Amazon [AMZN]
Task 6 · built 2026-07-27 · spot $231.39 · updated 2026-07-29 — no position verdict is issued; this is a dated event calendar only
Every entry is dated and carries an explicit upgrade/downgrade threshold. Ongoing maintenance hands off to
equity-research:catalyst-calendar and equity-research:thesis-tracker.
Updated 2026-07-29 — Criteria + two-horizon valuation. The numbered Gates are retired and replaced by named Criteria with types (BINDING / MEASURED) returning PASS / FAIL / INDETERMINATE. This memo no longer issues a position verdict — no Long, Short, Watchlist or Avoid. The expected-return-versus-cash-hurdle test is replaced by a reverse-DCF implied-path test plus a 12-month target, with sensitivity run over the exit multiple rather than over scenario probabilities. Momentum is entry timing only and vetoes nothing. Sections below that predate this update are retained as the historical record and are annotated where superseded.
Reading the retired "Gate N" numbering. Any
Gate Nbelow is historical. The mapping is: Gate 1 → Quality Criteria (BINDING) · Gate 1b → Short Mechanism Criteria (MEASURED) · Gate 2 / 2A / 2B / Path B → dissolved, absorbed by the Valuation Criteria · Gate 2C → Peer Spread Criteria (MEASURED) · Gate 3 → Catalyst Criteria (MEASURED) · Gate 4 / 4a → Valuation Criteria (BINDING) · Gate 5 → Liquidity Criteria (BINDING) · Gate 6 → Momentum Criteria (MEASURED, entry timing only) · (new) Downside Criteria (MEASURED). A Criteria returns PASS / FAIL / INDETERMINATE; a missing input is INDETERMINATE, never FAIL. Historical calibration items keep the old numbering on purpose — the record is the record.
Consensus: revenue $196.18bn (range $187.98–200.10bn, 46 analysts); EPS $1.82 (range $1.56–$2.10); AWS ~$40.5bn; North America ~$113.8bn.
| Metric | Watch for | Threshold and action |
|---|---|---|
| AWS revenue growth | 28% in Q1-2026 | < 20% with FY capex guidance still > $200bn → trigger E1, first short condition |
| > 30% → the AI spend is visibly earning; raise the bull weight from 20% | ||
| North America segment operating margin | 6.95% FY2025 | Contracting YoY would materially worsen the analysis — it would confirm the FY2025 gain was substantially the heavy-equipment life extension. Recorded as a test, not as a verdict change. |
| Reported vs operating EPS | $16.8bn Anthropic gain in Q1-2026 | Any new mark: restate EPS excluding it before comparing to consensus. Any downward mark → trigger E3 |
| Capex / PP&E additions | $142.4bn FY2025; run-rating ~$200bn | > $230bn implied FY2026 → depreciation model shifts up a full year |
| Depreciation commentary | Servers at 5–6 yrs | Any further EXTENSION of AI server life → downgrade to AVOID. Reverses the quality signal |
| Free cash flow (TTM) | $1.2bn | Negative TTM FCF → the first time in Amazon's modern history; a genuine regime marker |
| AWS RPO / backlog | Not currently disclosed | First-ever disclosure → materially changes the analysis. Would make the regime-change test applicable for the first time |
IV context: 21-Aug ATM IV is 44.6% against 31.4% realised — 1.42x. The options market prices a 9.3% move. No position is taken into this print precisely because the house has no view that the move exceeds what is priced.
Merger agreement signed 13-Apr-2026; ~$10.9bn including debt; cash/stock election with a 40% cash cap; accompanying agreements with Apple. Watch: regulatory clearance, the $110m operational-milestone adjustment, and whether Amazon begins capitalising Amazon Leo costs. Trigger: the first quarter in which Leo costs are capitalised rather than expensed produces a one-time, non-recurring operating-income benefit of unknown size — restate before comparing to consensus.
Six show-cause orders issued to all RTOs under FPA Section 206. Responses and subsequent rulemaking through H2 2026. Trigger: any ruling that materially restricts co-located or behind-the-meter data-centre load would constrain the ~9 GW implied-penetration requirement and is a genuine bear catalyst — the first one that would operate on physics rather than on sentiment.
The stated reassessment date for this name. Same metric grid as Q2, plus: - Whether FY2027 capex guidance is given, and at what level. - Whether 30-day FY2027 EPS revisions have turned net negative (trigger E2; currently +2.8% over 90 days).
The largest single swing factor in reported EPS and entirely outside Amazon's reporting calendar. A new round at a higher valuation adds billions of non-operating income; a down-round reverses it. Trigger E3. No dated event exists — which is itself a reason the earnings quality is low.
Amazon disclosed a >$20bn combined chips run-rate (Graviton + Trainium + Nitro) growing triple digits in Q1-2026. Trigger: any Trainium-specific revenue disclosure, or a named third-party Trainium customer outside Anthropic, would be the first hard evidence that Amazon's silicon is competitive outside its own captive demand.
Modelled D&A reaches $96.1bn (FY2028) and $119.9bn (FY2029) against EBITDA of $205.0bn and $231.9bn. This is arithmetic, not forecast. The crossover point at which D&A growth exceeds EBITDA growth — modelled at FY2027 — is when reported operating margin begins visibly falling. Watch each quarter for it arriving early.
$16.7bn of principal and interest due 2028, $203.5bn total to 2076. Amazon issued $53.8bn in March 2026 at 3.85%–6.05%. Trigger: any issuance at a materially wider spread signals the market repricing hyperscaler capex risk — a sector-level early-warning indicator, not just an Amazon one.
The implied-penetration statement requires roughly 9 GW of net new, fully-monetised AI capacity. Amazon added ~3 GW in FY2025 (derived from $96.5bn of AWS capex). Checkable annually against AWS PP&E additions in the 10-K R-files (R92), since Amazon will not publish GW directly.
| Trigger | Date | Effect |
|---|---|---|
| E1 — AWS growth < 20% with capex guide > $200bn | 30-Jul-26, then quarterly | First short condition |
| E2 — 30d FY2027 EPS revisions net negative | rolling | Second short condition |
| E3 — Anthropic down-round / mark reversal | undated | Third short condition |
| T1 — weekly close below the 200dma (~$234.6) with a lower-high sequence | rolling | Required tape condition — no short without it |
| L1 — price ≤ $150 with AWS growth ≥ 20% | rolling | Long conversion |
| L2 — first-ever AWS RPO / backlog disclosure | 10-K or any quarter | Long conversion; makes the regime-change test applicable |
| A1 — retail segment margin contracting YoY | 30-Jul-26 | Would materially worsen the analysis |
| A2 — further extension of AI server useful life | 10-K, ~Feb-27 | Would materially worsen the analysis |
Short conversion requires (E1 or E2 or E3) AND T1. The tape condition is mandatory because Gate 6 is currently neutral, and initiating against a neutral tape with no variant is the exact error this workflow exists to prevent.