Amazon reported third-quarter results on October 29 that split neatly in two: its cloud business grew faster than it has in nearly five years, while the rest of the report disappointed investors enough to send shares sharply lower in after-hours trading. Amazon Web Services sales rose 39 percent to $46.0 billion, yet total net sales of $201.9 billion and earnings of $1.89 per share both missed expectations. Free cash flow swung to a $42.4 billion outflow over the trailing twelve months, and Amazon held its 2026 capital spending forecast at about $220 billion, with its chief executive telling analysts that spending will rise meaningfully further next year.
AWS does the heavy lifting
AWS is now a $184 billion annualized revenue business after sales grew 39.3 percent in the quarter ended September 30, the unit’s fastest pace in nearly five years. Operating income at the cloud unit rose to $17.8 billion from $11.4 billion a year earlier, and AWS supplied well over half of Amazon’s total operating income for the quarter.
Demand, Amazon said, is not the constraint: chief executive Andy Jassy had already told investors that AWS capacity was largely sold out through 2027, a position this site covered when AWS capacity sold out through 2027. The sold-out order book explains why Amazon keeps building even as its cash flow turns negative.
The AI work inside that growth is visible in Amazon’s own operating details. Spending by Fortune 500 companies on Bedrock, AWS’s inference service, rose nearly sevenfold from a year earlier, while total Bedrock token volume grew nearly seventeenfold in the same period.
AgentCore, the company’s managed service for building agents, is used by tens of thousands of customers, who deployed twice as many agents in the third quarter as in the previous three quarters combined. New AWS agreements signed since the last report include Meta, NVIDIA, OpenAI, Salesforce and Oracle.
The miss behind the beat
Outside the cloud, the quarter was softer than the headline operating profit suggests. Total operating income rose to $28.1 billion from $17.4 billion, at a 13.9 percent margin that Amazon called its highest ever, but net income fell to $20.7 billion from $21.2 billion.
Amazon attributed that profit fall mainly to a $9.5 billion pre-tax gain on its Anthropic investment that had lifted the year-earlier quarter, so it is an accounting comparison rather than a collapse in the underlying business. North America sales rose 6 percent to $112.9 billion, while international sales rose 5 percent to $43.1 billion.
Advertising was the other disappointment, reaching $19.6 billion with growth slowing sharply from a year earlier. Amazon blamed the timing of Prime Day, which moved into the second quarter this year after sitting entirely in the third quarter of 2025, saying the shift alone cost about four percentage points of total sales growth.
On the consumer side, the company’s agentic shopping push continued to scale: Alexa for Shopping was used by more than 400 million customers over the past twelve months, and Amazon added more than 100 new shopping capabilities, as covered in our report on Alexa’s 100 new shopping actions.
Guidance for the holiday quarter did little to calm investors. Amazon forecast fourth-quarter net sales of $240.0 billion to $247.0 billion, with operating income of $28.5 billion to $33.5 billion. The midpoint of the sales range sits below the $244.51 billion consensus compiled by LSEG, as reported by Reuters.
The $220 billion question
The number that did the most damage was free cash flow. Amazon reported a trailing twelve-month outflow of $42.4 billion, against an inflow of $14.8 billion a year earlier, driven primarily by a $94.8 billion year-over-year increase in purchases of property and equipment, which Amazon said primarily reflects investment in artificial intelligence.
Amazon is not backing away from that spending. It reaffirmed about $220 billion of cash capital expenditure for 2026, and Jassy told analysts on the earnings call that capital spending will be meaningfully more next year, without giving a figure. He also projected AWS revenue growth of more than 40 percent in both 2027 and 2028.
Microsoft, by contrast, lowered its own capital spending forecast to $170 billion from $175 billion, as covered in our report on Microsoft’s cloud beat, showing that not every hyperscaler is accelerating at the same pace. Amazon’s bet is the larger and the less hedged of the two.
Why It Matters
This quarter turns Amazon’s AI bet into a cash-flow question that investors can now measure. The cloud business is plainly working: growth is accelerating, operating income is rising, and customers from Meta to OpenAI are signing new AWS agreements. The open question is timing, not demand.
The bill for building the infrastructure behind that growth has now swallowed a full year of free cash flow, and Amazon is promising to spend more, not less, in 2027. The debate is whether that demand arrives fast enough to pay for the data centers.
Amazon’s defense rests on a number that cuts the other way. Operating cash flow rose 29 percent to $168.3 billion over the trailing twelve months, so the business is generating cash at scale; it is the investment program, not the operations, that produces the outflow.
The same cash-burn pattern hit Alphabet this quarter, whose cloud business more than doubled while free cash flow went negative, as covered in our report on Alphabet’s cloud quarter. If Jassy’s growth forecast holds, the spending looks like investment ahead of demand; if growth cools, the outflow will look like an early warning.
Background
Amazon launched AWS in 2006 and built it into the world’s largest cloud provider. The third quarter shows both sides of the current buildout bet in a single report: a cloud business growing at its fastest pace in nearly five years, and a balance sheet absorbing the cost of keeping it that way.
Source: Amazon’s earnings release and Reuters report
0 Comments