# Investors Face Risks as AI Infrastructure Obsolescence Accelerates Rapid technological shifts threaten the long-term financial returns of massive AI infrastructure investments. - **Source:** [Business Insider](https://www.businessinsider.com) - **Published:** 2026-11-11 - **Topic:** [Finance](https://brifing.si/c/finance) - **Original article:** - **On Brifing:** https://brifing.si/s/ccdf8161-3b08-4bb5-bc68-71648f5ca80e ## Rapid Obsolescence Risks Capital The AI boom is driving massive capital expenditure across the technology sector, yet rapid innovation threatens to render infrastructure obsolete before it becomes profitable. Experts warn that the pace of computing hardware advancement is significantly faster than the multi-year timeline required to construct and fully operationalize large-scale data centers. ## Accounting Adjustments Signal Trouble Amazon recently shortened the estimated useful life of its servers and networking equipment from six years to five years. This accounting change, implemented in 2025, increased depreciation expenses by $1.4 billion and reduced net income by $1 billion, highlighting the financial burden of keeping pace with evolving AI hardware requirements. ## Massive Capital Expenditure Targets Major tech giants including Amazon, Microsoft, Alphabet, and Meta are planning a combined capital expenditure of up to $725 billion for 2026. This aggressive spending strategy faces scrutiny as investors worry about the long-term sustainability of these investments, especially amid rising interest rates and uncertain future demand for data capacity. ## Infrastructure Versus Computing Hardware There is a fundamental mismatch between the economic lives of different assets. While physical buildings and power connections can support multiple generations of technology, the computing hardware housed within them loses commercial value much faster. Investors must now determine how frequently this expensive hardware will require replacement to remain competitive. ## Uncertainty Over Future Scale Industry leaders remain divided on the total required scale of the AI buildout. Goldman Sachs CEO David Solomon noted that while current demand for computing power exceeds supply, it remains unclear whether the industry requires a $5 trillion or $10 trillion investment, complicating long-term financial planning for infrastructure-focused institutional investors. ## Historical Parallels to Fiber-Optics Analysts draw parallels between current AI infrastructure spending and the fiber-optic expansion of the late 1990s. While those networks became essential to the internet economy, many companies that financed their construction struggled financially. The economic value may ultimately shift toward businesses that utilize affordable computing power rather than those building it. ## Efficiency as a Double-Edged Sword More efficient AI models could potentially fuel demand by lowering costs and encouraging wider adoption. However, for infrastructure investors, returns depend on whether this growth in paid usage can compensate for falling rental prices and the high costs associated with electricity, financing, and the frequent replacement of aging computing hardware. ## Strategic Shift for Investors Investors are becoming increasingly discerning, moving away from treating AI as a single opportunity. Success now depends on securing power supplies, creditworthy customers, and robust contracts. The greatest vulnerability lies in borrowing heavily to buy hardware while relying on a small customer base and assuming current high rental prices will persist. ## Sources covering this story - https://www.businessinsider.com/ai-boom-investment-returns-risks-infrastructure-technology-outpacing-buildout-2026-11 — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._