Billionaire Bill Ackman Has 10.5% of His $19.5 Billion Portfolio in Amazon. What's the Thesis Behind Pershing Square's Bet?

Pershing Square trimmed its Amazon stake but still holds nearly 8.6M shares (~$2B, 10.5% of its portfolio), arguing the AI-driven capex build is investment in growth rather than a cost overhang. Ackman expects Amazon capex to rise to about $220B in 2026 (from ~ $132B in 2025), supporting AWS data-center expansion as AWS revenue grows from ~20% growth in 2024-25 to over 30% so far in 2026. He also projects Amazon can compound total earnings by at least 20% and cites strong retail momentum with ~15% unit volume growth in Q1 (highest since 2021).

Analysis

The market is likely to overfocus on the capex line item and underfocus on whether incremental spend is still earning an above-cost-of-capital return. In the next 1-2 quarters, that usually means the stocks that sell the picks-and-shovels of AI infrastructure can outperform the platform name itself, because hardware orders, power, and data-center buildout monetize immediately while the revenue lift lags. For AMZN, the key debate is not spending size but utilization: if the new capacity is filling quickly, the stock can re-rate as a durable compounder; if not, it starts to look like a high-growth business with utility-like returns.

Second-order pressure falls on cloud peers and adjacent suppliers forced to keep up with the pace of investment. That dynamic can keep industry margins competitive, but it also raises the bar for anyone trying to defend share without matching spend. Retail strength matters less for headline growth than for financing flexibility: it gives Amazon internal funding to sustain the buildout and continue taking share from omnichannel incumbents and marketplace sellers, which should show up first in category economics, not just revenue.

The contrarian risk is timing: spending can outrun monetization for several quarters, especially if rates stay elevated and investors demand faster free-cash-flow conversion. The thesis is falsified if AWS growth rolls back toward the low-20s while capex stays elevated, or if management has to keep raising investment without visible operating leverage. In that case, AMZN deserves a lower multiple as a capital-intensive infrastructure asset rather than a premium software-adjacent platform.

Key Decisions for Investors

• Buy AMZN on 3-5% pullbacks and express it with 6-9 month call spreads; best risk/reward is if the market sells capex before the next AWS print confirms utilization. Falsify the trade if AWS growth decelerates back below ~25% or FCF margin keeps compressing after the spend ramp.
• Run a tactical long NVDA / short AMZN pair for 1-2 quarters to capture the near-term beneficiary of AI infrastructure spending versus the longer monetization lag at the platform level. Keep sizing modest; the pair loses if AMZN proves the capex is already translating into durable margin expansion.
• Add exposure to data-center and power-infrastructure beneficiaries on weakness rather than chasing the headline name; the cleaner way to express the buildout is via suppliers and colocation assets, which monetize faster than AMZN's depreciation cycle.