When accounting issues become a hot topic on Wall Street, it’s rarely a good sign for the stock market. That’s why recent debates over the depreciation schedules of AI computing infrastructure among U.S. tech giants should give investors deeply immersed in the AI investment frenzy pause as they approach 2026.
Historically, financial reporting disputes were too obscure—and frankly, too dull—to capture much investor attention before the early 2000s. However, after accounting scandals at Enron, WorldCom, and Adelphia Communications exposed deceptive practices, "accounting fraud" became a recurring market concern, often preceding sharp market downturns. These cases triggered major reforms, including the Sarbanes-Oxley Act, which largely eliminated blatant fraud among U.S. public companies.
Yet, misleading but legal financial disclosures remain a significant risk. If a company’s reported profits are revealed to be disconnected from economic reality, its stock could plummet—and if the company is large enough (like today’s "Magnificent Seven"), it could drag down the broader market.
The S&P 500’s 30 trillion-dollar rally over the past three years has been driven largely by the world’s biggest tech firms (e.g., the Magnificent Seven), AI infrastructure players (e.g., Micron, TSMC, Broadcom), and energy suppliers (e.g., Constellation Energy). The Magnificent Seven—Apple, Microsoft, Alphabet, Tesla, NVIDIA, Amazon, and Meta Platforms—account for roughly 35% of the S&P 500 and are seen as the prime beneficiaries of the AI-driven technological revolution.
**A Growing Concern: The "Depreciation Debate"** Against this backdrop, a tweet by "The Big Short" investor Michael Burry has reignited worries about tech giants’ depreciation practices. Burry argued that Meta, Oracle, and even NVIDIA are overstating profitability by spreading semiconductor costs over extended periods. Longer depreciation schedules reduce annual expenses, boosting reported profits—but Burry contends that rapid technological obsolescence makes chips lose value faster than assumed.
NVIDIA, the $4 trillion AI chip leader dubbed "the world’s most important stock," isn’t alone. Oracle, Alphabet, Amazon, Meta, and Microsoft have also extended the assumed useful lives of major assets since 2020. Even IBM has seen its depreciation costs halve since 2020 while revenue grew, raising eyebrows.
**The Illusion of Value** Depreciation schedules don’t create real economic value—they’re accounting tools. Free cash flow, not expense allocation, determines shareholder value. Extending depreciation lowers annual charges, flattering earnings per share (EPS) and P/E ratios, but doesn’t increase cash flow. While tax benefits exist, financial and tax reporting are separate, so tweaking depreciation schedules doesn’t always reduce tax bills.
**Why Manipulate Depreciation?** In the 1980s–90s, EPS-linked executive pay made such moves understandable. Today, with compensation tied to stock performance, the motive may be to justify lofty valuations. At 36x P/E, tech stocks trade at a premium to the S&P 500’s 25x. Artificially lowering P/E ratios could ease pressure on CEOs and CFOs to defend valuations.
No evidence suggests current tech giants are committing fraud, but depreciation adjustments serve as "earnings optics tools"—managing expectations rather than improving fundamentals. However, with tech stocks priced for perfection, any depreciation-related skepticism could trigger a painful market correction, especially as the AI bull narrative hangs in the balance.
Comments