nvidia and Palantir. Financial media described the position as a direct challenge to the AI rally. Discussions intensified, and readers began looking for the reasoning behind such an extraordinary wager.
That reasoning arrived soon after.
During November 2025, Burry detailed what he considered fundamental flaws in how the AI boom is understood. The arguments covered accounting choices, cost assumptions, corporate incentives, and the psychology of heavy-capex technology cycles. They were not limited to Nvidia, but the company stood at the center because it supplies the hardware that fuels the race. His concerns soon reached a point where Nvidia responded.
The Accounting Debate at the Core of Burry’s Warning
A major part of Burry’s position concerns the useful life of AI hardware. Companies buying advanced GPUs often depreciate those assets over several years. These schedules allow the cost to be spread across time, reducing near-term expenses and raising reported earnings. Burry believes this practice obscures the economic truth. His view is that AI hardware becomes outdated too quickly to justify such long schedules. He estimates that a more realistic life might be closer to three years.
That difference matters because these chips are expensive. Long depreciation periods lower each quarter’s stated cost. Shorter periods would reduce earnings for many AI adopters and would shrink the apparent return on their AI investments. If these companies are counting on six years of productivity from assets that last only half as long, their real margins may be far thinner than they appear.
This is not an accusation of fraud. It is an argument about pace. AI hardware progresses quickly. A chip released today competes with a successor in a short window. If depreciation schedules fail to align with that reality, the gap eventually appears in write-downs or in reduced enthusiasm for new purchases. Burry’s analysis argues that this gap could become evident by 2026–2028. For a company that benefits from customers buying rapidly and aggressively, a slowdown could be significant.
How Incentives Drive Spending in the AI Era
Another layer of his warning concerns the motives that drive the ongoing AI build-out. Many companies invest not only for returns but also out of fear of being left behind. When a technology is framed as the future of work, decision-makers may feel pressure to deploy capital quickly, even if expected benefits are uncertain. Burry argues that this competitive tension can lead to spending that outruns economic justification.
The argument is straightforward. If companies are driven by fear of missing out, or by internal pressure to appear innovative, their procurement decisions may not match long-term performance. When the cost of capital remains manageable, these decisions feel safe. When conditions tighten, the same decisions become burdens.
Nvidia’s success is tied to that wave of demand. The question raised by Burry is whether that demand reflects durable economics or a cycle that could slow once budgets tighten or performance expectations shift.
The Disputed Territory of Vendor Financing
Some of the most controversial commentary involves the idea of vendor or circular financing. Analysts such as Jim Chanos have echoed concerns that incentives or financing structures may exist within the AI ecosystem that create artificial demand.
In this view, companies that benefit from AI infrastructure growth might be assisting purchasers in ways that inflate current sales. Nvidia has firmly denied this. The company states that it does not engage in such practices and that its demand comes from genuine, independent purchasing decisions.
This debate persists because vendor financing has existed in other industries. When demand surges, some companies support clients to sustain momentum. Whether any such patterns exist in the AI sector remains uncertain. Without evidence, the dispute remains one of suspicion versus denial. Any confirmation in the future would carry significant consequences. For now, it remains an unresolved question in the broader discussion.
Stock-Based Compensation and the Question of Owner Earnings
Burry also focuses on stock-based compensation. Many technology companies use this method to reward employees. The cost appears indirectly, often through dilution. Buybacks can offset dilution, but Burry argues that Nvidia’s stock-based pay has lowered long-term “owner earnings” more than headline figures suggest. The implication is not wrongdoing. The implication is that shareholders may not be receiving as much economic value as reported earnings indicate once compensation mechanics are considered.
This concern resonates in parts of the tech world because stock-based pay has grown across the sector. Investors increasingly look for clarity about how much cash returns remain after these adjustments. Nvidia defends its practices as standard and compliant. The debate reflects a difference between reported accounting results and long-term economic reality.
A Comparison Rooted in the Dot-Com Past
Burry’s most provocative analogy compares Nvidia not to fraudulent companies of the past but to Cisco during the dot-com surge. Cisco was a real company with real revenue and real innovation. Its products became essential to the internet. Yet its valuation soared far beyond sustainable levels. When spending cooled, the stock fell hard and took years to recover.
The analogy suggests that Nvidia may be central to AI’s future while still being priced in a way that leaves little room for normal fluctuations. Nvidia rejects this view. The company says that AI demand is broad, persistent, and supported by many industries, including data centers, automotive, edge computing, and enterprise use cases. The disagreement centers on whether this demand remains as strong when capital turns cautious.
Nvidia’s Memo and the Moment the Dispute Became Public
The conflict escalated when Nvidia drafted a memo addressing criticisms circulating in the market. The memo defended its accounting, denied any circular financing, and rejected comparisons to corporate failures of the past. Analysts reported that the document quoted Burry directly. That acknowledgment elevated a debate that had previously been one-sided.
Burry responded publicly, stating that the memo misrepresented parts of his argument and avoided the central point about depreciation by focusing on the wrong target. His thesis focused on the buyers of AI hardware, not on Nvidia’s internal practices. He reiterated his expectation that companies will take significant write-downs in the next few years as newer chips render current equipment less valuable.
The direct exchange between an individual investor and a company of Nvidia’s size was unusual. For many observers, that moment transformed a technical debate into a clear public dispute.
The Options That Turned the Debate Into a Statement
The scale of Burry’s put positions gave weight to his arguments. Notional value refers to the underlying value of the shares tied to an option contract. Holding puts with more than a billion dollars in notional exposure signals conviction, even if the actual capital outlay is far smaller. Burry later noted that the premium paid for these positions was around ten million dollars each. The impact of the positions lies not in cost but in exposure. If he proves correct, the payoff is significant. If not, the trade expires.
Notional options allow investors to express large views with limited upfront capital. The positions reflect belief in a substantial decline, not a mild correction. They also invite scrutiny. Investors and analysts now monitor filings to see whether he increases or reduces the exposure.
Where This Leaves the AI Boom
The debate reveals uncertainty about the durability of AI capital spending. If depreciation schedules tighten, if financing conditions change, or if enthusiasm cools, the effect would ripple across the sector. Nvidia’s strength is tied to demand that has grown rapidly. The company argues that demand extends far beyond data centers and that it can diversify across new markets. Burry questions whether the economics can sustain that expansion.
Investors now watch for clues. The decisions hyperscalers make about useful life and write-downs will matter. The pace of AI capex will matter. The existence or absence of any financing arrangements will matter. The question is not whether AI will remain central to technology. The question is whether the current rate of spending matches long-term value creation.
A Debate That Tells a Larger Story
The Burry–Nvidia dispute is not only a clash about valuation. It is a signal of the tension between extraordinary technological progress and the financial systems that support it. AI promises to transform entire sectors, including fintech, health care, and manufacturing. That promise does not eliminate the need for disciplined accounting and clear incentives.
Nvidia has become a symbol of the AI era. Burry has become a symbol of skepticism grounded in financial history. Their disagreement reflects competing visions of risk and reward during a moment of intense technological acceleration. Whether the coming years confirm one side or the other, the debate highlights how complex the economics of AI have become.
The world is living through a period when technology races ahead and financial frameworks try to keep pace. The result is a conflict that feels larger than a single company or a single investor. The truth will emerge through time, earnings reports, depreciation schedules, capital budgets, and the decisions that follow. The only certainty is that both sides believe they see the future clearly. The market will decide which version endures.
