Michael Burry Questions Big Tech Earnings Quality Through Extended Depreciation Practices

This analysis is based on recent market reports examining Michael Burry’s criticism of Big Tech accounting practices [1][2][3]. The famed investor has specifically targeted Meta, Alphabet, Amazon, and Microsoft for extending depreciation schedules on servers and GPUs, thereby reducing non-cash charges and boosting net income [1].

The controversy centers on depreciation schedule changes that significantly affect earnings quality. Meta has been most explicit, extending the estimated useful lives of “certain servers and network assets” from 4-5 years to 5.5 years effective January 2025, estimating this change would reduce 2025 depreciation expense by $2.9 billion [1][2]. Microsoft and Alphabet have made similar moves, while notably, Amazon took the opposite approach in February 2025, shortening its server equipment useful life to 5 years from 6 years [1].

Burry estimates that Big Tech will understate depreciation by approximately $176 billion between 2026 and 2028, with projected earnings inflation of 26.9% at Oracle and 20.8% at Meta [2][3]. This represents a material accounting adjustment that could fundamentally alter earnings perception across the technology sector.

The market reaction to these concerns has been notably divergent [0]:

• Meta (META): Down 13.90% in the past month, significantly underperforming
• Alphabet (GOOGL): Up 45.92% year-to-date, outperforming significantly
• Amazon (AMZN): Up 6.89% year-to-date
• Microsoft (MSFT): Up 21.39% year-to-date

This performance gap suggests investors are evaluating the accounting concerns differently across companies, potentially reflecting varying levels of exposure to depreciation issues and differing business models.

The four major AI infrastructure spenders are projected to boost combined capital expenditures by approximately 40% to $460 billion in the next 12 months [1]. Despite accounting changes, depreciation expenses are accelerating dramatically - combined depreciation for Alphabet, Microsoft, and Meta rose from about $10 billion in Q4 2023 to nearly $22 billion in the most recent quarter, with projections suggesting this could reach almost $30 billion by this time next year [1].

Paradoxically, despite rising depreciation costs, the Magnificent Seven companies are on pace to deliver 27% year-over-year earnings growth, nearly double the 14% expansion projected at the start of earnings season [1]. This disconnect between rising capital costs and accelerating earnings growth lies at the heart of Burry’s concerns.

The core debate centers on whether GPUs and servers truly have the extended useful lives that companies claim, versus their economic obsolescence in a rapidly advancing AI landscape [2][3]. Nvidia has accelerated its chip release cycle from every 18-24 months to annually, suggesting hardware becomes functionally obsolete faster than accounting schedules reflect [2]. This creates a fundamental mismatch between accounting depreciation and actual economic useful life.

Reports indicate data centers in Santa Clara and Northern Virginia are sitting idle waiting for grid connections, creating a scenario where billions in GPU assets are depreciating without generating revenue [2]. This raises questions about the efficiency of capital deployment and the timing of returns on massive AI investments.

Nearly half of all 401(k) money is now effectively tied to six megacap tech stocks, creating unprecedented concentration risk [2]. Historical parallels to previous capital spending booms in shale, fiber, and railroads that ended in overcapacity and low returns suggest caution may be warranted.

1. Accounting Quality Risk
   : If Burry’s $176 billion depreciation understatement estimate proves accurate, it could represent a material misstatement of earnings across multiple quarters [2][3].

2. Technological Acceleration Risk
   : The mismatch between accelerating chip development cycles (now annual) and extended depreciation schedules (5-6 years) creates a growing accounting gap [2].

3. Capital Intensity vs. Productivity
   : The market may be confusing massive spending with genuine growth, particularly when individual data center projects can cost $50 billion each [2].

4. Concentration Risk
   : The extreme concentration of investment in a handful of megacap stocks creates systemic risk if the AI thesis falters [2].

1. Quarterly Depreciation Trends
   : Accelerating depreciation expenses could signal the beginning of earnings pressure [1]

2. AI Revenue Generation
   : Actual revenue from AI investments versus capital spending patterns

3. Power Infrastructure Progress
   : Resolution of grid connection issues affecting data center utilization [2]

4. Competitive Dynamics
   : Whether AI spending translates to sustainable competitive advantages or merely raises the cost of competition

The analysis reveals significant concerns about earnings quality in Big Tech companies, particularly regarding depreciation practices for AI infrastructure. While companies maintain premium valuations (Meta: 26.22x P/E, Alphabet: 26.88x P/E, Amazon: 32.86x P/E, Microsoft: 36.00x P/E) [0], the sustainability of earnings growth remains questionable given the massive capital expenditures and potential accounting adjustments.

Supporting arguments for current valuations include Bank of America analysts’ view that market skepticism about AI capex indicates the trade is less crowded than critics claim, and surging ancillary segments like memory and optical components suggesting robust underlying demand [2]. However, the fundamental disconnect between accelerating depreciation costs and reported earnings growth warrants careful scrutiny.

The market’s divergent reaction across companies suggests investors are performing differentiated analysis of each company’s specific exposure to these accounting concerns and their underlying business models.