Michael Burry has questioned whether Microsoft’s AI spending is driven by necessity, suggesting the company may be using its financial figures as leverage, according to Seeking Alpha. He argues Big Tech is spending wildly because it assumes it will be granted an oligopoly too big to fail. Michael Burry reportedly owns Microsoft Corporation (NASDAQ:MSFT) shares and long-dated call options, so this is skepticism about spending, not a short thesis.
Separately, the company made an accounting change at the start of fiscal 2027: it now assumes its data centers and office buildings last 25 years instead of 15. It’s worth examining because it changes how that spending is reported. Microsoft’s AI spending is under scrutiny, but its cybersecurity business offers another investment angle. See where it stands in our list of the 10 Best Cybersecurity Stocks to Buy According to Analysts.

Why I Think Burry Has a Point
Microsoft counts finance leases in its capex but leaves out operating leases. One key test for a finance lease is whether it covers most of a building’s useful life. With buildings now expected to last 25 years, more leases fall short of that bar and become operating leases. As a result, Microsoft’s calendar 2026 capex forecast dropped from about $190 billion to about $175 billion. CFO Amy Hood said that outside of this change, the company’s spending plans were unchanged. So the $15 billion drop came from accounting. Microsoft has also signed a huge number of new data center leases. Leases that haven’t started yet jumped from $196.6 billion to $329.1 billion in just one quarter. Under the revised lease classifications, more future data center leases will be treated as operating leases rather than included in Microsoft’s reported capex. With the company’s future data center commitments rising despite a lower capex forecast, investors may want to see which other growth stocks billionaires are backing, including MSFT.
The Valuation Isn’t Quite What It Seems
On paper, Microsoft looks reasonably priced. The forward non-GAAP P/E of 26.80x sits about 12% below its 5-year average of 30.51x. Analysts expect EPS to grow 14% this fiscal year, then about 20% and 22% over the next two. The same accounting change also affects earnings. The issue is that the same accounting change affects both earnings and how spending is reported. For a new building, stretching its assumed life from 15 to 25 years cuts yearly depreciation by about 40%. Microsoft says the boost to fiscal 2027 operating income will be minimal, but as more buildings come online, earnings will look better than they would have under the old rules. Meanwhile, the $329.1 billion in leases yet to begin won’t appear on the balance sheet until they start, so investors have to track them separately. So earnings could look stronger, and obligations smaller, than they really are. That’s why I don’t think the stock is quite as cheap as its P/E makes it look.
Hedge fund interest cooled slightly. Funds holding Microsoft fell from 282 in Q1 to 273 in Q2. The value of their stakes, though, rose from $63.6 billion to $66.5 billion.
Microsoft did disclose this change on its earnings call. Still, Burry is right that the headline capex number no longer tells the full story, and Microsoft’s valuation may be more demanding than its reported earnings multiple suggests. I’d track total lease commitments alongside capex to see what Microsoft is really spending,
READ NEXT: Wall Street’s Biggest Micron Bull Just Raised Its Target Again. It Now Sees the Stock Nearly Tripling And Nvidia (NVDA) vs. Broadcom (AVGO): Which AI Stock Is a Better Buy After the OpenAI Scare?




