Amazon Bid to $279.12 as AWS Reaccelerates to 37% and the $496B Backlog Meets $220B of Capex

Amazon closed Tuesday at $277.42, down $6.60 or 2.32% from Monday's $284.02, with an intraday range of $275.82 to $281.07. Wednesday's premarket has it bid at $279.12, up $1.70 or 0.61%. At 10.79 billion shares outstanding, Tuesday's close puts market capitalization at $2.99 trillion — roughly $10 billion below the threshold the stock cleared 24 hours earlier.

The badge lasted one session. Amazon became the fifth company in history to reach a $3 trillion market capitalization on August 3, 2026, peaking at $3.06 trillion, joining Apple, Microsoft, Nvidia, and Alphabet in that group. Nvidia sits alone at the top with a valuation approaching $5 trillion. Amazon reached $2 trillion in June 2024, meaning it took just over two years to add the third trillion — and it gave the milestone back within a day on a filing that disclosed a pre-scheduled insider sale.

The move that got it there was violent. Shares rallied 15.32% on Friday, July 31 following the second-quarter release, then extended another 5% into Monday's session to close at $284.02. Amazon gained 17% during the week ended July 31 and is up roughly 24% year to date. Over the past four weeks the stock has added 8.86%. Over twelve months it is up 12.39%. The 52-week range runs from $196.00 to $287.20, putting Tuesday's close 3.4% below the high and 41.5% above the low.

Valuation now carries a normalized price-to-earnings ratio of 38.74. Consensus among covering analysts sits at an average price target of $332.56, built on 40 Buy ratings against a single Hold — a Strong Buy consensus with roughly 20% implied upside from the current print. Individual targets moved after the quarter: one lifted from $280 to $320, another from $300 to $325, and one reiterated at $335.

The relative performance is what makes the setup interesting. Amazon is up roughly 23% year to date against Microsoft's 1% — a 22-point gap between two companies selling essentially the same cloud infrastructure product into the same AI demand wave. Alphabet has run 98% over twelve months on the strength of proving AI monetization.

What the stock is actually pricing, and whether the earnings report that produced this run supports it, is a considerably harder question than the headline numbers suggest. The reported figure was $5.75 per share. The operating business earned $1.97.

The $200.6 Billion Quarter And What Actually Beat

Second-quarter net sales for the period ended June 30 came in at $200.61 billion, up 20% from $167.7 billion, against a consensus of $196.47 billion. That is the first quarter in Amazon's history above $200 billion in a single three-month period. Foreign exchange contributed a favorable $0.1 billion, so the 20% growth rate holds on a constant-currency basis.

Operating income climbed 43% to $27.46 billion from $19.2 billion. Profit growth running more than double revenue growth is the operating leverage story that has defined this company since the 2022 cost reset, and it survived a quarter in which capital spending accelerated hard.

The segment breakdown shows where the growth came from. North America, the largest unit, grew 16% to $116.2 billion. International grew 15% to $42.2 billion. Amazon Web Services grew 37% to $42.23 billion — matching the international segment in absolute revenue for the first time while growing at more than twice the rate. Advertising services grew 26% to $19.8 billion.

The adjusted earnings comparison is the one that reflects operations: $1.97 per share against a $1.82 consensus, a beat of 15 cents or 8.2%. That is a solid quarter. It is not the quarter the headline described.

Third-quarter guidance is where the print got complicated. Management guided net sales to $197.0 billion to $202.0 billion, implying 9% to 12% growth against a third-quarter 2025 base of roughly $180.7 billion. Consensus stood at $204.1 billion, meaning the $199.5 billion midpoint represents a $4.6 billion shortfall — a 2.25% miss on the top line.

The explanation is calendar mechanics rather than demand. Amazon moved this year's Prime Day discount event into June rather than its typical July slot, which pulled revenue into the second quarter and created a difficult third-quarter comparison against a period that included last year's event. Excluding the Prime Day shift in both years, third-quarter growth would run nearly 400 basis points higher — which would put underlying growth at 13% to 16% rather than 9% to 12%.

The market accepted that explanation, which is why a light guide produced a 15% rally rather than a selloff. Whether it should have depends on the number examined in the next section.

The $5.75 Headline Is 85% A Non-Cash Anthropic Mark

Reported net income for the quarter was $62.6 billion, or $5.75 per diluted share, against $18.2 billion a year earlier. That is a 244% increase in reported profit, and it is the single most misleading number produced by any company this earnings season.

Inside it sits a $53.4 billion pre-tax non-operating gain, attributed primarily to Amazon's investment stake in Anthropic. That gain equals 85.3% of the reported $62.6 billion net income figure. It is an accounting revaluation of a privately held equity position — not revenue, not operating profit, and not cash. Adjusted earnings of $1.97 per share represent 34.3% of the $5.75 headline, and the remaining 65.7% is the after-tax effect of that mark.

The consensus estimate the headline "beat" was $1.82. Comparing $5.75 against $1.82 produces a 216% beat that does not exist in any economic sense. Comparing $1.97 against $1.82 produces the actual result.

This matters for valuation in a direct, mechanical way. Anchoring a multiple to $5.75 of quarterly earnings implies an annualized run rate above $23 per share, which at $277.42 would put the stock near 12 times earnings — absurdly cheap for a company growing revenue 20%. Anchoring to the operating reality instead, the second-quarter operating income of $27.46 billion annualizes to roughly $110 billion, which is the number any defensible valuation has to work from. The normalized price-to-earnings ratio of 38.74 already reflects that adjustment; the headline does not.

The revaluation also cuts both ways going forward. A private-market mark that added $53.4 billion in one quarter can subtract in another, and it will do so with no connection whatsoever to how many instances AWS sells or how many packages the network delivers. Any quarter in which the Anthropic position is written down will produce a reported earnings collapse that means nothing about the business — the mirror image of what just happened.

The practical read for anyone modeling this stock is to discard the reported earnings line entirely and work from operating income, operating cash flow, and free cash flow. Two of those three tell a strong story. The third is the problem.

AWS At 37% Is The Only Number That Moved The Stock

The 15% single-day rally was not about retail and it was not about the reported earnings. It was about a cloud division that reaccelerated when the market had priced deceleration.

AWS revenue of $42.23 billion grew 37% year over year — 36.7% precisely — against a consensus of $40.54 billion and expectations for 31% growth. That is a $1.7 billion revenue beat and a six-point growth beat, and it represents the fastest AWS growth rate in 18 quarters, the strongest since 2021. Critically, the rate accelerated from 28% in the first quarter, which is a nine-point sequential acceleration on a base above $40 billion.

The profitability improved faster than the revenue. AWS operating income reached $16.62 billion, up from $10.2 billion a year earlier — a 64% increase — producing an operating margin of 39.4% against a $13.62 billion consensus. That is a $3 billion operating income beat in a single segment.

The annualized run rate now stands at $169 billion. On a trailing twelve-month basis AWS generated $148.4 billion in net sales and $54.7 billion in operating income. As a standalone business at that run rate it would rank 24th on the Fortune 500.

The composition detail is what justifies the AI framing. The AWS artificial intelligence business and the unit's custom chips business each eclipsed annualized run rates of more than $25 billion in the quarter, more than doubling from a year earlier. That is $50 billion of combined annualized revenue in businesses that barely existed at scale two years ago, growing at triple-digit rates. The chip line runs on Trainium accelerators and Graviton servers built in-house, which is margin the company keeps rather than pays to a third-party silicon vendor.

The customer wins during the quarter matter as much as the numbers. AWS announced arrangements with Meta and OpenAI — two of the largest independent compute buyers in the world, both of which had been assumed to be structurally aligned elsewhere. Landing either would be significant. Landing both in one quarter is the strongest available evidence that the capacity Amazon is building has committed demand behind it.

Amazon remains a larger force in cloud computing than any other company by absolute revenue. The growth-rate comparison is where it loses, and that is covered below.

The $496 Billion Backlog Is The Whole Bull Case

The disclosure that carried the stock past $3 trillion was not in the press release headline. AWS backlog — contracted work signed but not yet recognized as revenue — reached $496 billion during the quarter.

That figure is 3.35 times trailing twelve-month AWS revenue of $148.4 billion and 2.93 times the current $169 billion annualized run rate. A backlog covering nearly three years of current revenue converts the AWS growth question from a demand forecast into a delivery schedule. The revenue is contracted. The constraint is capacity.

That distinction is what re-rated the stock, because it changes how the capital spending program should be read. Capex against speculative demand is a bet. Capex against a $496 billion signed backlog is working capital. Management has described AWS as capacity-constrained, meaning the binding limit on near-term growth is how fast data centers come online rather than how many customers want the product.

The unit economics management laid out support the return argument. Once a data center opens with servers plugged in, revenue generation begins immediately, and the facility can be monetized for 30-plus years without repeating the startup capital. Servers and networking equipment typically break even in under three years. The shell and the power interconnect are the long-lived asset; the silicon inside is a three-year payback that gets refreshed across multiple generations against the same fixed investment.

If that framework holds, revenue growth eventually outpaces capital expenditure and return on invested capital rises rather than falls — which is the analytical basis for the argument that AWS growth has begun to justify the capex bet.

The obvious rebuttal is that backlog is not revenue and contracts are not cash. A $496 billion figure includes multi-year commitments from a small number of very large AI buyers whose own funding depends on capital markets that have repriced AI risk twice in the past year. Concentration risk inside that backlog is not disclosed. Neither are the terms, the ramp schedules, or the cancellation provisions.

The first genuine evidence on conversion arrives with third-quarter results. Until then, $496 billion is the number the entire re-rating rests on, and it is a number that cannot be verified from outside.

$220 Billion Of Capex And A Memory Price Problem

Management raised 2026 capital spending guidance to approximately $220 billion, up 10% from a prior forecast of roughly $200 billion. The increase was attributed largely to higher memory prices — a cost input Amazon does not control and cannot substitute away from at scale.

That detail deserves more attention than it received. A capex raise driven by demand is a bullish signal: more capacity for more contracted revenue. A capex raise driven by input cost inflation is the opposite — the same physical capacity for more money, which mechanically compresses the return on every dollar deployed. The $20 billion increase buys no incremental compute if it is all memory pricing.

Memory pricing has been running hot across the industry, which is why the two largest storage and NAND names carry option-implied moves of 15% into their reports this week and why one of them is up 423% in 2026. Amazon is on the wrong side of that trade as a buyer.

The absolute scale is difficult to contextualize. At $220 billion, Amazon's 2026 capital spending exceeds the full-year revenue of all but a few dozen companies on the planet. Trailing twelve-month purchases of property and equipment already reached $169 billion, up 64% year over year — a $66.1 billion increase in a single year, primarily reflecting artificial intelligence investment.

The corresponding cash generation is holding up. Operating cash flow increased 33% to $161.4 billion on a trailing twelve-month basis, from $121.1 billion in the comparable period. That is a genuinely large number and it is growing fast. The problem is that $161.4 billion of operating cash flow against $169 billion of capital expenditure produces a negative result, and the guided 2026 spend of $220 billion is $51 billion above current cash generation.

Closing that gap requires operating cash flow to grow roughly 36% from here, or capex to plateau, or Amazon to fund the difference from the balance sheet. The first is possible given AWS margin expansion. The second contradicts the capacity-constrained framing. The third is what has been happening.

The bull framing is that free cash flow compression is intentional infrastructure investment for AI workloads that already command 39.4% segment margins. That framing is defensible. It is also what every company says while spending more than it earns.

Free Cash Flow Swung To Negative $7.6 Billion

The single hardest number in this report: trailing twelve-month free cash flow is an outflow of $7.6 billion, against an inflow of $18.2 billion in the comparable period ended June 30, 2025. That is a $25.8 billion deterioration in twelve months, driven entirely by the property and equipment line.

Amazon is now a company that generates $161.4 billion of operating cash flow, grows revenue 20%, expands operating income 43%, and burns cash. Those five facts are simultaneously true, and reconciling them is the central analytical task for anyone holding the stock at 38.7 times normalized earnings.

The guided path makes it worse before it makes it better. Capital spending rising from $169 billion trailing to $220 billion guided means the free cash flow deficit widens through 2026 even if operating cash flow keeps compounding at 33%. Free cash flow pressure intensifies before it eases — that is arithmetic, not opinion.

The historical precedent cuts in Amazon's favor. The company ran negative free cash flow through the 2021 and 2022 fulfillment buildout and emerged with a materially larger, more profitable retail network. Investors who sold that drawdown missed the recovery. The company has never paid a cash dividend since its 1997 listing, preferring reinvestment and occasional repurchases, and the entire history of the stock is a story of accepting cash burn in exchange for asset accumulation.

The difference this time is the composition of the assets. Fulfillment centers and delivery vans depreciate against a retail business with visible, diversified, decades-old demand. AI data centers depreciate against contracted commitments from a handful of counterparties in an industry two years old. Management's 30-year monetization framework assumes the demand curve for accelerated compute persists across multiple server generations. That is a reasonable assumption. It is not a proven one.

The practical valuation consequence is that discounted cash flow models on this stock currently produce whatever the analyst assumes about 2028 and beyond, because near-term free cash flow is negative and contributes nothing. That is why price targets span from $205 to $485.91 on the same set of reported facts — a 137% spread that reflects assumption dispersion rather than information dispersion.

Advertising Is The Quiet $80 Billion Business

Lost inside the AWS story is a segment that grew 26% year over year to $19.8 billion and now runs at a $79.2 billion annualized pace — on track to pass $80 billion in annual revenue with software-like incremental margins.

Advertising is the highest-quality revenue line Amazon has. It requires no incremental fulfillment capacity, no data center capex, and no inventory. It monetizes traffic the retail business already generates and purchase intent the company already observes. Every dollar above the fixed cost of the ad platform is close to pure operating income, which is why the segment carries more valuation weight per revenue dollar than the retail gross merchandise it sits on top of.

The growth drivers cited for the quarter were AI shopping tools and Prime Video advertising placements. Both are structural rather than cyclical. AI-driven product discovery increases the number of monetizable surfaces inside the shopping funnel, and Prime Video inventory is a supply expansion into a premium video ad market Amazon entered only recently.

For context on scale, a $79 billion advertising run rate growing 26% places Amazon as the third-largest digital advertising business globally, behind only the two dominant search and social platforms. Standing alone at a market multiple appropriate to a high-margin advertising franchise, that business would be worth several hundred billion dollars on its own.

The retail side supporting it delivered its own operational proof point. Record delivery speeds were set for Prime members across the first half of the year, with more than 40% more items delivered same-day or overnight. Faster delivery drives purchase frequency, purchase frequency drives ad inventory, and ad inventory drives the highest-margin revenue in the company. The flywheel argument is not marketing here — it is visible in the 26% advertising growth rate against 16% North America retail growth.

The reason none of this moved the stock is that a $79 billion business growing 26% is simply less interesting than a $169 billion business reaccelerating to 37% with a $496 billion backlog behind it. That relative neglect is arguably the strongest part of the bull case, because advertising is the line item least dependent on the capex thesis working.

Google Cloud At 82% Is The Competitive Problem

The uncomfortable comparison sits one earnings report earlier. Google Cloud reported quarterly revenue up approximately 82% to almost $25 billion, after 63% growth in the first quarter, bringing the subsidiary's trailing twelve-month revenue toward $78 billion.

AWS grew 37% to $42.23 billion. In absolute dollars Amazon remains far larger — $42.23 billion against roughly $25 billion in the quarter, and $148.4 billion against $78 billion on a trailing basis. In growth terms it is being outrun by more than double.

The sequential dollar revenue growth comparison is where the risk concentrates. A competitor adding revenue faster in absolute quarterly increments while running at 60% of the scale raises the possibility of share shifts — a modest investor concern by the standards of a market that has spent 2026 assuming a stable hyperscaler duopoly with Amazon on top.

The strategic read is that the competition has moved from infrastructure to model access. Google Cloud's acceleration is being attributed to proving AI monetization, and Alphabet's 98% twelve-month stock return reflects the market rewarding that proof. Amazon's counter is a different architecture: custom silicon at a $25 billion annualized run rate, its own accelerator stack, and a foundational stake in Anthropic that produced the $53.4 billion mark discussed above.

That Anthropic relationship is more strategically important than the accounting gain it generated. A hyperscaler with a preferred position on a frontier model family has a distribution advantage that raw compute cannot replicate, and the market has not yet priced Amazon's model-layer exposure the way it has priced Alphabet's.

The third competitor is the one to watch on relative performance. Microsoft is up roughly 1% year to date against Amazon's 23% — a 22-point divergence between two businesses with similar cloud growth profiles. That gap says the market has already picked its second-place hyperscaler, and it is not the one that spent the year underperforming.

For the stock, the practical implication is that AWS needs to hold above 30% growth in the third quarter. A deceleration back toward the 28% first-quarter rate against a competitor compounding at 60-plus percent would invalidate the reacceleration narrative that produced the entire July 31 move.

The Insider Sale And The $284.02 Rejection

Tuesday's 2.32% decline came from a filing that contained no new information about the business. The founder disclosed plans to sell 15 million shares worth approximately $4.07 billion, executed under a pre-scheduled 10b5-1 plan — a calendar entry established in advance precisely so that the timing carries no signal about management's view.

The market treated it as a signal anyway, which is what happens when a stock has run 21% in three sessions and needs a reason to take profits. There have been 66 recent insider transactions skewed toward selling, which is unremarkable for a company whose executive compensation is overwhelmingly equity-based and whose share price has just made an all-time high.

The technical significance is more interesting than the fundamental one. The stock printed $284.02 on Monday, tagged $3.06 trillion in market capitalization, and reversed. The 52-week high sits at $287.20. That creates a defined resistance band between $284 and $287 that now has a failed test attached to it, and the round-number attention at $277 — the level that separates a $2.99 trillion valuation from a $3 trillion one — becomes a magnet that gets retested repeatedly.

The premarket bid of $279.12, up 0.61%, puts the stock back above the $278.05 level required for a $3 trillion capitalization. That threshold is going to function as a psychological pivot through the third quarter regardless of what the fundamentals do, because index committees, headline writers, and momentum algorithms all reference it.

Support beneath sits at Tuesday's low of $275.82, then in the low $260s where the stock consolidated on July 31 after touching $264.17. Below that, the pre-earnings base near $235 is the level the entire post-report gain was measured from — a full round trip that would require the AWS thesis to break.

Overbought conditions are flagged on momentum measures and the price-to-sales ratio sits above historical norms, both of which argue for consolidation rather than immediate continuation. A stock that gained 17% in a week does not usually add another 20% without pausing.

Valuation On Numbers That Will Not Repeat

The normalized price-to-earnings ratio of 38.74 is the honest starting point, and it already strips the Anthropic mark. Against 20% revenue growth and 43% operating income growth, that multiple is defensible. Against negative free cash flow, it requires faith.

Working from operating income rather than reported earnings, the second quarter's $27.46 billion annualizes to roughly $110 billion. At a $2.99 trillion capitalization, the stock trades at 27.2 times annualized operating income before tax, interest, and the depreciation charges that $220 billion of annual capex will generate over the coming years. That last point is the one bears press hardest: capitalized spending becomes depreciation, and $220 billion a year of it will weigh on reported operating margins for a decade regardless of how well the assets perform.

The published scenario ranges reflect how wide the honest uncertainty is. One framework built off the $110 billion operating income run rate sets a bull case at $395, a base case at $320, and a bear case at $205 against the $277.42 close — implying 42% upside, 15% upside, and 26% downside respectively. A more aggressive framework carries a twelve-month target of $432.91 with a bull scenario at $485.91 and a bear scenario at $355.96, meaning even its downside case prices the stock 28% higher. Consensus across covering analysts sits at $332.56, about 20% above spot, with 40 Buys against one Hold.

A range from $205 to $485.91 on the same reported quarter is a 137% spread. That is not analytical disagreement about the facts. It is disagreement about whether $220 billion a year of capital expenditure buys a $496 billion backlog at an acceptable return — and nobody outside the company can answer that from the current disclosure.

The segment mix frames the sum-of-parts argument. Retail-related revenue represents approximately 74% of the total, AWS 17%, and advertising services 9%. International operations constitute 22% of revenue, led by Germany, the United Kingdom, and Japan. On that mix, AWS contributes the overwhelming majority of operating income — $16.62 billion of the $27.46 billion total, or 60.5% — from 17% of revenue. Advertising contributes a disproportionate share of the remainder.

The reasonable conclusion is that Amazon is now a cloud and advertising company with a large, structurally low-margin logistics business attached. Valuing it on retail comparables is wrong. Valuing it on hyperscaler comparables makes the multiple look reasonable and the cash burn look normal.

The Levels And Targets That Decide The Rest Of 2026

The setup reduces to one resistance band, one support shelf, and one report. Resistance runs $284.02 to $287.20 — Monday's rejected high and the 52-week peak. Clearing $287.20 on a closing basis opens the path toward the $320 base-case cluster where multiple post-earnings targets have converged, roughly 15% above spot, with $332.56 the consensus average and $335 the high end of the near-term revisions.

Support starts at the $278.05 line that separates a $3 trillion capitalization from a $2.99 trillion one, then Tuesday's low at $275.82. Beneath that, the low-$260s consolidation zone around $264.17 is the first real shelf, and the pre-earnings base near $235 is the level a full retracement targets — a 15% decline that would require the AWS reacceleration to reverse.

The base case into the third-quarter report is a range between $275 and $287. Momentum measures are stretched after a 17% week, price-to-sales sits above historical norms, and the $277 round-number level will draw repeated retests through the quarter. A stock that has gained 24% year to date and just made an all-time high does not need a catalyst to consolidate.

The bull path requires third-quarter AWS growth to hold above 30% and the $496 billion backlog to show visible conversion into recognized revenue, alongside evidence that the $220 billion capex program is buying capacity rather than absorbing memory price inflation. That combination takes $287.20 out and makes the $320 to $335 band the destination, with the more aggressive $395 and $432.91 targets requiring durable AWS share gains against a competitor growing 82%.

The bear path needs only capex indigestion to become visible. AWS growth decelerating back toward 28%, a widening free cash flow deficit beyond the current negative $7.6 billion, or any write-down of the Anthropic position that produces a reported earnings collapse would each be sufficient. The $205 bear case implies 26% downside and assumes the market reprices a company burning cash at 38.7 times normalized earnings.

Amazon at $279.12 is a company that crossed $200 billion of quarterly revenue for the first time, reaccelerated its highest-margin segment to a four-year growth high, disclosed a backlog worth three years of cloud revenue, and burned $7.6 billion of cash doing it. The reported $5.75 in earnings was 85% an accounting mark. The third-quarter report is where the capex bet either starts paying or starts costing.

That's TradingNEWS