Rothschild downgrades two ‘Magnificent Seven’ giants, cites lower conviction on AI trade

Rothschild downgraded Microsoft and Amazon to neutral from buy, with analyst Alex Haissl urging caution on the generative-AI trade because GPU-led Gen‑AI deployments are materially more capital intensive than cloud 1.0. Haissl estimates GPU deployments require roughly six times more capex to generate equivalent value, prompting a cut to Microsoft's price target to $500 from $560 (implying ~1% downside; MSFT is up ~20% YTD) and flagging potential Office 365 leakage and weaker long‑term cash flow. Amazon's $250 target was left unchanged (about 7% upside; AMZN is up ~6% YTD) as Rothschild sees limited upside for AWS despite recent reacceleration and warns Gen‑AI is dilutive to returns even where AWS captures more stack value, indicating higher capital intensity and downward pressure on hyperscaler economics.

Analysis

Rothschild Asset Management downgraded Microsoft and Amazon to neutral from buy, citing a need for greater caution on the generative-AI investment thesis; Alex Haissl cut Microsoft’s price target to $500 from $560 (implying roughly 1% downside from Monday’s close) while leaving Amazon’s $250 target unchanged (about 7% upside). The firm quantifies a material change in the cost profile of AI-enabled cloud services, estimating GPU-led Gen‑AI deployments require roughly six times more capex to generate the same cloud 1.0 value, and notes sentiment has turned moderately negative for these hyperscalers.

Haissl argues Microsoft faces revenue leakage in Office 365 as third‑party models like OpenAI and Anthropic shift value away from Microsoft, and that higher capex and lower Gen‑AI value creation will pressure long‑term cash flow and earnings — the explicit drivers behind the price‑target cut. For Amazon, Rothschild acknowledges AWS has reaccelerated but sees limited upside relative to consensus and warns Gen‑AI is dilutive to returns even where AWS captures more of the stack.

The practical implication is that current market prices may be embedding cloud 1.0–era returns despite a structural increase in capital intensity; investors should therefore focus on capex trajectory, incremental margin on Gen‑AI revenues, and indicators of product‑level value capture rather than headline growth metrics when reassessing hyperscaler exposures.

Key Decisions for Investors

• Trim or reweight Microsoft exposure given the reduced price target, higher capex risk and potential Office 365 revenue leakage,
• Maintain a cautious, watchful position on Amazon—its $250 target implies limited upside and AWS fundamentals look constrained despite recent reacceleration,
• Monitor company disclosures for capex guidance, GPU utilization metrics, Gen‑AI revenue margins, and third‑party model partnerships as triggers to increase or reduce positions,
• Use downside protection or position sizing (options hedges or size reduction) on hyperscaler exposure until there is clearer evidence of sustainable Gen‑AI economics and improved cash flow conversion