The hidden danger of the "AI Bull Market Narrative" in 2026: The "depreciation trick" of American tech giants.

Usually, when an accounting issue becomes a hot topic on Wall Street for a period of time, it is definitely not a good sign for the stock market. That is why the recent discussions surrounding the depreciation "useful life" arrangements and plans of the AI computing infrastructure of American tech giants should make those investors who are deeply immersed in the AI investment frenzy cautious as they approach 2026. In general, matters debated by those who prepare company financial statements in the financial market are too obscure -- and frankly, too boring, so until the early years of this century, they hardly sparked active exploration interest among stock investors. However, since the early 2000s, accounting fraud issues at once high-flying companies like Enron, WorldCom, and Adelphia Communications that were exposed for using deceptive accounting practices in the capital markets have occasionally made headlines, causing heated debates, and shortly thereafter, the stock market may experience a rapid downturn. These companies quickly went from seemingly brilliant "admired by everyone" success to bankruptcy, triggering widespread market chain reactions. For a period of time, investors began to doubt whether they could trust any company's comprehensive representation of its profit potential and financial condition. This severe accounting fraud -- which is actually criminal fraud -- once triggered significant changes in the US stock market, the most famous being the Sarbanes-Oxley Act. It largely made open fraudulent practices in US listed companies a thing of the past. However, this does not mean that investors can relax their guard in the long term. Misleading but legal financial/accounting disclosures still pose significant risks to the market. If a company's financial reports reveal profits that are completely disconnected from its underlying economic fundamentals, the company's stock price may plummet significantly over a period of time. And if the company's market value is large enough (like the current Magnificent Seven), it could even drag down the broader US stock market as a whole. The S&P 500 index's approximately $30 trillion surge in the past three years, driven largely by the world's largest tech giants (such as the US's seven major tech giants), has also benefited from companies investing heavily in AI computing infrastructure (such as Micron, Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR, and Broadcom Inc.) and power system suppliers (such as Constellation Energy). The so-called "Magnificent Seven," which account for about 35% of the S&P 500 index's weight, include: Apple Inc., Microsoft Corporation, Alphabet Inc. Class C, Tesla, Inc., NVIDIA Corporation, Amazon.com, Inc., and Meta Platforms, the parent company of Facebook. They are the core driving force behind the S&P 500 index repeatedly reaching new highs and are seen by top Wall Street investment institutions as the most capable combination of delivering substantial returns to investors in the largest technology revolution since the Internet era. A major concern -- a tweet by "Big Short" stirs market discussions on depreciation As the S&P 500 index's market value weight becomes increasingly concentrated on the aforementioned "seven major tech giants" (about 35%-40%) and the macro background of the "AI bubble narrative" pervades the market, concerns arising from a tweet by the "Big Short" regarding the depreciation dispute of American tech giants are worrying. Recently, the exaggerated "AI cyclic investment" led by OpenAI and unprecedented large-scale debt-fueled AI data center construction by tech giants like Oracle Corporation are jeopardizing their financial fundamentals, causing the market to increasingly worry about the "bursting of the AI bubble." As the market becomes more anxious about the panic surrounding the AI bubble storm that blew up in November, especially as 2026 approaches, there may even be doubts about the irrational AI bubble taking shape and coming closer to bursting. Several Wall Street giants, including Goldman Sachs Group, Inc., Bank of America Corp., Yardeni Research, and Morgan Stanley, have recently stated in their annual summaries and outlooks that the market's growing skepticism about the high valuations of tech stocks and the substantial AI investments being able to bring considerable returns is driving the market to focus more on traditional cyclical sectors like industrials and energy, rather than the high-valued tech giants like NVIDIA Corporation and Amazon.com, Inc., which are at the core of the "AI bubble." Yardeni Research founder Ed Yardeni even suggested for the first time since 2010 that investors "reduce" the seven major tech giants relative to the rest of the S&P 500 index. Take NVIDIA Corporation (NVDA.US) as an example -- this "AI chip overlord" with a market value exceeding $4 trillion ranks first on the global market value ranking and carries the title of the "most important stock on Earth." The company's stock price and market value movements are closely related to whether the logic of the "AI bull market narrative" can continue smoothly, and even to whether the global stock market bull market can continue. Renowned investment manager Michael Burry, known as the "Big Short," whose successful bet against the US housing market on the eve of the 2008 global financial crisis was depicted in the Hollywood movie "The Big Short," released a tweet showing that he believes that American tech giants such as Meta, Oracle Corporation, and even NVIDIA Corporation itself may have overestimated their profit capabilities by amortizing their semiconductor procurement/production costs over too many years, thus reducing their profit impact. The longer the depreciation period, the lower the annual depreciation expense and the less impact on profits. Burry believes that due to technological advances, the value of chips will decline faster than the numbers assumed by the company's depreciation arrangements. On the other hand, investors who are bullish on NVIDIA Corporation believe that extending the depreciation period merely reflects the actual useful life of these chips in terms of wear and tear. NVIDIA Corporation is not the only tech company being scrutinized for its depreciation practices that impact its "Big Short" profit statements. Burry also pointed his finger at cloud computing giant Oracle Corporation (ORCL.US) in his tweet. In addition, since 2020, as capital expenditures of global tech companies have soared, including Alphabet Inc. Class C parent company Alphabet (GOOGL.US), Amazon.com, Inc. (AMZN.US), Meta parent company Facebook (META.US), and Microsoft Corporation (MSFT.US), among other "seven major tech giants," have been extending the "assumed useful life" of certain major heavy assets. Moreover, there is evidence suggesting that the American old-school tech giant IBM (IBM.US), also known as the "Blue Giant," is participating in this "accounting game." According to the research firm Stock Analysis on Net (which mainly analyzes financial statements of large US companies), IBM's total depreciation expenses have decreased from $4.2 billion in 2020 to just $2.2 billion in 2024, while revenue has increased from $52.3 billion to $62.8 billion. This apparent mismatch does not reflect IBM's fundamental economic growth pace. However, asset disposal could also account for changes in IBM's revenue data, as well as adjustments to other depreciation policies; but it cannot be denied that extending the assumed useful life of assets may be a major contributing factor. The "storm" hidden in the tax return When it comes to the depreciation arrangements of tech companies, it is important to remember one thing: they have no real importance in actually creating shareholder value. Specifically, how the depreciation period is set in financial reports (depreciation plan) does not create or increase a company's real economic value. Shareholder value (intrinsic value) is determined by future free cash flows, not by the "allocation method" of accounting expenses. DCF/FCF valuation considers enterprise value as the present value of future cash flows. Depreciation in financial reports is a non-cash expense: it simply spreads past capital expenditures (buying equipment/building data centers) over future periods for accounting purposes. Lengthening the depreciation period simply makes "the depreciation expense calculated each year smaller, and the accounting profit larger," but it does not increase the free cash flows in the cash flow statement. It is important to remember that these accounting techniques are important for tax purposes, but the financial statements that these tech companies submit to the Internal Revenue Service (IRS) are completely different from the financial statements they provide to investors. Therefore, the main channel through which depreciation could really affect cash flow is through taxes (depreciation for tax deduction), but the financial statements submitted to the IRS operate on a different basis compared to the financial statements provided to investors. In a tech company's annual report, the depreciation expense is just an accounting entry in the statement. Lowering the depreciation expense by changing assumptions will not increase free cash flow, and therefore will not enhance the company's real economic value. As a result, stock prices usually do not experience significant increases or rapid declines due to these "cosmetic" actions, which is not surprising. So, these actions to "extend the depreciation period" are more about making earnings per share (EPS) look better, making the P/E denominator larger, and making valuation multiples seem less expensive; this is "superficial decoration" rather than a means of operational improvement to create cash flows. So why would they possibly manipulate the depreciation arrangements? In the 1980s and 1990s, this was easier to understand because CEO compensation was often tied to reported earnings per share (EPS). However, today, most public companies pay CEO compensation based on stock price/market value rather than simple EPS data. Perhaps these accounting frauds or reasonable techniques for embellishment are a way to lower seemingly "high" price-to-earnings multiples, making some investors believe that even after significant rises in stock price, the valuation remains low -- tech executives have had to repeatedly explain this in recent years. According to statistics from institutions, as of December 18, based on Wall Street analysts' consensus estimates, the average price-to-earnings ratio (P/E) for information technology stocks (IT stocks) currently stands at a high 36x, while the entire S&P 500 index is only at 25x. By raising the denominator, the P/E ratio can appear less high, making life a little easier for CEOs, CFOs, and investor relations officers at high-valued tech giants. However, the spread of these practices does not help investors trying to gauge whether the stock is priced reasonably, but rather misleads them into thinking that the company's P/E valuation is still appropriate. It is important to note that there is no evidence suggesting that today's large tech companies engage in fraudulent/accounting manipulative reporting practices. Overall, adjustments to the depreciation period are more like "financial statement cosmetic management/communication tools" used to serve valuation, manage market expectations, and the "profit quality" narrative led by CEOs (and some metrics), rather than creating cash flows. For example, depreciation affects core profit indicators such as EBIT, and many credit metrics (such as interest coverage ratios) use EBIT as the numerator to measure debt repayment capability; therefore, reducing depreciation and increasing EBIT may make these ratios look better. However, these tech stocks that account for a significant weight in the S&P 500 index and whose market movements have a huge impact on the global stock market are almost entirely priced based on "perfect profit expectations" in the market; thus, if the debate over depreciation heats up, investors may start reevaluating the actual profit levels of these companies. As valuations have reached historical highs, the AI bull market narrative has maintained investor bullish sentiment almost single-handedly, and the stock market has extremely high concentration, any factor perceived as "bad news" is nerve-wracking, and any negative controversy over depreciation could potentially trigger a painful "bear market edge" pullback.