Amazon Shares Fall After Q3 Earnings Miss Despite Record AWS Growth and $220 Billion AI Spending Plan

Amazon delivered a quarter that, on its own terms, looked strong in several important places, and the market still sold its shares. The company reported net sales of $201.9 billion for the three months to the end of September 2026, a rise of 12 percent from a year earlier, while its cloud division grew at its fastest pace in almost five years. Operating income jumped sharply, margins reached a record level, and management pointed to a holiday quarter that could bring in as much as $247 billion. None of that was enough to prevent a fall in the share price after the results were released on October 29, 2026.

The reaction says as much about expectations as it does about Amazon itself. Investors had already priced in a great deal of success for the cloud business, and they were watching closely for any sign that the enormous sums being poured into artificial intelligence infrastructure might weigh on profits, cash flow, or the balance sheet. When earnings per share and total sales both came in slightly below what analysts had forecast, traders focused on the shortfall rather than the strength underneath it. Amazon Q3 earnings AWS growth October 2026

What Amazon Reported

The headline numbers

Amazon’s third-quarter net sales reached $201.94 billion, compared with $180.17 billion in the same quarter a year earlier, according to the company’s earnings release. Net income was $20.69 billion, or $1.89 per share, down from $21.19 billion, or $1.95 per share, in the year-earlier period. Operating income rose to $28.1 billion, a large increase from $17.4 billion a year before, and the operating margin reached 13.9 percent, the highest the company has reported. Those figures describe a business that is still expanding and, on an operating basis, becoming more profitable even as it spends heavily elsewhere.

The difficulty for Amazon was not the direction of the numbers but their position relative to forecasts. Analysts had expected earnings of about $1.97 per share on sales of roughly $202.2 billion, according to estimates cited in coverage of the release. Amazon missed both marks, although only narrowly on sales, and the earnings gap of eight cents per share was enough to set the tone for the evening. In a market that had already rewarded technology shares generously through much of the year, a small miss can carry a disproportionate weight, especially when it arrives alongside a very large spending plan.

Why the miss mattered

Earnings misses matter most when they confirm a worry that investors already hold, and in Amazon’s case that worry was cost. The company is building data centres, buying chips, and securing power at a pace that would have seemed extraordinary only a few years ago, and every dollar spent on that build-out is a dollar that does not show up as profit or free cash flow in the near term. A result that falls short of expectations therefore invites a simple question: is the spending starting to bite sooner than shareholders hoped, or is this simply a quarter in which timing and one-off items clouded an otherwise healthy picture?

Amazon’s own commentary pointed toward the second explanation, emphasising operating strength and the momentum in cloud computing rather than the headline miss. Yet the market’s first response was sceptical, with shares falling in regular trading on October 29 and extending the decline after hours. Reports of the after-hours move varied, with some putting it at around 6 percent and others higher, a reminder that early trading after an earnings release can be volatile and that a single percentage figure should not be treated as settled until regular trading resumes. The direction, however, was consistent across reports: down, and by more than the size of the sales miss alone might suggest.

AWS Carries the Quarter

Growth at its fastest pace in years

The clearest bright spot in the quarter was Amazon Web Services, the cloud division that has become the engine of the group’s profits. AWS sales rose to $46.0 billion, an increase of 39.3 percent from a year earlier, up from a 37 percent growth rate in the previous quarter and the fastest pace the division has recorded in nearly five years. The result also beat the roughly $45.7 billion that analysts had expected, giving investors at least one headline figure that came in ahead of forecasts. Management said the division is now running at an annualised revenue rate of about $184 billion, a scale that few technology businesses anywhere can match.

That acceleration matters because cloud growth had, for a period, been slowing as customers tightened budgets and worked through earlier capacity purchases. The return to faster growth suggests that demand for computing power, storage, and artificial intelligence services is again outpacing caution, at least for the largest providers. It also helps explain why Amazon feels able to keep spending: when the division that earns most of the group’s operating profit is growing close to 40 percent, the case for adding capacity looks stronger than it would if growth were fading. AWS accounted for roughly 63 percent of Amazon’s operating profit in the quarter, according to analysis of the results, underlining how central it has become.

What is driving cloud demand

Much of the demand is tied to artificial intelligence, although Amazon, like its rivals, does not break out AI revenue in a way that allows a precise split between traditional cloud work and newer AI workloads. Companies are renting computing capacity to train and run large models, to process vast amounts of data, and to build AI features into their own products, and they often prefer to rent that capacity rather than build data centres themselves. Amazon has signed new agreements with large customers, including Meta and Salesforce, according to coverage of the earnings call, and management projected that AWS revenue growth would exceed 40 percent in both 2027 and 2028.

Those forward projections are ambitious, and they should be read as management expectations rather than guarantees. They depend on continued demand, on Amazon being able to bring new capacity online on time, and on customers staying with AWS rather than shifting work to rivals or to their own infrastructure. Still, the projections help explain the spending plan: if growth really does accelerate beyond 40 percent, Amazon will need considerably more computing power than it has today. If it does not, the company risks having built capacity ahead of demand, with depreciation and interest costs arriving before the revenue to cover them.

The Spending Question

A $220 billion capital plan

Amazon kept its full-year 2026 capital spending guidance at about $220 billion, a figure that was raised in July from an earlier forecast of about $200 billion. Most of that money is going toward data centres and the equipment inside them, including chips, networking gear, and power systems needed to run AI workloads. Chief executive Andy Jassy has previously told investors that Amazon still does not have enough capacity to meet all the AI demand it is seeing, even at this elevated level of spending, a statement that frames the investment as a response to demand rather than a speculative bet.

The scale of the plan is difficult to grasp without comparison. At $220 billion, a single year’s capital spending is larger than the annual revenue of most large companies, and it represents a sharp increase on Amazon’s spending in earlier years. Rising memory chip prices have added to the cost, according to coverage of the July update, which means that each new data centre is more expensive to equip than it would have been a year ago. For supporters of the strategy, that is simply the price of staying competitive in AI. For sceptics, it raises the risk that Amazon is committing itself to a cost base that will be hard to reduce if demand cools.

Cash flow turns negative

The most striking financial consequence of the spending surge is visible in cash flow. Over the twelve months to the end of the quarter, Amazon recorded a free cash flow outflow of $42.4 billion, compared with an inflow of $14.8 billion a year earlier, according to analysis of the results. That swing of more than $57 billion reflects the sheer size of capital expenditure, which in the third quarter alone reached $77.3 billion, well above the roughly $57.5 billion analysts had expected. Long-term debt has also risen sharply since December, to about $173 billion, while quarterly interest expense more than tripled to about $1.7 billion.

None of this means Amazon is short of money in the ordinary sense: the company remains highly profitable on an operating basis and can borrow at scale. But negative free cash flow changes the conversation around a stock. Investors who once valued Amazon for its ability to generate cash must now accept a period in which cash is being consumed by construction, with the payoff expected in future years through higher cloud revenue. That trade-off is at the heart of the current debate about AI spending across the technology sector, and Amazon’s results have become one of the clearest examples of it.

Retail and Advertising in Detail

North America and international sales

Away from the cloud, Amazon’s retail businesses grew more modestly. North America sales rose 6 percent to $112.9 billion, slightly below the roughly $113 billion analysts had expected, while international sales grew 5 percent to $43.1 billion, also a touch below forecasts. Those figures reflect a consumer environment that is steady rather than booming, with shoppers still spending but carefully, and with competition in online retail remaining intense. Amazon moved its large summer sale from July to June this year, which pulled some spending into the earlier quarter and makes year-on-year comparisons for the third quarter slightly harder to read.

The retail result also helps explain why the overall sales miss was small in percentage terms but still noticeable. When the largest part of the business grows at 6 percent and the international arm at 5 percent, total growth depends heavily on AWS to lift the average to 12 percent. That is exactly what happened, but it also means Amazon’s headline growth rate is now more sensitive to cloud performance than to the shopping habits of any single country. A slowdown in either retail segment would be cushioned by cloud growth, just as a cloud slowdown would be harder for retail alone to offset.

Advertising and the holiday build-up

Advertising, another fast-growing part of Amazon, grew by about 12 percent in the quarter, according to analysis of the results, with management expecting a stronger rate in the fourth quarter. Advertising has become an important profit contributor because it allows sellers and brands to pay for prominence on Amazon’s shopping pages, turning traffic into revenue without the inventory costs of retail itself. Growth of 12 percent is solid, but it is slower than the cloud division’s pace, and some analysts had hoped for a faster acceleration as the holiday season approached.

The holiday build-up itself is already under way. Amazon said earlier in October that it would hire 300,000 seasonal workers in the United States for the holiday period, up from 250,000 last year, a sign that it expects a busy final quarter in its warehouses and delivery network. Seasonal hiring on that scale is both a cost and a signal: it adds to wage bills in the short term, but it also suggests management does not expect demand to collapse. How efficiently those workers are deployed, and whether automation offsets some of the cost, will be one of the details to watch when fourth-quarter results arrive.

The Holiday Quarter Outlook

Guidance for the fourth quarter

For the fourth quarter, Amazon forecast net sales of between $240 billion and $247 billion, an increase of 12 to 16 percent from a year earlier, and operating income of between $28.5 billion and $33.5 billion, according to coverage of the guidance. The sales range brackets the average analyst estimate of about $244 billion, which is why some commentators described the outlook as cautious rather than weak: it neither promises a breakout holiday season nor warns of a sharp slowdown. The breadth of the range, $7 billion from bottom to top, also leaves room for the usual uncertainties of the holiday period, from weather and delivery disruptions to last-minute shopping patterns.

One specific headwind was flagged on the earnings call. Chief financial officer David Zinsner told Reuters that an explosion at Intel’s Kiryat Gat fabrication plant in Israel during October had reduced fourth-quarter revenue expectations by approximately $300 million, according to coverage of his remarks. That figure is small beside a $240 billion-plus quarter, but it illustrates how supply-chain events far from Amazon’s own warehouses can still affect its results, particularly when chip supply is tight. It also serves as a reminder that guidance is built on assumptions that can change quickly.

Seasonal hiring and delivery pressure

The fourth quarter is always Amazon’s busiest, and this year the pressure is compounded by the need to handle both retail orders and the computing demands of the holiday period. Shoppers expect fast delivery, sellers expect their products to be visible, and advertisers expect their campaigns to reach customers at the moment of purchase. Meeting all three expectations at once requires warehouses, transport, and systems to work in concert, and any bottleneck in one area can quickly become visible in customer reviews and seller complaints.

Amazon’s decision to hire more seasonal workers than last year suggests it is preparing for volume rather than hoping for it. The risk is that higher staffing costs, combined with the capital spending programme, could squeeze margins in the quarter even if sales meet expectations. The opportunity is that a smooth holiday season could rebuild confidence after the third-quarter miss, showing that the company can both invest heavily and operate efficiently. Investors will therefore look beyond the sales total in January and ask how much profit each holiday dollar produced.

Why Investors Sold Anyway

Expectations were already high

To understand the share price fall, it helps to remember what had happened before the results. Technology shares had risen strongly through much of 2026 on enthusiasm for artificial intelligence, and Amazon was no exception, closing regular trading on October 29 at $243.79, down 4.25 percent on the day even before the earnings release. When a stock has already risen a long way, good results are often not enough: they must be better than good to justify further gains. A small miss, in that context, gives traders who were looking for a reason to take profits exactly the excuse they need.

The pattern is familiar in earnings season. Companies that beat expectations on every measure tend to rise, companies that miss on the headline measures tend to fall, and companies that deliver a mixed picture — a beat in one division, a miss overall — often fall hardest because different investors focus on different parts of the story. Amazon’s mixed picture, with AWS ahead and the group total behind, allowed both optimists and pessimists to find support for their view, but the immediate trading was dominated by those who focused on the miss and the spending.

A wider tech sell-off

Amazon did not fall in isolation. Coverage of October 30 noted declines in technology stocks in Asia and weakness in Nasdaq futures, alongside a sharp fall in Alphabet shares during the October 29 session in New York. When the whole sector is under pressure, individual results are judged more harshly, because investors are already asking whether valuations across the industry have run ahead of profits. A company that might have been forgiven a small miss in a calm market can be punished for the same miss when sentiment is fragile.

That wider context does not excuse Amazon’s miss, but it does caution against reading the share price fall as a verdict on Amazon alone. Sector-wide sell-offs often reverse quickly if the next set of results reassures investors, just as they can deepen if further disappointments follow. For now, the safest reading is that Amazon’s results arrived at an unforgiving moment, when the market was primed to focus on costs and cash flow rather than on growth. Whether that moment passes will depend as much on other technology companies’ results as on Amazon’s own.

What This Means for Amazon’s AI Bet

Capacity, chips and power

The central strategic question raised by these results is whether Amazon is building the right amount of capacity at the right time. Data centres take years to plan, permit, and equip, and the chips and power supplies they need are in short supply. Building too slowly risks losing cloud customers to rivals who can offer capacity sooner; building too quickly risks owning expensive facilities that sit underused if demand grows more slowly than expected. Amazon’s $220 billion plan is, in effect, a bet that demand will stay strong enough to fill the capacity it is adding.

Power is an increasingly important part of that calculation. Large AI data centres consume enormous amounts of electricity, and securing reliable, affordable power has become as much of a constraint as securing chips. Amazon, like other cloud providers, has been investing in energy supply and efficiency, but those investments add further costs and complexity. If power constraints slow the pace at which new capacity can be brought online, Amazon’s spending could rise without a matching rise in revenue in the short term, even if the long-term demand case remains intact.

The case for patience — and the case against it

The case for patience rests on AWS growth itself. A division growing close to 40 percent, with an annualised run rate of $184 billion and projected growth above 40 percent in the next two years, can justify substantial investment if those projections prove even broadly correct. Operating margins at record levels suggest the underlying business model is sound, and the holiday quarter guidance, while not spectacular, points to continued expansion. From this perspective, a period of negative free cash flow is the cost of building a larger, more profitable business for the next decade.

The case against patience rests on the balance sheet and on history. Technology cycles have repeatedly shown that periods of heavy investment can end with overcapacity, price wars, and write-downs, particularly when many companies invest at the same time in the same technologies. Debt has risen sharply, interest costs have tripled, and free cash flow has swung deeply negative, all while the share price had already risen a long way. From this perspective, investors are right to demand clearer evidence that each new dollar of spending will earn an adequate return, rather than simply trusting that growth will eventually cover the cost.

What to Watch Next

Signals in the holiday quarter

The next test will come with fourth-quarter results, when investors will be able to judge whether holiday sales reached the top or the bottom of the $240 billion to $247 billion range, and whether operating income held up despite seasonal hiring and capital spending. They will also watch AWS growth closely: if it stays near or above 40 percent, the investment case strengthens; if it slows sharply, questions about overbuilding will grow louder. Advertising growth, retail margins, and the pace of capital spending in the quarter will all provide additional clues.

Beyond the headline figures, watch for commentary on capacity utilisation and customer commitments. If Amazon can point to long-term contracts, rising utilisation, and customers expanding their use of AWS, the spending plan will look more like a response to locked-in demand. If commentary turns vague, or if capital spending rises again without a matching rise in committed revenue, scepticism is likely to persist. The company’s own language on the next earnings call will therefore matter almost as much as the numbers themselves.

Questions still unanswered

Several important questions remain unanswered after this quarter. The final after-hours share price move varied by source and by timing, so the market’s settled verdict will only become clear in regular trading. The precise split between AI and traditional cloud demand is not disclosed in a way that allows an exact calculation, and the full effect of the Intel plant explosion on chip supply beyond the $300 million revenue figure is not yet known. The impact of recent workforce reductions, including smaller cuts reported in early October, on costs and morale is also unclear.

More broadly, it is not yet known whether the wider technology sell-off that accompanied Amazon’s results was a brief correction or the start of a more sustained reassessment of AI valuations. Amazon cannot control that backdrop, but it can influence how investors view its own place within it by delivering on its AWS growth projections and by showing discipline on spending. Until the holiday quarter provides that evidence, Amazon’s story will remain a contest between a cloud business growing at its fastest pace in years and a spending plan that asks investors to wait for the payoff. That tension, more than any single number, is what the third quarter revealed.

Comments

comments