Amazon’s $200 billion AI bet tests investors’ nerves as earnings rise below expectations

Amazon’s latest quarter delivered what many portfolio managers say they want from megacap tech: robust top-line growth, a re-accelerating cloud franchise and a detailed plan for artificial-intelligence infrastructure. The market’s reaction on Thursday suggested something more complicated.

The company reported fourth-quarter 2025 revenue of $213.4 billion, up 14% from a year earlier and slightly ahead of Wall Street expectations around $211 billion. Net income rose to $21.2 billion, or $1.95 a share, modestly above last year’s $1.86 but just shy of consensus forecasts near $1.97.  

The Wall Street Journal noted that the miss on earnings, combined with a strikingly aggressive investment plan, sent the shares down as much as 8–10% in after-hours trading, even as Amazon showcased its strongest growth in key businesses since before the Fed’s tightening cycle began.

At the center of the reaction is a number that is now likely to loom large for institutional investors, highlighted in the earnings release by Amazon president and CEO Andy Jassy: “With such strong demand for our existing offerings and seminal opportunities like AI, chips, robotics, and low earth orbit satellites, we expect to invest about $200 billion in capital expenditures across Amazon in 2026, and anticipate strong long-term return on invested capital.”

That $200 billion would represent a sharp jump from an estimated $125 billion in capex this year, but Jassy’s statement sought to justify the spend.   

“AWS growing 24% (our fastest growth in 13 quarters), Advertising growing 22%, Stores growing briskly across North America and International, our chips business growing triple digit percentages year-over-year—this growth is happening because we’re continuing to innovate at a rapid rate, and identify and knock down customer problems,” he said.  

Underneath the headline spending figure, the operating story was mostly solid. Amazon Web Services posted fourth-quarter revenue of about $35.6 billion, up 24% year over year, the fastest pace in 13 quarters and ahead of estimates around $35 billion.  AWS operating income climbed to $12.5 billion from $10.6 billion a year earlier, underscoring that cloud remains the profit engine funding Amazon’s broader bets.  

At the consolidated level, operating income rose to $25.0 billion from $21.2 billion despite more than $2.4 billion in special charges tied to tax settlements in Europe, severance costs and impairments in the physical retail footprint. Excluding those items, operating profit would have reached $27.4 billion.  

For the full year, revenue climbed 12% to $716.9 billion, while net income jumped to $77.7 billion from $59.2 billion. Operating cash flow increased 20% to $139.5 billion. Free cash flow, however, dropped sharply to $11.2 billion from $38.2 billion, almost entirely because of a roughly $50.7 billion increase in spending on property and equipment, “primarily” for AI-related investments, according to the company.  

That divergence between earnings strength and free cash flow deterioration sits at the heart of the current AI-investment debate. For allocators who have used free cash flow yield as a key discipline in sizing exposure to the “Magnificent Seven,” Amazon is now asking for patience.

With investors already concerned about the magnitude of investment in AI, the risk is not just that one company overspends, but that the entire ecosystem invests ahead of actual monetization, pressuring returns across the group.

Amazon’s commentary makes clear that its AI ambitions span not only cloud compute and custom silicon, but also robotics and satellite connectivity.  

For allocators, that breadth can look like diversification within a single issuer—but also like concentration risk at the portfolio level. Many balanced and growth mandates already have high weights to a handful of AI-centric platforms. Amazon’s outsize capex plan intensifies the question of how much “AI infrastructure risk” a diversified portfolio should carry relative to benchmarks.

For longer-horizon investors, the question is less whether Amazon can grow into its AI footprint, and more about timing and execution risk. If AWS growth sustains its recent pace, advertising continues to compound at double digits, and retail margins hold their gains from post-pandemic restructuring, the 2026–27 payoff from today’s capex could be substantial. If enterprise AI adoption proves slower or more price-competitive than hoped, the same spending could compress returns for years.