Billionaire Bill Ackman Just Sold Amazon and Bought Microsoft Stock. Which Cloud Computing Behemoth Is the Better Buy Today?

Pershing Square trimmed its Amazon stake by over 25% in Q2 while increasing Microsoft by nearly 10%, but the article argues Amazon is the better buy now. Since Q2, Microsoft is up ~30% and Amazon ~10% after strong earnings; Amazon’s Q2 operating income rose 43% (AWS up 64% YoY) as Amazon plans ~$220B of data center capex in 2026, supporting faster growth despite a higher operating valuation premium on paper.

Analysis

The real signal is not “which cloud name is cheaper,” but which management team can convert hyperscale capex into durable incremental margin. AMZN has more operating leverage because AWS is still scaling from a smaller base; that means each additional point of utilization can matter more to earnings power than it does for MSFT, where the market already prices in a more mature scarcity premium. Over the next 1-3 months, the key catalyst is whether capex commentary keeps rising faster than depreciation and power availability, because that is what determines whether this becomes a margin story or just a race to stand still.

Second-order beneficiaries are the picks-and-shovels names: semis, networking, and power infrastructure should keep seeing demand even if hyperscaler shares consolidate. NVDA remains the cleanest indirect winner if the build-out stays compute-heavy, but the market is already leaning into that trade, so upside is more about capex durability than surprise. The biggest loser in a relative sense may be the “good enough” cloud incumbent that keeps spending but fails to show a matching step-up in operating income; that is where multiple compression risk lives.

Contrarian view: the consensus is underpricing the drag from depreciation, energy, and lower initial utilization on these new data centers. If AI workloads do not fill capacity fast enough, reported revenue can lag headline spend for several quarters, which would hit free cash flow before it shows up in revenue trends. Falsifier for the AMZN-over-MSFT thesis: any deceleration in AWS growth, a capex cut or freeze, or MSFT re-accelerating Azure growth without a proportional margin hit.

Key Decisions for Investors

• Bias long AMZN vs MSFT on a 3-6 month horizon, but only on a pullback: target a 5-7% retracement in the pair before entering; thesis breaks if AWS growth slows materially or if MSFT turns capex into faster Azure monetization.
• Use NVDA as the cleaner second-order beneficiary of hyperscaler capex, but prefer buying weakness after any broad AI selloff rather than chasing momentum; watch for hyperscaler capex guide revisions as the trigger.
• Do not overread the 13F rotation as a fresh signal; treat PSHZF positioning as stale and avoid front-running a move that the market may already have arbitraged via the post-earnings rally.
• If expressing the view with options, favor a risk-defined AMZN call spread over outright stock to capture 3-6 month operating-leverage upside while limiting damage if depreciation and power costs compress margins.