Analysts Scrutinize Big Tech's 'Capex Illusion'
Analysts are warning of a "capex illusion" regarding the massive infrastructure spending by Big Tech companies on cloud and data centers. An analysis suggests that headline capital expenditure figures can obscure a company's true cash flow and profitability, potentially making some tech giants appear less cash-generative than they seem.
- The surge in capital expenditures is primarily fueled by an "AI arms race," with combined 2026 spending forecasts for Alphabet, Amazon, Meta, and Microsoft reaching approximately $650 billion. Amazon alone has a spending forecast of $200 billion. - A key part of the "illusion" involves accounting changes, specifically extending the depreciable lifespan of servers. For instance, by increasing the "useful life" of its servers to six years, Google saved $3.9 billion in depreciation in a single year, while a similar move by Amazon saved $900 million in one quarter. - Meta has been particularly aggressive with this strategy, extending the useful life of its servers three times in three years, from 4 years to 5.5 years, saving an estimated $2.9 billion in 2025 alone. This practice has raised concerns about deteriorating earnings quality. - Despite the massive cash outlays, some analysts argue that focusing on operating cash flow still shows these companies to be highly profitable. From 2013 to 2022, Big Tech collectively generated $2.3 trillion in operating cash flow. However, others point out that free cash flow, a measure that subtracts capital expenditures, is expected to "crater" due to the spending surge. - This level of concentrated spending is beginning to have broader economic effects, including straining energy supplies and raising concerns about distorting economic data due to the sheer scale of investment by a few wealthy companies. - Investor sentiment is showing signs of strain, with markets beginning to question the long-term return on these massive AI investments. The scale of spending has drawn comparisons to the fiber-optic build-out of the late 1990s, which preceded a market downturn. - While headline profits may appear strong, some analyses suggest that companies like Amazon and Google are spending more on capital expenditures for AI than they are making in net income. For example, Amazon's free cash flow was projected to turn negative in 2026 due to its $200 billion capex plan.
