Article published Sep 28, 2026. Prices below use latest available snapshots.
AMZN Price Structure
Loading timeframes…
Amazon is pulling back inside a long uptrend: down 7.6% since August 28, though still up 2.5% since June 29 and 2.2% above its 200-day average. At the 2026-09-28 settled close, AMZN $246.15 was -2.8% versus its 20-day average and -7.6% over 30 days. The newest verified quarter ends 2026-06-30 (filed 2026-07-31); revenue was $200.61B. Against 2025-06-30 revenue of $167.70B, that is +19.6% year over year. The build is outrunning cash, and debt is filling the gap. Over the trailing twelve months to June, capex of $173.0 billion exceeded operating cash flow of $161.4 billion. Long-term debt was $133.0 billion at June 30, up from $68.8 billion at year-end. The 10-year Treasury yield is at a 19-year high (CNBC, September 25). The standing call is 🟡 Wait for the retest to hold. Entry $242–$247 · stop $229 · first target $287 · medium conviction.
Quick Snapshot (as of 2026-09-28)
| Question | Answer |
|---|---|
| The call | 🟡 Wait for the retest to hold. Entry $242–$247 · stop $229 · first target $287 · medium conviction. |
| What it is | The largest online store and third-party marketplace, the largest cloud-infrastructure provider (AWS), and an advertising business that sells space in front of shoppers searching that store. |
| Why now | Prior case (2026-09-25): AI agents have reached the store. Meta launched Muse, an agent that can operate a browser and complete a purchase once the user approves it, on September 8, and Amazon blocked it around September 20 (Bloomberg, September 21). Over the same weeks AWS signed a custom-chip deal with Qualcomm (September 8) and a backup-generator deal with Generac (September 16). |
| Proof | Latest verified period ended 2026-06-30, filed 2026-07-31: revenue $200.61B. Prior case: June quarter (SEC 10-Q): revenue $200.6 billion, up 19.6%, the fifth straight quarter of faster growth. Operating margin was 13.7%, the highest in eight quarters on file. AWS grew 37% to $42.2 billion, per Amazon's release. |
| Objection | The build is outrunning cash, and debt is filling the gap. Over the trailing twelve months to June, capex of $173.0 billion exceeded operating cash flow of $161.4 billion. Long-term debt was $133.0 billion at June 30, up from $68.8 billion at year-end. The 10-year Treasury yield is at a 19-year high (CNBC, September 25). |
| Stock setup | The long trend remains intact; -2.8% versus the 20-day, +2.2% versus the 200-day, RSI 41.2. |
Action Matrix (reviewed 2026-09-28; levels are standing conditions)
| Signal | If | Then |
|---|---|---|
| 🟢 Buy | A week of daily closes inside $242–$247. The top of the zone is the $247.30 support shelf; the bottom is the $242.70 volume-weighted average price measured from the July low. | Entry. The retest of the July earnings gap has held above the 200-day ($240.85), and the stop sits just under the heaviest-volume price from before the report. |
| 🟡 Wait | Price between $247 and $257 (the 50-day and the $256.80 top of the value area), or a close above $257 before October 29 | No edge in the middle of the range. A breakout just before the report means chasing into a binary event. |
| 🔴 Out | A daily close below $229 | The July gap has filled below the $232.10 pre-report volume node and the 200-day has failed. The long trend is under test. |
| 🟠 Target | $287 | The 52-week high ($287.20), set in the days after the July report. Trim or trail from there. |
| 🔴 Thesis breaks | The October 29 report shows AWS growth slowing from 37% while capex guidance rises again, or advertising growth slowing as agent shopping spreads | The compute build loses its demand evidence, or agents are drawing shoppers away from the store's ad-funded search pages. |
The Story Right Now (as of 2026-09-28)
Amazon is pulling back inside a long uptrend: down 7.6% since August 28, though still up 2.5% since June 29 and 2.2% above its 200-day average.
Data as of 2026-09-28 close: AMZN $246.15, RSI 41.2, weak-down, -2.8% vs SMA20, +2.2% vs SMA200, 7D -4.8%, 30D -7.6%, 1Y +10.8%, -14.3% from the 52-week high.
Condition check. The $246.15 close sits inside the prior $242–$247 retest zone, but the required week of daily closes there is not established by one session. The condition remains pending, and the October 29 report is the next business test.
Business and mechanism. The largest online store and third-party marketplace, the largest cloud-infrastructure provider (AWS), and an advertising business that sells space in front of shoppers searching that store.
Prior case (2026-09-25). AI agents have reached the store. Meta launched Muse, an agent that can operate a browser and complete a purchase once the user approves it, on September 8, and Amazon blocked it around September 20 (Bloomberg, September 21). Over the same weeks AWS signed a custom-chip deal with Qualcomm (September 8) and a backup-generator deal with Generac (September 16).
Prior proof (2026-09-25). June quarter (SEC 10-Q): revenue $200.6 billion, up 19.6%, the fifth straight quarter of faster growth. Operating margin was 13.7%, the highest in eight quarters on file. AWS grew 37% to $42.2 billion, per Amazon's release.
The countercase. The build is outrunning cash, and debt is filling the gap. Over the trailing twelve months to June, capex of $173.0 billion exceeded operating cash flow of $161.4 billion. Long-term debt was $133.0 billion at June 30, up from $68.8 billion at year-end. The 10-year Treasury yield is at a 19-year high (CNBC, September 25).
Newest statement check. The newest verified quarter ends 2026-06-30 (filed 2026-07-31); revenue was $200.61B. Against 2025-06-30 revenue of $167.70B, that is +19.6% year over year.
Cash conversion and spending. The April–June SEC 10-Q reports $45.39B of operating cash and $54.21B of cash purchases of productive assets, a 119.4% capex-to-operating-cash ratio for that quarter. The cash comparison does not include every form of lease-financed infrastructure and should not be read as total economic investment.
News and attention check. The September 28 hosted pull returned five quote, press-release index or market-data pages and no dated issuer event for the session. The cash-buildout thesis remains tied to the June filing; no cause is assigned to this close. No company event for AMZN was recorded between September 24 and 28. 4 social posts from tracked accounts mentioned it between September 24 and 28; that is attention, not company news.
The call: 🟡 Wait for the retest to hold. Entry $242–$247 · stop $229 · first target $287 · medium conviction.
Extended analysis
Price Data (live)
These values come from the desk's settled-close summaries and update with each close. The dated sections above keep the numbers from their own date.
| Metric | AMZN |
|---|---|
| As of (close) | 2026-10-16 |
| Price ($) | 250 |
| RSI (14) | 45 |
| vs SMA20 (%) | -1.3 |
| vs SMA50 (%) | -2.8 |
| vs SMA200 (%) | 3 |
| SMA20 ($) | 253.23 |
| SMA50 ($) | 257.08 |
| SMA200 ($) | 242.73 |
| 1D (%) | -1.87 |
| 7D (%) | -4.74 |
| 30D (%) | 1.64 |
| 3M (%) | 1.12 |
| 1Y (%) | 16.57 |
| 2Y (%) | 33.77 |
| 1Y vs S&P 500 (pts) | 1.03 |
| From 52-week high (%) | -12.95 |
Prior Research (prepared by 2026-09-25; current call and filed-period check above)
Summary
Amazon runs three profit engines on one base of infrastructure. The store, which combines Amazon's own retail with a third-party marketplace, brings in shoppers. On marketplace sales Amazon is "not the seller of record" and earns "fixed fees, a percentage of sales, per-unit activity fees, interest, or some combination thereof" (10-K). Advertising sells access to those shoppers through sponsored ads, display and video. AWS rents compute, storage, databases and AI services to developers and enterprises. The company reports three segments: North America, International and AWS.
AWS sells relief from scarcity. AI workloads need chips, memory, power and buildings that take years to put together, and AWS builds them ahead of demand at a scale that analyst Ben Bajarin says competitors still have to catch up to (via TheBuildout, September 11, attributed). The store faces a newer problem. AI agents that shop on a user's behalf can reach Amazon's shelves without passing through Amazon's search pages, which is where the advertising revenue is earned.
The decisive uncertainty is whether a build running at $173 billion a year turns into cash before financing costs and agent-driven disintermediation compress the multiple. The June quarter shows the demand. The cash-flow statement shows Amazon spending ahead of that demand and borrowing to do it.
Why This Moment Is Different
The same three weeks brought a consumer agent to Amazon's store, a set of AWS supply deals and a bond-market rout. Taken together they cover both halves of the investment case.
| Date | Event | Source |
|---|---|---|
| September 8 | Meta launches Muse, with a free tier or a $20 or $100 monthly subscription depending on usage | CNBC |
| September 8 | Qualcomm issues Amazon warrants for up to $4 billion of its stock as part of an AI-infrastructure deal; Qualcomm will make custom chips for Amazon | CNBC; Yahoo Finance, September 9 |
| September 10 | Amazon lets its advertisers into ChatGPT, "hesitant to open its sprawling webstore to external AI platforms" | CNBC |
| September 15 | AWS says it cannot restore service to its Bahrain and UAE facilities six months after the Iran strikes | CNBC |
| September 16 | Amazon receives warrants for up to $340 million of Generac stock in a backup-generator supply deal | CNBC |
| September 21 | Amazon blocks Muse from shopping on Amazon.com | Bloomberg; Forbes; The Register |
| September 21 | Shopify partners with Muse on payment technology | MarketWatch |
| September 22 | "Retailers are picking sides on AI shopping agents" | Semafor |
| September 23 | Meta Connect: $1,299 VR glasses and the Muse Charm pendant | CNBC, September 24 |
| September 24–25 | 30-year Treasury yield highest since 2004; 10-year at a 19-year high | CNBC |
Amazon and Muse: the Stated Reason and the Ads Motive
What Muse is. Meta's agent carries out tasks: it can operate a browser, fill in forms and ask for approval before a purchase (coverage of the launch and the block, September 8–21). Downloads reached 730,000 in about five days and 2.5 million in about thirteen (CNBC, September 21). According to reporting TheBuildout attributes to @RihardJarc, each user is given a dedicated virtual machine. The Meta side, including the re-rating, the estimates and the capex, is covered in the desk's deep dive "Meta's AI Agent Re-Rated the Stock 34% in a Month. The Estimates Never Moved." (September 15, refreshed September 22).
What Amazon said. Coverage of the block (Bloomberg, September 21; GeekWire) describes a warning under Amazon's Conditions of Use with three stated grounds. Meta did not notify Amazon, the agent does not identify itself, and it appears to capture and store customer credentials. The security concern has outside company. Hacker News carried Ars Technica's "Meta's Muse has a serious 0-day" and a developer post titled "I asked Meta's Muse for its filesystem and it sent…". Neither establishes Amazon's specific credential claim. Coverage of the block also notes that Amazon had earlier sued Perplexity over its Comet agent and had acted against Google's and OpenAI's shopping agents. The Muse block therefore follows an existing pattern.
What Amazon did not say. The motive the coverage emphasizes is advertising, summed up in Forbes's September 23 headline, "Amazon's $68 billion reason to block Meta's Muse". Sponsored placements are sold on the pages where a person searches and browses. An agent that compares products and completes the purchase itself never sees those pages. Amazon's first-quarter release put advertising above $70 billion over the prior twelve months. The desk reads the evidence as supporting both motives at once. The stated security grounds are real enough to cite in public, and blocking also protects the advertising revenue sold on those pages.
The line Amazon is drawing. On September 10 Amazon let its advertisers into ChatGPT (CNBC). Amazon is willing to send its advertising demand into outside AI products on its own terms, but it will not let outside agents shop its store without identifying themselves. The desk infers from those two facts that Amazon's objection is to losing control of the shopping session, not to AI surfaces in general. Shopify took the opposite position and partnered with Muse on payments (MarketWatch, September 21). Semafor called it "the first sign of battle lines being drawn in the way we spend money online" (September 22).
Amazon's Own Agent and Shopping-AI Efforts
The company-side record on this is thin. Analysts on the first-quarter call asked about Amazon's agentic-commerce initiatives (StockStory's summary of the call). The second-quarter release, as summarized from Amazon's investor-relations page, highlighted advertising-agent metrics and wider adoption of Trainium and Graviton. Chris Camillo writes that "Amazon itself is simultaneously building agentic shopping that reaches beyond Amazon" (@ChrisCamillo, September 22, attributed). No filing or release on hand names or sizes Amazon's own shopping assistants or agents, so none is described here. The October 29 report and call are the place to get one.
Threat or Moat
The agent can bypass Amazon's search pages but still has to buy from Amazon's shelves and use its delivery. The 10-K names "selection, price, and convenience, including fast and reliable fulfillment" as the principal competitive factors in retail. An agent choosing among sellers still needs the item in stock and delivered quickly, and those are the things Amazon's fulfillment network provides. Camillo makes the same argument: "$AMZN's logistics, fulfillment + distribution moats don't suddenly become worthless because an AI agent sits upstream of the transaction."
The part an agent can route around is the advertising funnel, which would hit margin rather than volume. If agents take a meaningful share of purchase decisions, Amazon can still sell the units while earning less advertising on each sale. Its 10-K already lists technologies that "restrict online advertising or affect our ability to customize advertising" as a factor in its results. Citrini's September 24 thread describes how disruption usually shows up in an incumbent's stock: "technological disruption manifests first as multiple compression, current earnings tend to look reassuring" (@citrini, attributed). If agents erode the ad funnel, that is the pattern to expect at Amazon.
AWS benefits from agents regardless of which one wins. @RihardJarc's thesis, as TheBuildout reports it, is that agents running on a dedicated virtual machine per user create demand for cloud machine capacity, which favors AWS's scale (September 15, attributed). Meta's own reported WhatsApp delay for lack of hardware supports the scarcity side of that argument, though not an AWS contract.
Desk read (as of 2026-09-25): for now, agentic shopping is a threat to one margin line and a demand tailwind for AWS. Amazon's control of the shelf and the delivery holds up. The advertising take on each sale is the exposed part. The measurable test is advertising growth in the October 29 report against AWS growth. Muse's adoption makes both halves larger at once.
AWS in the Compute and Power Story
Capex. Purchases of property and equipment were $131.8 billion in 2025 and $98.4 billion in the first half of 2026: $44.2 billion in the first quarter and $54.2 billion in the second (SEC XBRL). Amazon guided to about $200 billion for 2026 at its February report, and the shares fell on that guide (CNBC, February 5). Barron's said the stock jumped after the July report "despite a massive AI capex hike" (July 31). A September 7 supply-chain report (UDN via @dnystedt on TheBuildout, which labels it a rumor) put the new figure at $220 billion. No filing or transcript on hand confirms that number. Depreciation and amortization rose to $20.0 billion in the June quarter from $15.2 billion a year earlier, so the build is already weighing on reported costs.
Custom silicon. First-quarter coverage (EarningsLabs; TIKR) put Amazon's chip business above a $20 billion annual run rate. With the second-quarter results, Amazon put its AI and chip businesses above $25 billion run rates. StockStory's summary of the first-quarter call reports an AWS backlog of $364 billion that excludes a new $100 billion Anthropic deal. The September supply-chain reports are attributed and not verified. They say AWS secured more TSMC 3-nanometer capacity (Digitimes via @paurooteri, September 14), that its Trainium 3.5 will use Intel's EMIB-M packaging (@jukan05, September 11, labeled a rumor), and that Qualcomm will co-develop custom inference chips and optical interconnects with a procurement ceiling reported at $60 billion (Nikkei via TheBuildout, September 8). Only the Qualcomm warrants are confirmed by mainstream coverage (CNBC).
Power. Amazon's disclosed on-site power deals point to backup generation and gas. Generac expects to deliver $2.4 billion of backup generators to Amazon over 2027–2028, with warrants that vest toward an $8 billion procurement threshold (TheBuildout via @ShanuMathew93; the $340 million warrant figure and the 40%-plus Generac jump are CNBC's). AWS is reported to be advancing large-scale natural-gas projects for its data centers (TheBuildout, September 8). Four U.S. localities paused new data-center approvals between September 8 and 14, while an Amazon project in Gilroy, California, was exempt (TheBuildout, September 15). Amazon's sustainability chief told Semafor the company should be held accountable on data-center impact (September 22). No direct relationship between Amazon and Bloom Energy was found in filings or coverage. The Bloom contracts on file are with Oracle, Brookfield, AEP and Equinix.
Anchor customer. Anthropic is the AI lab named alongside AWS's backlog: StockStory's summary of the first-quarter call reports a new $100 billion Anthropic deal. Two facts from this week bear on that relationship without resolving it. Anthropic signed a $12 billion computing deal with Akamai (Bloomberg, September 24), and a federal appeals court upheld the Pentagon's designation of Anthropic as a supply-chain risk (CNBC, September 25). No source quantifies the effect on AWS.
Financing and physical risk. Long-term debt was $68.8 billion at the end of 2025, $122.6 billion at March 31 and $133.0 billion at June 30, with $67.0 billion of long-term debt proceeds in the half (SEC XBRL). Amazon filed prospectus supplements for securities offerings on September 9 and 11 and an 8-K on September 14. Their terms are not reviewed here. Separately, AWS said it cannot restore service at facilities in Bahrain and the UAE six months after drone strikes in March (CNBC, September 15). Gulf data centers carry live war risk.
How the Business Makes Money
The February 10-K (fiscal 2025) describes customer sets of consumers, sellers, developers and enterprises, content creators, advertisers and employees, served through three reporting segments.
| Engine | What it sells | Who pays | Economic role |
|---|---|---|---|
| Store (North America, International) | Selection, price and fast delivery across online and physical stores; Prime subscriptions; devices (Kindle, Fire, Echo, Ring) | Consumers; third-party sellers pay fees for selling and fulfillment | The traffic engine. Seller fees make marketplace sales a services business. |
| Advertising | Sponsored ads, display and video | Sellers, vendors, publishers, authors | The layer sold on top of shopper traffic. It is the part Muse can bypass. |
| AWS | Compute, storage, database, analytics, AI and machine-learning services | Developers and enterprises, start-ups, government, academia | The largest profit engine and the part consuming the capex. |
The 10-K lists 1,576,000 full- and part-time employees at year-end 2025. Its competition section is unusually direct about AI. Competitors include "virtual assistants, and other online and app-based means of discovering, using, or acquiring goods and services... including through artificial intelligence". The filing also says that AI "facilitate[s] competitive entry and comparison shopping".
What the Filing Says the Company Is Becoming
The filed risk factors describe the company Amazon is building. It must predict data-center demand, because failures to do so "result in excess or insufficient fulfillment or data center capacity, service interruptions, increased costs, and impairment charges". It relies on "a limited group of suppliers for semiconductor products, including products related to artificial intelligence infrastructure such as graphics processing units". Its results depend on the availability and price of "land, water, and energy", and it warns that returns on "automation, artificial intelligence, and machine learning technologies" may not recoup "significant" investments. That describes an infrastructure company that also runs a store.
Satellites extend the same pattern. The second-quarter 10-Q describes an agreement to acquire Globalstar for about $10.9 billion including debt, paid in cash or Amazon stock capped at $90 a share, with service agreements with Apple. Closing is expected in 2027, subject to regulatory approvals and satellite milestones. The deal folds into Amazon Leo, the company's satellite-internet service.
What the Street and Public Conversation Say
The sell-side record on hand is thin and mostly dated. Morningstar held a $260 fair value after the February report (history, February 9). UBS put Amazon on its September list of tech picks for the data-center buildout (MarketWatch, September 18). At the July report, EPS of $5.75 compared with a consensus near $1.8, a gap the non-operating gain explains more than operations do. Third-quarter revenue guidance of $197–$202 billion came in below the $203.9 billion consensus (StreetInsider, July 31).
Public conversation about Amazon's side of the Muse story is close to absent. A September 25 social search for "AMZN Amazon Muse" returned nothing relevant. Meta's Muse sweep the same day surfaced the Amazon block as a Hacker News story (Forbes link, 152 points; The Register, 19 points), well below Muse's own launch post (666 points) and the security posts. On X, Camillo's argument that the moat survives drew the most engagement of the Amazon takes on file (836 likes). @ParadisLabs called Amazon one of the "top three Mag 7's" alongside Nvidia and Meta (September 24, 66 likes, with the author saying he is buying). That is sentiment, not corroboration.
Insiders sold and none bought. Executive Chair Jeff Bezos sold 1,209,649 shares, about $346.5 million, in a Form 4 filed August 5. Between August 25 and September 3, six insiders sold 65,848 shares, about $17.1 million, among them CEO Andy Jassy, AWS CEO Matt Garman and CFO Brian Olsavsky (SEC Form 4s). The desk's insider screen ranks Amazon as the largest seller in its tracked universe. Whether these sales ran under trading plans was not checked. Short-interest data is unavailable.
The Conviction Chain
| Link | Evidence today | What strengthens it | What breaks it |
|---|---|---|---|
| AWS demand outruns supply | AWS +37% in the June quarter, its fastest growth in 18 quarters (release). Capacity is reserved into 2027–2028 (call highlights, Yahoo Finance). Q1 backlog was $364 billion excluding the $100 billion Anthropic deal (StockStory). | Q3 AWS growth holds in the mid-30s and backlog rises again | AWS slows while capex guidance rises |
| The build converts to cash | Twelve-month operating cash flow of $161.4 billion against $173.0 billion of equipment purchases. Depreciation is up to $20.0 billion a quarter. | Operating cash flow grows faster than capex, and capex guidance holds | Another capex raise with flat cash flow and more debt |
| Custom silicon bends the cost curve | AI and chip businesses above $25 billion run rates; the Qualcomm deal; reported TSMC 3nm capacity | Trainium customers disclosed beyond one lab | Packaging or foundry slips, or the anchor customer's legal and compute-diversification risk grows |
| The ad funnel survives agents | Advertising above $70 billion over twelve months (Q1 release). Amazon has blocked unidentified agents and exported its ad demand to ChatGPT. | Q3 advertising growth steady while agents spread | Advertising growth slows, or agent-mediated retail (Shopify's route) takes measurable share |
| The stock can compound from here | +10.8% on the year and +2.2% above the 200-day (as of 2026-09-28); about 19.7x trailing GAAP earnings (inflated by non-operating gains) and 28.6x trailing operating income (as of 2026-09-28 at $246.15; price-only repricing of September 25 share and earnings basis) | The $242–$247 base holds, then a close through $287 | A close below $229 |
Conviction is medium, and the weakest link is cash conversion. The stock's February slide followed the $200 billion capex guide (CNBC, February 5), and the capex bill has only grown since.
Financial Proof
Figures are from Amazon's SEC 10-Q and 10-K XBRL. Massive's quarterly revenue matches on every quarter, and the growth rates and margins are script-computed. "Equipment purchases" means gross purchases of property and equipment.
| Quarter | Revenue ($B) | YoY | Operating income ($B) | Operating margin | Non-operating income ($B) | Net income ($B) | Diluted EPS | Operating cash flow ($B) | Equipment purchases ($B) | Source |
|---|---|---|---|---|---|---|---|---|---|---|
| Q3 2024 | 158.9 | 11.0% | 17.4 | 11.0% | 0.6 | 15.3 | 1.43 | 26.0 | 22.6 | SEC 10-Q |
| Q4 2024 | 187.8 | 10.5% | 21.2 | 11.3% | 1.1 | 20.0 | 1.86 | 45.6 | 27.8 | SEC 10-K |
| Q1 2025 | 155.7 | 8.6% | 18.4 | 11.8% | 3.3 | 17.1 | 1.59 | 17.0 | 25.0 | SEC 10-Q |
| Q2 2025 | 167.7 | 13.3% | 19.2 | 11.4% | 1.7 | 18.2 | 1.68 | 32.5 | 32.2 | SEC 10-Q |
| Q3 2025 | 180.2 | 13.4% | 17.4 | 9.7% | 10.7 | 21.2 | 1.95 | 35.5 | 35.1 | SEC 10-Q |
| Q4 2025 | 213.4 | 13.6% | 25.0 | 11.7% | 1.6 | 21.2 | 1.95 | 54.5 | 39.5 | SEC 10-K |
| Q1 2026 | 181.5 | 16.6% | 23.9 | 13.1% | 16.0 | 30.3 | 2.78 | 26.0 | 44.2 | SEC 10-Q |
| Q2 2026 | 200.6 | 19.6% | 27.5 | 13.7% | 53.4 | 62.6 | 5.75 | 45.4 | 54.2 | SEC 10-Q |
Operations are improving: revenue growth has quickened for five straight quarters and operating margin has risen for three. The earnings line needs restraint. Non-operating income of $53.4 billion made up most of the June quarter's $62.6 billion of net income, and the filed figures do not break that line into its parts. The source of the gain is not established here, so operating income is the cleaner measure of the business. Operating cash flow is not inflated by customer prepayments. Unearned revenue fell by $0.28 billion in the quarter and by $0.75 billion over the trailing twelve months. Supplier payables added $9.4 billion in the quarter and $13.9 billion over the trailing twelve months, which is ordinary retail working capital. On trailing-twelve-month filed figures, capex (gross equipment purchases) of $173.0 billion exceeded operating cash flow of $161.4 billion, which leaves free cash flow negative. Amazon's own free-cash-flow definition nets sale proceeds and incentives, so it reads somewhat differently.
Valuation, as of 2026-09-25 at $249.67: 10,786,313,572 shares on the 10-Q cover (July 22) give an equity value of about $2.69 trillion. That is about 20x trailing-twelve-month GAAP net income of $135.3 billion, a figure flattered by the non-operating gains, and about 29x trailing operating income of $93.7 billion (script-computed from the filed quarters). Q3 2026 results are scheduled for October 29.
Balance sheet (June 30, 2026, SEC 10-Q): cash of $78.2 billion and short-term investments of $44.8 billion, against long-term debt of $133.0 billion. Q3 guidance calls for revenue of $197–$202 billion and operating income of $22.5–$26.5 billion (Amazon's release).
The Stock Setup and What to Watch (as of 2026-09-25)
The desk's September 25 settled-close summaries place Amazon at $249.67. That is under the 20-day ($254.31) and 50-day ($256.17) averages and above the 200-day ($240.85), with the golden cross intact and the regime labeled pullback. The structure runs as follows: support shelf at $247.30; average price measured from the July low at $242.70; heaviest pre-report volume at $232.10; value area from $207.40 to $256.80; eight weeks of base under a $287.20 high. The one-year trend is intact (+14.45%, +3.7% above the 200-day), so this is a deep pullback in a working trend, not a breakdown. The weakness is relative: -2.77 points against the S&P 500 over the year. Amazon is in none of the Momentum Board tiers as of September 25.
Where the connected names sit (live). This is a trend-question view, so it carries the one-year and 200-day columns.
| Name | As of | Price ($) | 30D (%) | 1Y (%) | vs SMA200 (%) | 1Y vs S&P 500 (pts) | Link to Amazon on file |
|---|---|---|---|---|---|---|---|
| Amazon | 2026-09-28 | 246.15 | -7.61 | 10.79 | 2.2 | -4.56 | — |
| Meta | 2026-09-28 | 715.62 | 23.81 | -3.74 | 14.2 | -19.1 | Muse blocked from Amazon.com (Bloomberg, September 21) |
| Qualcomm | 2026-09-28 | 187.48 | 14.18 | 13.42 | 10.8 | -1.94 | Custom chips; warrants to Amazon (CNBC, September 8) |
| Generac | 2026-09-28 | 205.21 | 11.65 | 23.75 | -3 | 8.4 | Backup generators; warrants to Amazon (CNBC, September 16) |
| Shopify | 2026-09-28 | 144.01 | -5.81 | -3.35 | 10.9 | -18.71 | Took the other side: partnered with Muse (MarketWatch, September 21) |
| Bloom Energy | 2026-09-28 | 262.87 | 24.72 | 257.16 | 29 | 241.8 | No direct relationship found |
| Cloudflare | 2026-09-28 | 353.99 | 18.06 | 61.95 | 52.4 | 46.59 | No direct relationship found |
| Palantir | 2026-09-28 | 187.48 | 0.64 | 4.82 | 23.3 | -10.54 | No direct relationship found; shares a UBS pick list (MarketWatch, September 18) |
What to watch:
- October 29, after the close (scheduled): third-quarter report. The three numbers that decide the call are AWS growth against 37%, advertising growth, and the capex guide against the $200 billion February figure and the reported $220 billion.
- Agent policy. Whether Amazon opens a sanctioned route for outside agents, as Shopify did, or keeps blocking them, and whether Muse's rollout reaches WhatsApp.
- Financing terms. The coupons on the September offerings, and whether rates at multi-decade highs change the pace of issuance.
- Anthropic. Any change to the $100 billion AWS relationship after the Pentagon designation ruling and the Akamai deal.
Kill tests: a daily close below $229, or a third-quarter report that shows AWS slowing while capex rises.
Cross-References
- Meta: the desk's deep dive "Meta's AI Agent Re-Rated the Stock 34% in a Month. The Estimates Never Moved." (September 15, refreshed September 22) covers Muse, Meta's capex guide and MTIA silicon. This profile covers the Amazon side of the same standoff.
- Bloom Energy, Cloudflare, Palantir: the September 23 profiles of each. No direct Amazon relationship was found in filings or coverage.
- Campaigns: Amazon is not an instrument in any active campaign. The "AI power: generation and grid equipment" campaign counts a verified off-grid or behind-the-meter commissioning among its failure conditions. AWS's reported gas projects would matter there only once they are commissioned. The "AI spending: the supplier earnings reversal case" campaign reviews at the October hyperscaler reports. Amazon's third-quarter capex guide is one of its inputs.
- Scans: the September 25 AI Infrastructure report noted that Amazon's 200-day cushion, which had narrowed for three cycles, stabilized at +3.7%. Amazon is also covered by the Focus, Magnificent 7, FAANG, AI Scan and Tech watchlists and by several buildout lists (custom AI silicon, AI data-center operators, nuclear power for data centers).
Financial Proof
Filed-period check (as of 2026-09-28): The newest verified quarter ended 2026-06-30 and was filed 2026-07-31. Revenue was $200.61B; GAAP net income was $62.65B; diluted EPS was 5.75. Source: Massive income statements and the filing. Dated historical tables below retain their own period basis.
The recent filed revenue series shows the direction of the business; missing cells are left unavailable.
| Period end | Revenue | YoY | Operating income | GAAP net income | Filing date |
|---|---|---|---|---|---|
| 2024-09-30 | $158.88B | — | $17.41B | $15.33B | 2024-11-01 |
| 2025-03-31 | $155.67B | — | $18.41B | $17.13B | 2025-05-02 |
| 2025-06-30 | $167.70B | — | $19.17B | $18.16B | 2025-08-01 |
| 2025-09-30 | $180.17B | +13.4% | $17.42B | $21.19B | 2025-10-31 |
| 2026-06-30 | $200.61B | +19.6% | $27.46B | $62.65B | 2026-07-31 |
Research Log
| Date | Update |
|---|---|
| 2026-05-29 | AMZN $270.64, RSI 59.9, constructive. Base target $259 exceeded. |
| 2026-06-16 | AMZN $246, RSI 43.1, below SMA20; -7.12% 30D. AWS thesis intact; wait for RSI stabilization. |
| 2026-06-26 | AMZN $232.69, RSI 40, down; -3.64% 7D, -14.41% 30D, +16.73% 3M. Continued slide — now at the SMA200 test ($232.77 ≈ price). Golden cross intact, regime "pullback." RSI 40 is the oversold watch level. Quality-dip rotation frame: if SMA200 holds, this is the risk/reward inflection. If breaks, $215-225 entry zone next. No new fundamentals re-researched. |
| 2026-07-11 | AMZN $245.34, RSI 51, weak-down/basing; +0.91% 7D, +1.59% 30D, +2.27% 3M. SMA200 test from 06/26 resolved bullish — price now +5.1% above SMA200. Regime upgraded from "pullback" to "basing." RSI reset to neutral. -11.93% from 52wk high (narrowed from -16.47%). No new fundamentals re-researched. |
| 2026-07-17 | AMZN $247.23, RSI 51.4, weak-down/basing; +0.77% 7D, +1.16% 30D, -0.42% 3M. Essentially flat vs 07/11 — basing regime holding steady, no fresh deterioration. -11.25% from 52wk high. No new fundamentals re-researched. |
| 2026-07-25 | AMZN $232.11, RSI 37.2, down/pullback; -6.12% 7D, -0.92% 30D, -12.08% 3M. Basing pattern broken — back below both SMA20 (-4.8%) and SMA200 (-1.1%). -16.12% from 52wk high (widened from -11.25%). No new fundamentals re-researched. |
| 2026-07-28 | AMZN $230.86, RSI 36.3, down/pullback; -6.74% 7D, -3.86% 30D, -12.23% 3M. Pullback continuing, no reversal signal — still below SMA20 (-5.1%) and SMA200 (-1.6%). -17.12% from 52wk high (widened from -16.12%). Q2 2026 earnings call scheduled, not yet reported. No new fundamentals re-researched. |
| 2026-07-31 | AMZN $271.58, RSI 67.7, strong-up; 7D +17.00%; 30D +12.36%; 3M +1.24%; -2.51% from 52wk high. No new fundamentals re-researched. |
| 2026-08-08 | AMZN $274.48, RSI 64, strong-up; 7D +1.07%; 30D +12.67%; 3M +0.66%; -5.07% from 52wk high. No new fundamentals re-researched. |
| 2026-08-14 | AMZN $262.65, RSI 53.4, strong-up; 7D -4.31%; 30D +3.02%; 3M -0.56%; -7.68% from 52wk high. No new fundamentals re-researched. |
| 2026-08-22 | AMZN $258.63, RSI 49.7, strong-up; 7D -1.52%; 30D +5.63%; 3M -2.89%; -9.95% from 52wk high. 1Y +16.53%, vs SMA200 +8.50%. No new fundamentals re-researched. |
| 2026-08-28 | AMZN $266.43, RSI 55.9, strong-up; 7D +2.99%; 30D +17.55%; 3M -1.56%; -7.23% from 52wk high. 1Y +15.04%, vs SMA200 +11.6%. No new fundamentals re-researched. |
| 2026-09-04 | AMZN $258.51, RSI 49.5, strong-up; 7D -2.97%; 30D -5.19%; 3M +5.07%; -9.99% from 52wk high. 1Y +9.69%, vs SMA200 +8.10%. News pulled 2026-09-04 shows no new print this week (Q2 2026 reported 07-31, already reflected); coverage flags an ongoing warehouse-strike/regulatory-scrutiny watch alongside continued AWS growth, but nothing dated to this window explains the move. No new fundamentals re-researched. |
| 2026-09-11 | AMZN $256.78, RSI 48.9, strong-up; 7D -0.67%; 30D -3.93%; 3M +7.64%; -10.59% from 52wk high. 1Y +11.67%, vs SMA200 +7.10%. News pulled 2026-09-11: routine coverage (board appointment, earnings calendar) with no item dated to this week's small move. No new fundamentals re-researched. |
| 2026-09-18 | AMZN $253.71, RSI 48.0, weak-down (downgraded from strong-up); 7D -1.20%; 30D -4.56%; 3M +3.81%; -11.66% from 52wk high. 1Y +9.72% with alpha1y -5.29 — the weakest long-trend read of the large-cap group here, though still +5.6% above the 200-day. News pulled 2026-09-18: quote pages and the July quarter release, nothing dated on this move. Stock-events checked: none in this window. No newly filed period on disk this cycle. Removed a held-position row from the snapshot table — recommendation surfaces are book-blind. |
| 2026-09-23 | AMZN $249.27, RSI 44.1, -2.3% vs SMA20, 30D -4.88%. Still under the short averages, +3.6% above the 200-day; 1Y +12.94% with 1Y vs S&P 500 -2.83 pts. Wednesday -2.24% on 1.29x volume; news not pulled for this move (receipt fetched 2026-09-22 carries only July-quarter items). Entry zone moved from $215–$225 to the $240–$247 support area. Stock-events checked: none in the window; Q3 report scheduled 10-29. Filed-period check: newest quarter on disk is Q1 2026 (filed 2026-04-30); Q2 2026 reported July 30 is not on disk and no table here cites it. Call at that top: 🟡 wait for support to hold, entry $240–$247, stop $229, first target $287, medium conviction; objection was capex running ahead of cash, with CNBC's September 23 report of the Muse block noted as context, not a cause. |
| 2026-09-28 | AMZN $246.15, RSI 41.2, -2.8% vs SMA20, 30D -7.6%. The long trend remains intact; The newest verified quarter ends 2026-06-30 (filed 2026-07-31); revenue was $200.61B. Against 2025-06-30 revenue of $167.70B, that is +19.6% year over year. The September 28 hosted pull returned five quote, press-release index or market-data pages and no dated issuer event for the session. The cash-buildout thesis remains tied to the June filing; no cause is assigned to this close. No company event for AMZN was recorded between September 24 and 28. |
Sources
Primary filings and company releases
- Amazon 10-Q for the quarter ended June 30, 2026 (filed 2026-07-31): quarterly series, trailing cash flow, debt, cash, share count, Globalstar terms.
- Amazon 10-Q for the quarter ended March 31, 2026.
- Amazon 10-K for fiscal 2025 (filed 2026-02-06): business, competition and risk-factor quotations.
- Amazon Q2 2026 results release, EX-99.1 and investor-relations release: AWS $42.2 billion, +37%, fastest in 18 quarters; Q3 guidance; AI and chip run rates.
- Amazon Q1 2026 results release: advertising above $70 billion over the trailing twelve months; AWS +28%.
- SEC Form 4s: Bezos, filed 2026-08-05; Herrington, filed 2026-09-03.
- September filings not read for terms: 424B5, 2026-09-09; 424B5, 2026-09-11; 8-K, 2026-09-14; 8-K, 2026-09-09.
Muse and agentic shopping
- Bloomberg: Amazon blocks Meta's Muse AI agent from its retail site (2026-09-21)
- GeekWire: Amazon blocks Meta's Muse AI assistant in new standoff over agentic shopping
- Forbes: Amazon blocks Meta's new Muse AI agent from shopping on Amazon.com (2026-09-21)
- Forbes: Amazon's $68 billion reason to block Meta's Muse (2026-09-23)
- Yahoo Finance: Amazon blocks Meta's Muse AI
- The Register: Amazon shows Meta's Muse AI shopping agent the door (2026-09-21)
- CNBC: Meta's standoff with Amazon over Muse (2026-09-23)
- MarketWatch: Amazon and Shopify make starkly different moves in the brewing battle over AI shopping (2026-09-21)
- Semafor: Retailers are picking sides on AI shopping agents (2026-09-22)
- CNBC: Meta pushes into personal AI agents (2026-09-08)
- CNBC: Muse downloads are surging (2026-09-21)
- CNBC: Zuckerberg debuts $1,299 Meta VR Glasses and Muse Charm (2026-09-24)
- CNBC: Amazon lets its advertisers into ChatGPT (2026-09-10)
- Ars Technica: Muse has a serious 0-day; mouse.dev: Muse runtime export (Hacker News headlines only)
- X: @ChrisCamillo, 2026-09-22 and thread; @citrini, 2026-09-24; @ParadisLabs, 2026-09-24; @RihardJarc on Muse and AWS, 2026-09-15; @RihardJarc on Muse VMs, 2026-09-08; @zephyr_z9 on the WhatsApp delay, 2026-09-21
AWS compute, power and financing
- CNBC: Qualcomm issues warrants to Amazon (2026-09-08); Yahoo Finance: Qualcomm to make custom chips for Amazon (2026-09-09); Nikkei; @zubo_semicon
- CNBC: Amazon obtains right to buy stock in Generac (2026-09-16); Bloomberg: Generac surges on $8 billion Amazon data-center deal (2026-09-17); @ShanuMathew93 on Generac
- @ShanuMathew93 on AWS natural-gas projects (2026-09-08); @ShanuMathew93 on local data-center pauses (2026-09-15)
- @dnystedt on the $220 billion capex report (2026-09-07) and UDN; @dnystedt on Bajarin (2026-09-11)
- @paurooteri on TSMC capacity (2026-09-14); @jukan05 on Trainium 3.5 packaging (2026-09-11)
- Semafor: Amazon sustainability chief on data centers (2026-09-22)
- CNBC: AWS can't restore service to Bahrain, UAE facilities (2026-09-15)
- Bloomberg: Akamai surges on $12 billion Anthropic deal (2026-09-24); CNBC: appeals court upholds Pentagon designation of Anthropic (2026-09-25)
- EarningsLabs: Amazon Q1 2026; TIKR (2026-04-28); StockStory via TradingView: analyst questions from Amazon's Q1 call; Yahoo Finance: Q2 2026 call highlights
- Barron's: Amazon stock jumps despite a massive AI capex hike (2026-07-31); StreetInsider: Q2 results vs consensus (2026-07-31); CNBC: shares fall on $200 billion 2026 capex guidance (2026-02-05); Morningstar (2026-02-09); MarketWatch: UBS tech picks (2026-09-18)
Macro
- CNBC: 30-year Treasury yield highest since 2004 (2026-09-24); CNBC: 10-year yield ends a volatile week (2026-09-25); Bloomberg: new tech issuance repricing existing debt (2026-09-24); CNBC: Fed delivers its first hike in three years (2026-09-16)
Internal receipts
- Tape: ai-infra (AMZN row, 2026-09-25 close, Schwab; support shelf, volume profile, anchored VWAP, 52-week pivot); cohort rows from
tech.json(META),thebuildout-advanced-packaging-2d3d-leaders.json(QCOM),thebuildout-power-distribution-pdus-busways.json(GNRC),focus.json(SHOP, NET),war-sensitivity.json(BE),ai-infra.json(PLTR). - Financials: company-facts and
key-financials.json(SEC XBRL through the Q2 2026 10-Q); income-statements,cash-flow-statements.json,balance-sheets.json,10k-business.json,10k-risk_factors.json; the desk's own tooling (2026-09-25; FCF -$11.625B single-source SEC; OCF flagged disputed against Massive's non-quarter value, quarterly figures agree). - News: latest (fetched 2026-09-25T20:44Z, ok, quote pages only) and
news/articles.jsonl. - Events: 2026-07-30-amzn-earnings, 2026-10-29-amzn-earnings (scheduled, after close).
- Insiders: filing-watch (generated 2026-09-25); 2026-09-03-amzn-4-0001936006-26-000024,
2026-09-09-amzn-424b5-0001104659-26-106115.md,2026-09-11-amzn-424b5-0001104659-26-107122.md,2026-09-14-amzn-8-k-0001104659-26-107526.md. - Captures: 2026-09-{07,08,09,10,11,14,15,16,17,21,22,23,24,25}-*
(files named in the Sources links above); 2026-09-22-ChrisCamillo-nobody-not-even-the-companies-themselves-know-ho-51525960,2026-09-24-citrini-how-do-companies-act-when-they-are-being-existen-74767166.md,2026-09-24-ParadisLabs-as-of-right-now-it-feels-like-the-top-three-mag--05005600.md,2026-09-25-social-search-amzn-amazon-muse.md,2026-09-25-social-search-meta-muse.md`. - Scans and prior work: 2026-09-25-ai-infrastructure, 2026-09-25-monster-scan, 2026-09-25-insider-scan; 2026-09-15-meta-deep-dive; 2026-09-23-{BE,NET,PLTR}.md
; campaignsai-power-bottleneck,ai-capex-digestion,ai-power-delivery` (market-only show).
2026-09-28 refresh receipts
2026-09-28 settled-close receipt: *, AMZN row dated 2026-09-28; tape values above are bound to the current summary fields.
2026-09-28 financial-period check: income-statements and key-financials; newest verified 2026-06-30 period, filed 2026-07-31 (Massive statement / SEC period cross-check).
2026-09-28 news receipt check: latest, fetched 2026-09-28; fresh successful receipt; event-quality review summarized above.
2026-09-28 stock-event scan:
research/stock-events/, no AMZN row dated 2026-09-24 through 2026-09-28.2026-09-28 captured-follows scan: tweets has 4 September 24–28 document(s) tagging AMZN.
Retrospective social-search receipt (captured 2026-09-29): 2026-09-29-social-search-amzn; quality/attention check only, not a September 28 event or price source.