Amazon’s latest results show that the artificial intelligence boom is becoming an infrastructure and cash-flow test, not merely a technology opportunity. The company’s cloud business is growing rapidly, but the cost of building the data centres, power systems and technology infrastructure needed to support that growth is absorbing much of the cash generated by operations.
Amazon shares fell as much as 7% in after-hours trading after the company reported third-quarter results that missed Wall Street expectations and offered a cautious outlook for the holiday quarter. The stock had already declined 4.3% during regular trading to close at $243.79. The immediate market reaction focused on weaker earnings, higher capital spending and concerns about how quickly artificial intelligence investments will translate into stronger cash generation.
The urban significance lies in the physical system behind the numbers. Amazon has forecast capital expenditure of $220 billion for 2026, largely to build AI data centres and expand its technology infrastructure. Chief executive Andy Jassy told analysts that spending would be “meaningfully more” next year, although he did not provide a specific figure. That scale of investment places data centres at the centre of a wider infrastructure question: how cities and regions will accommodate large digital facilities that require land, electricity, water, networks and workers.
Amazon Web Services recorded its fastest growth in nearly five years. Revenue rose 39% year-on-year to $46 billion in the quarter ended September 30, accelerating from 37% growth in the previous quarter and exceeding analysts’ estimate of $45.61 billion. Jassy said AWS could become a business generating more than $1 trillion in annual revenue within five to 10 years and described customer demand for its AI infrastructure as strong.
That growth is being pursued through a capital-intensive model. Amazon’s spending on capital projects during the quarter reached $77.26 billion, about $20 billion above estimates compiled by LSEG. The company reported a trailing 12-month cash burn of $42.4 billion in the September quarter. Portfolio manager Dave Wagner of Aptus Capital Advisors said operations generated substantial cash, but capital expenditure “absolutely swallowed it”.
The figures illustrate the physical cost of digital expansion. A data centre is not simply a warehouse for servers. It depends on substantial construction, electrical systems, cooling equipment, fibre connectivity and continuing energy supply. The supplied results do not establish the location, capacity or resource requirements of Amazon’s planned facilities, but they show that the company’s AI strategy is inseparable from a major expansion of physical technology infrastructure.
This creates a different kind of pressure from the one visible in quarterly earnings. Investors are asking whether the infrastructure can produce adequate financial returns. Communities near data centres face a parallel question about how the facilities affect local development, employment, education, water use and energy systems. Amazon acknowledged that community relationship by saying earlier this month that it would invest more than $1 billion over five years in communities where it operates data centres.
The company said that programme would include funding for education, job training, water conservation and energy-related initiatives. Those areas indicate the range of local issues connected to data-centre expansion. Construction and operations can generate economic activity and employment, while water and energy requirements bring the facilities into the domain of public infrastructure and local resource management. The input does not provide details of the programme’s individual projects, funding allocation or measurable targets, so its practical effect cannot yet be assessed from the available evidence.
Amazon’s results also show how the data-centre question is spreading across the technology industry. Alphabet shares fell 12% on Thursday after investors reacted to its latest results and rising expenditure on property and equipment. Microsoft lowered its capital-expenditure outlook to $170 billion from $175 billion. The comparison suggests that investor scrutiny is not limited to one company; it is being applied to the broader race to build data centres and deploy AI chips.
For cities, that race changes the relationship between technology companies and infrastructure providers. Digital demand is often experienced by users as an online service, but the service depends on large fixed assets in specific locations. The source material does not identify the authorities responsible for approving Amazon’s facilities or explain how their electricity, water and land requirements are being governed. It does, however, make clear that the company’s expansion is sufficiently large to affect local communities and to become a material component of its capital programme.
The financial results also reveal the tension between growth and infrastructure efficiency. Amazon’s total net sales rose 12% to $201.94 billion from $180.17 billion a year earlier, while net income declined to $20.69 billion, or $1.89 per share, from $21.19 billion, or $1.95 per share. The company’s operating performance remains substantial, but the decline in net income alongside higher infrastructure spending has made the cost of expansion more visible.
AWS was not the only business shaping the results. Advertising revenue rose 12% to $19.6 billion, below analysts’ expectation of 18% growth and marking the first miss in seven quarters, according to Reuters. Amazon attributed part of the weakness to the timing of its Prime Day shopping event, which occurred in the second quarter in most major markets, including the United States. The company said that, excluding the calendar change, year-on-year revenue growth would have been approximately four percentage points higher.
Amazon forecast fourth-quarter net sales of between $240 billion and $247 billion, representing growth of 12% to 16% from a year earlier. The midpoint of $243.5 billion was slightly below analysts’ estimate of $244.51 billion. The forecast covers the holiday shopping season, a crucial period for its retail business, as consumers face elevated living costs and higher fuel prices.
The retail business is also being expanded through faster deliveries, automation and AI-powered warehouse technology. That extends the infrastructure story beyond data centres. Amazon’s AI spending is connected not only to cloud computing but also to the systems that support storage, logistics and delivery. The supplied material does not quantify those facilities or explain how spending is divided between them, but it confirms that the company is pursuing technology-led expansion across both digital and physical operations.
The central urban question is therefore not whether AI requires infrastructure. It is how the costs and benefits of that infrastructure are distributed. Amazon’s results confirm the scale of the investment and the pressure it is placing on cash flow. They also show that the company recognises a responsibility to invest in communities around its data centres. What remains unclear from the available material is how those investments will be measured, how resource demands will be managed and whether the expected commercial returns will justify the physical expansion.
The next significant signals will come from Amazon’s capital-spending decisions, AWS growth and the implementation of its community investment programme. For cities and infrastructure stakeholders, the company’s forecast of $220 billion in 2026 capital expenditure makes the development of AI facilities a financial issue, a construction issue and a governance issue at the same time.
