Google, Meta and Amazon are facing an expensive AI risk that nobody is paying attention to

Bank of America warns that Big Tech firms like Google, Meta, and Amazon face underestimated risks from rapidly accelerating depreciation costs associated with their substantial AI infrastructure investments. Despite an expected $274 billion in collective AI capex this year, analyst Justin Post projects depreciation and amortization expenses to grow significantly faster (33% in 2026, 30% in 2027) than revenue (13% and 12%), threatening profit margins. This discrepancy, driven by potentially shorter useful lives for AI assets and the risk of overbuilding, implies current valuations may be inflated and necessitates accelerated AI monetization to mitigate financial impact.

Analysis

A Bank of America analysis highlights a significant, underappreciated risk to Alphabet, Meta, and Amazon, where escalating depreciation from massive AI capital expenditures threatens to compress future margins. The firms are projected to collectively spend $274 billion on capex by year-end, but BofA estimates their combined depreciation and amortization expenses will grow at 33% in 2026 and 30% in 2027, far outpacing projected revenue growth of 13% and 12% respectively. This divergence suggests Wall Street consensus is underestimating future costs, with BofA's 2027 depreciation forecasts exceeding Street estimates by $7 billion for Alphabet, $5.9 billion for Amazon, and $3.5 billion for Meta. The core drivers of this risk include the potentially shorter useful life of AI assets like GPUs due to rapid innovation, a reversal of the previous trend of extending asset lifespans (as seen with Amazon's recent revision from six to five years for some equipment), and the risk of overbuilding leading to commoditization. Consequently, current profitability estimates may be inflated, posing a risk to valuations if AI monetization fails to accelerate and cover these rapidly rising non-cash expenses.

Key Decisions for Investors

• Investors should revise their earnings models for GOOGL, META, and AMZN to stress-test the impact of higher depreciation expenses, aligning with analyst projections of 30%+ growth, to gauge the potential for significant margin compression.
• Closely monitor company-specific AI monetization progress, as the widening gap between capex-driven depreciation growth and revenue growth makes accelerated monetization a critical catalyst for sustaining current valuations.
• Pay special attention to any future downward revisions in the stated 'useful life' of server and networking equipment in financial disclosures, as this would serve as a direct confirmation of the accelerated depreciation thesis.
• Be cautious of current valuations for these hyperscalers, as they may not fully price in the margin compression from accelerating depreciation, creating a risk of a valuation de-rating if consensus estimates adjust towards Bank of America's more pessimistic outlook.