Amazon’s Cloud Is Booming. Where Is the Cash Going?

Amazon’s cloud business is booming. Its cash-flow statement tells a more complicated story. The company is selling more cloud services than ever—but building the computing capacity behind AI is absorbing extraordinary amounts of cash. The question is not whether customers want AI. It is how long the investment will take to pay for itself.

WHAT MATTERS

What changed? Amazon reported $46.0 billion in AWS sales in the third quarter of 2026, up 39% year over year. Its companywide free cash flow, measured across the 12 months ended September 30, swung from positive $14.8 billion a year earlier to negative $42.4 billion.

Why does money care? Revenue and accounting profit can grow while cash available after infrastructure spending declines. Investors ultimately need Amazon’s enormous investment to produce attractive returns—not just bigger data centers.

Who is affected? Shareholders, cloud and AI customers, smaller businesses buying AI tools, suppliers, workers and communities hosting energy-intensive infrastructure.

What happens next? Watch AWS sales growth, operating margins, infrastructure spending and the cash generated after those investments—not just announcements about new AI products.

The MNS takeaway: AI demand is real. So is the bill for serving it. Growth without a credible path to cash generation remains an open financial question.

WHAT HAPPENED

In results released October 29, 2026, Amazon reported $201.9 billion in quarterly net sales and $28.1 billion in operating income for the three months ended September 30. Its operating margin reached 13.9%, the highest in the company’s history.

AWS was the standout business: $46.0 billion in revenue, up approximately 39%, and $17.8 billion in operating income. Companies are paying Amazon for cloud capacity that supports everything from ordinary websites and software to advanced AI workloads.

But the cash picture spans a different period. Over the trailing 12 months, Amazon generated $168.3 billion in cash from operations and spent approximately $210.7 billion on property and equipment, net of specified proceeds and incentives. That left negative $42.4 billion in free cash flow under Amazon’s disclosed definition. A year earlier, free cash flow had been positive $14.8 billion.

Amazon said the change was driven primarily by rising investment in artificial intelligence. The company reported a $94.8 billion year-over-year increase in the relevant property-and-equipment cash spending measure. This is not a claim that AWS itself lost $42.4 billion, nor that the company recorded a $42.4 billion accounting loss.

WHY IT MATTERS

The AI industry has entered a phase in which a useful product also requires an expensive physical network: specialized chips, data centers, land, power, cooling, fibre connections and teams to keep everything running. Much of that infrastructure must be paid for before its full economic return is known.

AWS’s revenue growth is evidence of real demand. Yet a dollar of revenue is not a dollar of free cash flow. Hardware can become obsolete; electricity and maintenance are recurring costs; and customers may demand lower prices as competition expands. A successful strategy must connect usage → revenue → operating profit → cash after reinvestment.

There is an important balance: capital spending can temporarily depress free cash flow while creating productive assets for future years. Negative free cash flow alone is not evidence of distress. The risk is that spending stays high while returns disappoint—or that the cost of keeping up with competitors rises faster than the value customers are willing to pay.

MONEY CONSEQUENCE

• For shareholders: compare Amazon’s free-cash-flow trend with AWS revenue growth and operating income over several quarters. A strong margin today does not guarantee an attractive return on the next data center.
• For businesses buying AI: subscription or API fees are only part of the cost. Add integration, training, staff review, data controls, downtime and the cost of switching providers.
• For workers: new tools can change job tasks and hiring priorities, but a large capital-spending figure does not establish a specific number of jobs lost or created.
• For households and communities: data centers may bring investment and employment, while putting pressure on local electricity and infrastructure. Actual bill impacts depend on the local grid, regulations and contracts—not on a company’s global spending total.

A practical example: if a company introduces an AI assistant, the real return is not how impressive the demo looks. It is whether verified time savings or additional revenue exceed software fees, setup costs and the human work still needed to check the output.

CONTEXT

Amazon is not alone. Alphabet and Microsoft are also investing heavily in AI capacity, with very different cash-flow profiles. Comparing them requires care: quarterly capital spending, trailing-12-month free cash flow, operating income and long-term commitments are different measures. A fair comparison uses consistent definitions and reporting periods.

For investors who follow AI stocks, the central question is shifting from Who can build the most capacity? to Who can turn that capacity into dependable cash? Read our comparison of Alphabet and Microsoft and The AI Boom Has a Balance Sheet for the wider financial picture.

Amazon’s management expects strong demand for cloud services and has outlined ambitious growth forecasts. Those are company expectations, not guaranteed outcomes. The underlying financial test will be visible in subsequent filings, not in a prediction made after one earnings call.

WHAT TO WATCH NEXT

• AWS revenue and operating margin: does customer demand keep growing without sacrificing profitability?
• Free cash flow across consistent periods: is cash after infrastructure investment recovering as new capacity earns revenue?
• Capital-spending guidance: how much more will Amazon commit, and what is the expected useful life of the equipment?
• AI unit economics: are workloads becoming cheaper to run, or does demand require even more costly infrastructure?
• Customer value: do businesses report measurable returns from AI adoption, rather than buying capacity simply to avoid falling behind?

The bottom line: Amazon has demonstrated strong cloud demand and substantial operating profits. It has not yet answered the full question of how much cash the next stage of AI expansion will produce after its costs. That is the number worth following.

SOURCES / CONTEXT

• Amazon Investor Relations — Third Quarter 2026 Results, October 29 (primary source for revenue, segment operating income and trailing-12-month free cash flow).
• SEC — Amazon Q3 2026 earnings exhibit (financial tables and reporting periods).
• Reuters — market reaction and Amazon’s AI investment outlook, October 29 (independent reporting).

Source note: AWS revenue and profit refer to the third quarter; free cash flow refers to the trailing 12 months ended September 30, 2026. They must not be presented as if they cover the same period. Money no Sofá | Updated October 31, 2026. Not investment advice.

ASK NOVA

Ask NOVA what this could mean for your money: “How can I calculate whether paying for AI is saving my business more than it costs?”

Best Money Tool: an AI Payback Calculator. Enter setup cost, monthly subscriptions, verified hours saved, realistic labour value, additional revenue and oversight costs. Compare a base case with a downside case before making a commitment.