Michael Burry's NVIDIA Short Position: AI Chip Depreciation Accounting Controversy

This analysis is based on a Reddit discussion [1] published on November 14, 2025, highlighting Michael Burry’s bearish position on NVIDIA (NVDA) centered on aggressive depreciation assumptions for AI infrastructure.

The central issue revolves around depreciation assumptions for AI chips and servers. Major cloud providers including Microsoft, Google, Oracle, Meta, and Amazon currently depreciate AI infrastructure over 5-6 years [2]. However, Burry argues the actual useful life is only 2-3 years due to rapid technological obsolescence [2]. This discrepancy creates significant financial implications:

• Financial Impact
  : Burry estimates that between 2026-2028, these accounting choices could understate depreciation by approximately $176 billion [2]
• Profit Inflation
  : Oracle’s profits could appear 27% higher, Meta’s 21% higher under more conservative assumptions [2]
• NVIDIA’s Position
  : The company benefits from current practices but faces risks if accounting standards change

NVIDIA’s stock has shown recent volatility amid these concerns. As of November 14, 2025, NVDA was trading at $188.10 (+0.66%) with a market cap of $4.58T [0]. However, the stock experienced significant pressure including a 3.58% drop on November 13th, falling from its peak above $200 earlier in November [0]. This reflects growing concerns about AI infrastructure spending and valuation levels.

: NVIDIA has accelerated its hardware release cycle from every 2-3 years to annually [2]. CEO Jensen Huang himself stated that when Blackwell chips ship in volume, “you couldn’t give Hoppers away” [2], acknowledging rapid obsolescence.

: CoreWeave reports strong demand for older chips (A100s fully booked, H100s at 95% of original price) [2], suggesting some residual value, though this may not represent broader market conditions.

1. Accounting Standards Risk
   : If auditors or regulators force more conservative depreciation assumptions, earnings across the AI sector could see substantial downward revisions [2]

2. Hardware Lifecycle Economics
   : The debate highlights a fundamental mismatch between accounting practices and technological reality in the AI era, where Moore’s Law acceleration meets traditional depreciation schedules

3. Customer Cash Flow Dynamics
   : The original Reddit concern about customers facing “cash squeezes” from rapid hardware turnover represents a legitimate demand-side risk for NVIDIA

• Earnings Quality
  : The controversy raises questions about earnings quality across major tech companies, potentially affecting investor confidence in the entire AI sector
• Investment Thesis Validity
  : Current AI infrastructure investment models may need reassessment if hardware turnover costs are higher than currently accounted for
• Second-Hand Market Development
  : The viability of a robust secondary market for AI chips becomes crucial for justifying longer depreciation periods

1. Accounting Restatement Risk
   : Barclays has already cut earnings forecasts of AI firms by up to 10% for 2025 to account for more realistic depreciation assumptions [2]

2. Hardware Obsolescence Acceleration
   : The annual release cycle suggests faster-than-expected obsolescence could make current accounting assumptions untenable [2]

3. Customer Cash Flow Strain
   : If AI companies face cash pressures from rapid hardware turnover, they may cut orders, directly impacting NVIDIA’s revenue [1][2]

4. Regulatory Scrutiny
   : As Burry’s concerns gain traction, regulatory bodies may examine these accounting practices more closely [2]

1. Second-Hand Market Development
   : Companies that can effectively create and manage secondary markets for AI chips may mitigate depreciation concerns

2. Accounting Innovation
   : Opportunities exist for developing more accurate depreciation models that reflect rapid technological change

3. Customer Financing Solutions
   : NVIDIA could develop financing solutions that help customers manage cash flow pressures from rapid hardware turnover

NVIDIA’s fundamental metrics remain strong [0]:

• Revenue Breakdown (FY2025)
  : Data Center: $115.19B (88.3%), Gaming: $11.35B (8.7%)
• Profitability
  : Net Profit Margin: 52.41%, Operating Margin: 58.09%
• Valuation
  : P/E Ratio: 53.09x, Market Cap: $4.59T

• Analyst Consensus
  : Overwhelmingly positive with 73.4% Buy ratings and a consensus target of $235.00 (+24.7% from current) [0]
• Recent Pressure
  : Tech sector selloff intensified on November 13-14, with NVDA dropping alongside other AI stocks [0]
• Additional Factors
  : SoftBank’s $5.8 billion Nvidia stake sale added to downward pressure [0]

1. Q3 2025 Earnings (November 20)
   : Watch for any changes in depreciation guidance or customer comments on hardware lifecycle [0]
2. Auditor Scrutiny
   : Monitor 10-K filings for any auditor challenges to depreciation assumptions
3. Second-Hand GPU Market
   : Track pricing and demand trends for older generation AI chips
4. Customer CapEx Patterns
   : Watch for changes in hyperscaler spending patterns or comments on hardware refresh cycles

The depreciation controversy represents a significant challenge to current AI infrastructure investment thesis, requiring careful monitoring of accounting practices, technological advancement rates, and customer cash flow dynamics.