AMZN › analyze
AMZN · Analyze
Price: $277.42 · Market cap: $2.99T · Sector: Consumer Cyclical / Internet Retail · Beta 1.45 Verdict: 🟡 HOLD — conviction [6.0] · First analysis file ever written on this name.
A portfolio-specific passage was removed from the public build.
The one-paragraph version
Amazon's Q2 2026 was a genuine operating beat: revenue $200.6B (+20%), operating income $27.5B (+43%), and AWS accelerating to +36.7% — its fastest in 18 quarters — with operating income up 63% and margins expanding 650bps to 39.4%. Consensus had modelled 31% AWS growth. That is real, it is operational, and it deserves a re-rating. But the stock added roughly $500B of market value on a print where $4.90 of the $5.75 EPS was an unrealized mark on Anthropic — an asset Amazon cannot sell, priced off a secondary market that brokers describe as having almost no sellers. Meanwhile trailing twelve-month free cash flow is −$7.6B against +$18.2B a year ago, FY2026 capex was raised to $220B because memory got more expensive, and 76% of that capex goes to AWS while retail capex is flat. The honest verdict: the AWS re-rating is real and roughly 1.5–2x too large.
🚩 The headline multiple is an artifact — the fourth instance found today
| ttm to 2026-06-30 | |
|---|---|
| Revenue | $775.68B |
| Operating income | $93.71B → 12.1% margin |
| Reported net income | $135.28B → 17.4% net margin ← above the operating margin |
| Implied pre-tax non-operating income | ≈ $52.6B |
A net margin above the operating margin cannot come from operations. In Q2 2026 alone, Amazon booked $53.4B of pre-tax non-operating gain, "primarily from our investments in Anthropic."
| Basis | ttm EPS | P/E at $277.42 |
|---|---|---|
| Headline GAAP (what every screener showed) | $12.44 | 22.2x |
| Core, ex non-operating marks | ~$6.84 | ~40.6x |
| Consensus forward | $10.33 | 26.9x ← the honest anchor |
The decisive tell: consensus forward EPS ($10.33) is BELOW trailing EPS ($12.44). Forward estimates exclude future marks; trailing includes them. When forward earnings are lower than trailing at a company growing revenue 20%, the trailing number is not earnings. Consensus already understands this. The screener does not.
Anthropic carrying value: $16.2B (Dec-2025) → $190.4B (Jun-2026), on roughly $13–18B invested. That is Level 3, non-marketable, ownership-capped and lock-up restricted — unsellable. $190.4B ÷ the May-2026 $965B primary = 19.7% ownership; ÷ the $1.2T July secondary = 15.9%; third-party estimates put Amazon "mid-to-high teens." So Amazon is marking at or above the last primary, in a market with essentially no supply — where absence of sellers raises price mechanically without new information.
⚠️ Unreconciled and flagged. Third-party reporting of the Q1 10-Q puts the combined carrying value at $74.2B at 2026-03-31. $74.2B → $190.4B is a +$116.2B step-up against a $53.4B disclosed gain plus ~$5B of new investment — leaving a ~$58B gap that could not be closed from outside the filing. Either the third-party figure is wrong, part of the step-up ran outside the non-operating line, or the disclosed gain is net of something. This should be checked against the actual Q2 10-Q footnote before any conclusion rests on the exact number. The direction and order of magnitude are not in doubt.
Filed as the fourth confirmed instance today — after GOOGL, NOW and AMAT — in pitfall-unrealized-equity-marks-break-headline-pe.
1. General Health (§1)
Free cash flow has gone negative for the third time in six years
| FY2023 | FY2024 | FY2025 | ttm Jun-26 | |
|---|---|---|---|---|
| Operating cash flow | $84.95B | $115.88B | $139.51B | $161.40B |
| Capex | −$52.73B | −$83.00B | −$131.82B | −$169.01B |
| Free cash flow | $32.22B | $32.88B | $7.70B | −$7.60B |
FY2026 capex guidance was raised to ~$220B from ~$200B. CFO Olsavsky, plainly: "The higher cost of memory pushing this number up." — the same memory supercycle this agency analysed on MU today is now a direct cost line for the hyperscalers. 1H capex was $98.4B, so 2H implies ~$122B.
Context that matters, in Amazon's favour: this is the third FCF-negative episode in six years (2021, 2022, now), and the first two were capex cycles that resolved into strong FCF recovery in 2023–24. The pattern is not new.
What IS new, and what should worry an owner: 1. Scale — the FY2026 capex guide ($220B) exceeds the entire prior decade of free cash flow (~$151B, 2016–2025). 2. Funding — the 2021–22 cycle was self-funded. This one is not. Long-term debt went $65.6B → $128.9B in six months; H1 long-term debt proceeds were $67.0B; the filing states "We expect to undertake additional financing activities in 2026"; a $25B bond sale followed in July. Total debt ~$251.6B. Jassy declined to say how the incremental $20B is funded — "Nothing to share." 3. Reinvestment quality — the 2021–22 capex went into fulfillment, where returns were demonstrable. This goes into AI infrastructure, where AWS's own return on segment assets is measurably falling.
🚩 Where the capex actually goes — disclosed in the 10-Q segment note, and rarely quoted
Net additions to property and equipment:
| Segment | H1-2025 | H1-2026 | YoY | % of H1-26 |
|---|---|---|---|---|
| AWS | $36.51B | $90.12B | +147% | 76% |
| North America | $16.37B | $23.27B | +42% | 20% |
| International | $4.04B | $4.27B | +6% | 4% |
| Corporate | $1.30B | $1.00B | −23% | 1% |
| Total | $58.22B | $118.66B | +104% |
AWS is 76% of capex, up from 63% a year ago; retail capex was nearly flat (+6% in Q2). This is no longer "Amazon investing in logistics." It is an AI-infrastructure company bolted onto a retailer — and the money has left the segment where the historical moat was built.
🚩 Capital allocation — the ten-year record, and it is unique among megacaps
| 2016–2025 cumulative | |
|---|---|
| Reported net income | ~$246B |
| Free cash flow | ~$151B |
| Cash returned to shareholders | $0 |
Zero buybacks in H1 2026 (treasury stock unchanged at $(7,837)M). No dividend, ever. SBC ~$19.5B/yr. Diluted share count +2.0%/yr.
Amazon has generated owner cash — ~$151B over the decade at 61% conversion. That is real. But it has never distributed a dollar, and it is the only megacap where the shareholder's claim gets smaller every year while the balance sheet gets more levered. MSFT, GOOGL, META and AAPL all fund comparable AI builds and retire shares. Over five years the dilution is a ~10% drag compounding against the thesis.
The fair characterisation: a company that reinvests genuinely well, now reinvesting at a scale it cannot self-fund, into a lower-return asset class than the one that built it. The absence of a buyback is defensible. The absence of any track record of stopping is the real issue — an owner's return depends entirely on terminal reinvestment quality, not on cash coming back.
Growth, normalized
⚠️ Prime Day moved from July 2025 into June 2026. That inflates Q2's +20% and deflates the Q3 guide.
| Q2-2026 reported | +20% |
| Q3-2026 guide midpoint (~$199.5B) | +11% |
| Blended | +15.0% ← the honest underlying rate |
Q3 operating income is guided to $22.5–26.5B — below the $27.5B Q2 actual — while AWS, 60.5% of profit, is accelerating. Something in the cost base is consuming a rising quarter.
2. Moat (§2) — WIDE, trend stable, composition deteriorating
Segments — Amazon is three businesses and they do not share a moat
Q2 2026:
| Segment | Revenue | YoY | Op income | Margin | % of revenue | % of op income |
|---|---|---|---|---|---|---|
| North America | $116.18B | +16% | $9.12B | 7.9% | 57.9% | 33.2% |
| International | $42.20B | +15% | $1.72B | 4.1% | 21.0% | 6.3% |
| AWS | $42.23B | +37% | $16.62B | 39.4% | 21.1% | 60.5% |
AWS is 21.1% of revenue and 60.5% of operating income. On a ttm basis AWS earns $54.7B on $148.4B (36.8%) while all of retail earns $39.0B on $627.3B (6.2%).
🚩 And the retail number is worse than it looks. North America's 7.9% includes a $640M one-off IEEPA tariff refund; ex that, NA margin was 7.3% — down from 7.5%. International is flat at 4.1%. The ~30,000 corporate roles cut since October 2025 have not shown up in segment margin. Strip advertising out of the retail segments and the core retail business earns roughly $3B of operating income on ~$560B of revenue — about 0.5%.
By product line: Advertising $19.81B +26.2% (near-100% incremental margin, closed-loop purchase data — the strongest moat in the company), 3P seller services $46.78B +15.9%, Online stores $70.43B +14.6%, Subscriptions $13.73B +12.5%, Physical stores $5.79B +3.6%.
The central economic fact: AWS margin up, AWS returns down
| FY2024 | FY2025 | ttm Jun-26 | |
|---|---|---|---|
| AWS operating margin | 32.9% (Q2-25) | — | 39.4% (Q2-26, +650bps) |
| AWS return on segment assets | ~20% | 18.0% | 15.6% |
| AWS asset turns | 0.51x | — | 0.42x |
Margin per revenue dollar is rising; revenue per asset dollar is falling faster. The synthesis: AI cloud is a structurally more capital-intensive, lower-ROIC business than the classic cloud it displaces. AWS is not becoming a worse business — it is becoming a heavier one, and the terminal ROIC of AI-era AWS will sit below that of 2019-era AWS.
Consolidated incremental ROIC on FY2025's capital was only ~6.6% ($9.1B incremental NOPAT on $138.1B incremental invested capital) — at or below any plausible WACC. The bull calls that a lag. That is also exactly what year two of a capital-destroying boom looks like from the inside, and no one can distinguish them in real time.
✅ A genuine credit on earnings quality
Amazon shortened server and networking useful life from six years to five, effective 2026-01-01, absorbing a headwind (~$0.7–1.3B) voluntarily — while peers extended lives to flatter margins. The +650bps of AWS margin expansion happened despite a shortened life. Amazon's depreciation is the most conservative in the group, and that materially strengthens the credibility of the margin figure. (Contrast the concern raised at Alphabet, where a six-year server life has never been revisited.)
The depreciation risk here is therefore volume, not accounting: AWS took ~$167B/yr of capex at the FY2026 run-rate, and as those cohorts land on a 5-year life, incremental annual AWS depreciation rises by roughly $70–90B over three years. To hold a 39.4% margin while AWS revenue compounds ~25% to ~$290B, gross profit must rise ~$140B on ~$142B of incremental revenue — a ~100% incremental margin. That is impossible. AWS margins must compress from here; only the slope is in question.
Moat sources, by segment
| Source | Retail marketplace | Advertising | AWS |
|---|---|---|---|
| Network effects | Present, strong — buyers→sellers→selection | Present (derivative) | Weak |
| Switching costs | Weak for buyers, moderate for sellers (FBA, reviews, rank) | Weak — budgets move quarterly | Present, strong — data gravity, egress fees, IAM rewrites. The deepest moat Amazon owns. |
| Cost advantage | Present — fulfillment density compounds | Present, extreme — ~100% incremental margin | Present but narrowing — Trainium real, offset by memory inflation |
| Intangibles / brand | Present — Prime 93–95% retention | Present — closed-loop data no walled garden can match | Moderate, eroding to Azure's bundle |
| Efficient scale | Absent | Absent | Present — capital intensity now excludes all but 4–5 firms |
The strongest moat in the company is advertising — highest margin, fastest growth among mature lines, near-zero incremental cost, fed by data nobody else has. It receives almost no capex. The 76% goes to the segment with the weakest network effects. That is the "composition deteriorating" finding in one sentence.
🚩 The circularity, quantified
| Amazon cash in | Their AWS commitment | |
|---|---|---|
| Anthropic | $18.0B (+ a $20.0B credit facility) | ">$100B over 10 years" |
| OpenAI | $50.0B ($28.7B funded at 6/30) | $38B existing + $100B/8yr = $138B |
| Total | ~$68B | ~$238B+ of the $496B backlog = ~48% |
⚠️ Read the facility clause carefully: the $20B Anthropic facility has no amount available at inception and becomes drawable as Amazon reaches compute-delivery milestones. Amazon lends Anthropic money in proportion to the compute Amazon sells Anthropic. That is vendor financing and should be named as such.
So the $496B backlog — the single most-cited bull statistic — is roughly half owed by two loss-making entities whose ability to pay depends on Amazon's own capital and on continued external fundraising. The other ~$258B, from ordinary enterprises, is genuinely high quality. (Industry-wide, not Amazon-specific: Microsoft's $625B RPO is ~45% OpenAI.)
Adversarial stress-test — where each attack is actually landing
| Attacker | Landing? |
|---|---|
| Walmart | Landing hardest. Q1 FY27: global eCommerce +26%, US delivery +45%, marketplace +50%, advertising +37–44% vs Amazon's +26%. US eCommerce share 8.4% → 9.6%. Stores-as-fulfillment gives 93% same-day reach at near-zero marginal capex. Compressing Amazon's growth rate in one category, not its moat. |
| Shopify + OpenAI | Landing quietly — the most under-priced attack. AI-driven traffic to Shopify stores +8x YoY, orders from AI search +13x; 2M+ stores in ChatGPT product discovery. Amazon's retail moat is a demand-aggregation moat — it owns where product search starts. Amazon has restricted third-party agents on its site: protects today's take rate, cedes tomorrow's surface. This is the one attack aimed at the moat itself. |
| Microsoft / Google | Landing on share, failing on economics. AWS ~28–31% (from ~32% in 2021); Google Cloud 14% growing 63%; Azure 21%. AWS is still losing relative share while growing faster in absolute terms than at any point since 2021 — a good problem in a market compounding 35%. |
| Shein / Temu | Defeated — but not by Amazon. US de minimis repeal destroyed the economics (Temu US DAUs −52%, ad spend −95%). Amazon was rescued by trade policy, not by competitive advantage. Do not score this as moat evidence — if the rule reverses, the threat returns intact. |
Regulatory — an unusually dense window
| Matter | Status | Risk |
|---|---|---|
| FTC v. Amazon (marketplace monopoly, +18 states) | Trial 2026-10-13 — ~10 weeks away | The real one. Seller price parity, Buy Box degradation, Prime/FBA tying, "Project Nessie." A liability finding would reprice the retail segment even if remedies take years. |
| EU DMA — AWS as gatekeeper | Preliminary findings June 2026; decision Nov 2026 | Underrated. Would ban AWS self-preferencing and mandate portability — a direct attack on switching costs, AWS's strongest moat. Fines to 10% of global turnover. |
| FTC ROSCA / Prime dark patterns | Settled, $2.5B | Closed |
| Xockets ITC complaint | Investigation instituted June 2026 | Low probability, high severity tail |
✅ Kuiper/Leo resolved: the FCC granted a waiver on 2026-06-05 extending the 1,618-satellite milestone to 2027-07-30. Satellites launched after 2026-07-30 forfeit 2020/21 spectrum-priority rights, restorable at 50% deployment and automatically by March 2028. A real penalty, not a licence loss. Kuiper remains a ~$1B/quarter cash drain with no revenue.
Moat: WIDE. Trend: STABLE, with the composition deteriorating. The moat is not shrinking; capital is being reallocated from the strong part (advertising, marketplace) to the contested part (AI infrastructure). That shows up as declining consolidated ROIC long before it shows up as competitive weakness.
3. Valuation (§3)
Dividend models are N/A. Graham gives $120 on the inflated $12.44 EPS and is unusable — it capitalizes the Anthropic mark and rests on an equity base that now carries $122.3B of Level 3 marks. Recorded and set aside.
Sum of the parts (FY2027E basis, $B)
| Piece | Method | Bear | Base | Bull |
|---|---|---|---|---|
| AWS | FY27 NOPAT × multiple | 855 (19x $45B) | 1,620 (27x $60B) | 2,190 (30x $73B) |
| Advertising | FY27 NOPAT × multiple | 390 (13x $30B) | 595 (18x $33B) | 790 (22x $36B) |
| Retail ex-ads | × ~$610B revenue | 170 (0.28x) | 255 (0.42x) | 395 (0.65x) |
| Equity stakes ($190.4B + ~$12B) | haircut, net of ~$42B DTL | 50 (−70%) | 88 (−45%) | 160 (−20%) |
| Kuiper/Leo | burn vs option value | −25 | 0 | +35 |
| Net debt | −155 | −155 | −155 | |
| Equity value | 1,285 | 2,403 | 3,415 | |
| Per share | ÷10.78B | $119 | $223 | $317 |
A "base-plus" case — AWS holding 38% margin through FY2027 as depreciation lands more slowly, ads at 21x — gives ~$258/share. Defensible base band: $225–265.
Reverse-DCF cross-check — and this is where the honesty is
At the current ~$3.15T enterprise value, a 24x normalized-FCF multiple implies the market is underwriting ~$131B of annual free cash flow against −$7.6B ttm — a $139B swing, and a ~13% FCF margin on ~$1.0T of FY2029 revenue. Amazon's best-ever FCF margin is ~5% (2020). It is reachable only through mix — AWS + ads going from ~31% to ~45% of revenue at 40%+ margins.
The entire valuation reduces to one assumption:
| FY2029 capex as % of revenue | Value/share |
|---|---|
| 13% | ~$290 |
| 15.5% | ~$245 |
| 18% | ~$200 |
Five points of capex intensity on a $1T revenue base is $50B/yr of FCF — roughly $100/share. Nobody knows this number, including management, and Jassy explicitly declined to discuss funding.
Fair value
$225 – $290, midpoint ~$255. Bull case $315–340; bear tail $120–150. At $277.42 the stock sits in the top quartile of its own fair-value range.
4. Sentiment — optimism tipping into euphoria
- 62 analysts, ZERO sells, recommendation mean 1.35, mean target $323.29 (+16.5%). A dozen target raises within hours of the print (Benchmark $370→$400, JPMorgan $330→$365, Morgan Stanley $330→$335).
- Short interest 0.99% of shares out. There is no wall of worry left to climb.
- Price 1.3% below the all-time high, +18% above the 200-DMA.
- 🚩 Insider activity — the Yahoo trap again. Yahoo reports 2,379,344 shares "purchased" across 57 transactions, net +1.64M. This is false: there is not a single open-market purchase in the entire dataset back to 2024. Every entry is a sale, a $0.00 stock gift, or a null-value RSU vest. Jassy's 50,000 (2026-05-21) and 49,680 (2026-02-20) "buys" are vests, each followed within days by sales. Actual behaviour: uniform programmatic selling, no buying at any price, with the heaviest officer selling clustered at $255–275 — right at today's price. Bezos's last open-market sales were ~25.0M shares for ~$5.65B in July 2025; since then only stock gifts. (Third instance of this Yahoo miscount today — see
pitfall-yahoo-insider-purchases-counts-rsu-grants.)
Sentiment verdict: OPTIMISM tipping into EUPHORIA on the AWS narrative. The distinction that saves it from full euphoria: consensus is not capitalizing the Anthropic gain — forward P/E (26.9x) sits above trailing (22.3x). The euphoria is aimed at AWS reacceleration, and AWS reacceleration is real.
5. Verdict
🟡 HOLD — conviction [6.0]. Do not initiate new capital at $277.42.
Was the +20% move an appropriate re-rating, or the market capitalizing a paper gain?
Both — roughly half and half, tilted toward over-extrapolation.
The legitimate part is real and should not be minimized. Consensus modelled 31% AWS growth; Amazon delivered 36.7% with operating income up 63%, a fifth consecutive quarter of acceleration, a $496B backlog, and AI and custom silicon each above a $25B run-rate. Upgrading AWS's growth path and terminal value on that evidence justifies roughly +8 to +12% on the whole company — $240–360B.
The stock added roughly $500B.
The residual $150–260B has no fundamental anchor. It traces to a $5.75 EPS print against ~$2.00 consensus — a 185% "beat" — of which $4.90 was a mark on an asset Amazon cannot sell. That headline propagated mechanically through screens, quant value factors, and every "cheapest megacap at 22.3x" note written that week. The re-rating is real, and it is roughly 1.5–2x too large.
Why HOLD and not TRIM — and why the skew is the argument
Upside to the bull case is +21%. The bear case is −55%, and it requires only that AWS revert to 20% growth — which is where it was four quarters ago. When numerator and denominator compress together (growth falls and the multiple falls), the downside is not symmetric.
Against that: the business quality is genuinely elite, the AWS beat was operational, advertising at +26% is the most underappreciated line in the company, and Amazon has looked expensive on FCF at every prior capex cycle and selling on that has always been wrong (2018–19 and 2020–22 both produced FCF air pockets that resolved into structurally higher earnings). The mechanism is identical: build ahead of a demand curve you can see and competitors cannot fund. Jassy says AWS is capacity-constrained through 2027 — Amazon is turning away revenue, which is the strongest possible evidence the spend is demand-led rather than speculative.
Conviction [6.0]. The debate round adjudicated 5.5; I have set 6.0. The reasoning for moving it up: on a like-for-like basis AMZN trades ~9% above its fair-value midpoint while GOOGL (rated 6.5 today) trades ~13% above its own, and Amazon's growth and segment quality are at least comparable. The reasoning for staying below GOOGL's 6.5: Amazon has never returned a dollar of capital in a decade, dilutes 2%/yr, runs negative FCF, carries $251.6B of debt, and roughly 48% of its flagship backlog is owed by companies it funds. A company whose headline multiple is an artifact does not earn a 7.
A portfolio-specific passage was removed from the public build.
Break triggers (falsifiable, dated)
| # | Trigger | Date | Falsifies |
|---|---|---|---|
| 1 | AWS operating margin <36% (≥340bps sequential compression), or consolidated op income below the $22.5B guide floor | Q3 print ~2026-10-29 | Depreciation-wave denial → FV to $200–225. Conversely AWS ≥38% margin on ≥33% growth falsifies the bear thesis → FV $290–320 |
| 2 | Anthropic IPO prices or trades below ~$700B | reported "a go", H2'26–H1'27 | The mark. ~30% cut ≈ $57B pre-tax / ~$4.10/sh of negative EPS. This is the event that tests a valuation set by a no-seller secondary market. |
| 3 | FY2027 capex guided ≥$270B while AWS growth guides below 30% | ~Feb 2027 | The incremental-ROIC case outright. Secondary tell in the same filing: a further cut to server life below 5 years — management confirming the bear case in its own footnotes. |
Also dated and material: FTC monopoly trial 2026-10-13 · EU DMA cloud gatekeeper decision Nov 2026. Two structural-risk decisions inside 100 days.
The single number that would most change this answer
The dollar amount of the $496B RPO — and of quarterly AWS revenue — owed by counterparties in which Amazon holds an equity stake or an outstanding funding commitment.
The depreciation question resolves itself within two prints whether Amazon discloses it or not. The circularity question is unanswerable from outside, and it determines whether 36.7% growth is a market or a subsidy. If circular AWS revenue is under ~$5B of $148B ttm, conviction goes to 7.0 and the verdict to ACCUMULATE on a pullback. If it is over ~$25B, conviction goes to 3.5 and the verdict to AVOID. No other disclosure has that range of outcomes.
Data-quality notes
- 🚩 Yahoo's 22.3x trailing P/E is unusable — ~45% of ttm GAAP EPS is a non-cash mark on an unsellable asset. Use 26.9x forward, or ~40.6x core trailing.
- 🚩 Yahoo's
operatingMarginsis the LATEST QUARTER, not ttm (13.689% = the Jun-2026 quarter; ttm is 12.1%). Verified against NOW and MU as same-day controls. - 🚩 Yahoo's statement tables lagged the Q2 print by a full quarter while the snapshot fields (EPS, P/E, revenue) were current — mixing them silently produces a one-quarter-stale ttm. The Jun-2026 quarter was derived by difference and reconciled exactly to the reported $200.6B revenue before use.
- 🚩 Yahoo's
insider_purchasesis false — see §4. Third instance today. - 🚩 Yahoo's
freeCashflowfield shows +$3.22B against Amazon's own −$7.6B; different definitions (finance-lease principal). Use the company's. - ⚠️ The ~$58B carrying-value reconciliation gap described above is unresolved and should be checked against the Q2 10-Q footnote.
- ⚠️ NA operating margin ex-one-offs is 7.3% using the verified $640M IEEPA refund; a second source suggested ~$1.2B of one-offs (implying ~6.8%). The conservative, sourced figure is used.
- ⚠️ Advertising and Kuiper margins are not disclosed; both are estimates.
Sources
Yahoo Finance MCP (quarterly statements, recommendations, holder info) and .mcp/fin.py, retrieved 2026-08-04 · Amazon Q2 2026 10-Q (SEC) · Amazon Q2 2026 earnings release · CNBC — capex to $220B · Q2 2026 call transcript · GeekWire — AWS booming, FCF negative · TechTimes — $496B backlog · Anthropic — Amazon 5GW compute · TechCrunch — Anthropic $5B/$100B · Bloomberg — additional $5B into Anthropic · Anthropic $1.2T secondary · Calcbench — Amazon server lifespans · Synergy via Computer Weekly — cloud +35% · Walmart Q1 FY27 · Shopify — agentic commerce · Marketplace Pulse — Temu/Shein · Bloomberg Law — FTC trial · FCC DA 26-553 — Kuiper waiver · TheNextWeb — EU DMA cloud gatekeepers · Benzinga — Jassy funding dodge · CNBC — Big Tech AI-stake gains distort earnings (2026-08-03)
Related agency knowledge: pitfall-unrealized-equity-marks-break-headline-pe · pitfall-yahoo-insider-purchases-counts-rsu-grants · principle-primary-source-beats-vendor · principle-down-a-lot-is-not-cheap · companion analysis GOOGL 2026-08-04 · MU 2026-08-04 (the memory cycle raising Amazon's capex)