Amazon and Microsoft's AI bet hit a wall: Wall Street's rare bear makes his move

Rothschild & Co. analyst Alex Haissl downgraded Amazon and Microsoft from Buy to Neutral, arguing that “cloud 1.0” economics no longer apply as GPU-driven AI infrastructure is far more capital‑intensive and unable to generate commensurate revenue (Rothschild cites roughly $40bn of GPU cost per GW versus $10bn of revenue), with longer depreciation cycles (5–6 years), weaker pricing power and a risk of overbuilding as capex accelerates into 2026. He cut Microsoft’s price target to $500 (leaving Amazon at $250) despite overwhelming Street bullishness, warning that hyperscalers can’t simply pass rising costs to customers without eroding demand or margins. While currently an outlier, the downgrade highlights a clear watch item for December earnings and early‑2026 results—if management commentary shows accelerating capex or deteriorating unit economics, broader sentiment and valuations for AI‑heavy names could quickly shift.

Analysis

Rothschild & Co. analyst Alex Haissl downgraded Amazon and Microsoft from Buy to Neutral, arguing the market still prices “cloud 1.0” economics while AI infrastructure is far more capital‑intensive; the note cites hyperscalers’ collective $349 billion pledge to AI infrastructure and a calculated mismatch of roughly $40 billion of GPU cost per gigawatt versus only $10 billion of revenue per gigawatt. Haissl cut Microsoft’s price target from $560 to $500 and left Amazon at $250, marking a rare reversal from an analyst who had been bullish since 2022.

The report highlights structural differences: longer assumed depreciation cycles of 5–6 years versus ~3 years in the prior cloud era, weaker pricing power because hyperscalers cannot easily pass GPU costs to customers, and a risk of overbuilding as capex accelerates into 2026. Market reaction has already compressed valuations—AI names endured a 5–6% pullback amid a $1.8 trillion decline in the Nasdaq 100 since late October—yet Wall Street consensus remains overwhelmingly bullish (Microsoft: 71 buys/2 holds; Amazon: 80 buys/5 holds).

The downgrade is currently an outlier but is actionable as a sentiment inflection indicator: the key near‑term catalysts are December earnings calls and Q1 2026 results for management commentary on AI unit economics and capex cadence. If guidance continues to accelerate without clear ROI milestones or if other major banks follow with downgrades, expect faster margin compression and a broader reassessment of valuations for AI‑heavy cloud names.

Key Decisions for Investors

• Consider trimming positions in Amazon and Microsoft toward neutral weight ahead of December earnings if you lack conviction in improving AI unit economics
• Avoid adding to long positions at current valuations until management provides concrete ROI milestones or a clear capex inflection point
• Use targeted hedges (put options or index hedges) to protect exposure to AI‑heavy names against downside risk from accelerating capex and margin compression
• Monitor December earnings and early‑2026 commentary closely; if other major sell‑side firms issue similar downgrades, reprice positions quickly to reflect a potential regime change in cloud economics