22.4% of Billionaire Bill Ackman’s Pershing Square Capital Management Is Invested in These 2 Artificial Intelligence (AI) Stocks

Microsoft reported fiscal Q4 2026 revenue growth of 18% to $90.0 billion, while cloud revenue rose 27% to $59.3 billion and cloud RPO surged 84% to $678 billion, signaling sustained AI-led demand. Amazon's Q2 net sales increased 20% to $200.6 billion, with AWS revenue up 37%—its fastest growth in 18 quarters—and advertising sales up 26% to $19.8 billion. Amazon raised projected FY2026 capex to $220 billion from $200 billion amid sold-out cloud capacity through end-2026 and a $496 billion backlog growing at a triple-digit rate.

Analysis

The relevant divergence is not AI demand but monetization efficiency. MSFT has the cleaner route to incremental margin expansion if proprietary inference silicon displaces third-party compute costs while Copilot and Azure attach rates rise; however, its valuation will remain sensitive to proof that AI revenue is recurring rather than consumption pulled forward by large enterprise commitments. AMZN has greater operating leverage because AWS growth, advertising, and retail fulfillment productivity can compound simultaneously, but the larger capital program makes free-cash-flow conversion—not revenue growth—the key multiple determinant.

Over the next 1-3 months, both stocks need hyperscaler capex to translate into disclosed utilization, backlog conversion, and stable gross margins. The second-order beneficiary is NVDA only if incremental capacity continues to require accelerated hardware at the same intensity; custom silicon adoption by MSFT and other clouds is a medium-term risk to NVDA's inference mix, even if training demand remains robust. Memory tightness is a more immediate pass-through risk for AMZN: capacity additions can be economically rational while still depressing near-term AWS and consolidated FCF.

Consensus likely underweights AMZN's diversification of AI returns: advertising and retail automation can fund cloud investment, reducing dependence on AWS margin alone. Conversely, consensus may be too willing to capitalize reported backlog at face value for both companies; long-duration commitments carry renewal, deployment, and pricing risk, particularly if enterprise AI pilots fail to produce measurable labor or revenue ROI. A broad rise in long-end yields or a capex-guidance increase without a corresponding utilization/FCF bridge would be the clearest catalyst for multiple compression.

Key Decisions for Investors

• Prefer long AMZN / short MSFT on a 3-6 month relative basis, sized modestly: AMZN has more unpriced cross-business earnings levers, while MSFT faces a higher burden to demonstrate AI margin durability. Reassess if AWS growth decelerates materially or AMZN's quarterly free cash flow fails to improve despite capacity deployment.
• Maintain a core MSFT long only on weakness following capex-driven selloffs, not after backlog headlines. Add if management quantifies AI gross-margin stabilization or Maia-driven unit-cost reductions; exit/add hedge if Azure growth slows while infrastructure capex continues to rise.
• Use NVDA as a hedge/watch against the long-cloud basket rather than a fresh directional add: sustained hyperscaler capacity commitments support it near term, but evidence of custom-chip inference migration or lower GPU content per workload is a 6-18 month relative-risk signal.
• Set an earnings alert for AMZN: a widening gap between AWS revenue growth and consolidated free-cash-flow growth is thesis-negative even if demand commentary remains strong. The market is likely to penalize another capex step-up absent explicit returns, utilization, or margin disclosure.