Quarter: Q4 FY25 (Dec 2025) | Reported: Feb 5, 2026 Revenue: $213.4B | YoY: +13.6% ^^ | Beat: +1.85% vs guide mid AWS: $35.6B | YoY: +24% ^^!! | ARR: $142B Op Income: $25.0B (GAAP) / 27.4B(adj.ex−2.4B charges) EPS: $1.95 | TTM FCF: 11.2B|* * MarketCap : ** 2,200B
Here's the thing... Amazon is the most comprehensive "scale in platform" story in the public markets today. IMHO, nobody else comes close in terms of sheer breadth of BUILDING BLOCKS that compound on each other.
$142B ARR cloud platform — THE infrastructure layer the internet runs on. Hundreds of services that are building blocks other companies build ON. Every AWS service makes every other AWS service more valuable (flywheel). Self-serve GTM — you can spin up an EC2 instance with a credit card. No SI dependency. Green flag.
This is the one the market is underappreciating. Amazon is building its own chips — PICKS AND SHOVELS for the AI era — INSIDE their own cloud platform:
DING DING DING — Azure and GCP buy NVIDIA GPUs at market price. Amazon makes chips delivering 30-40% better price-performance, giving customers better prices AND Amazon better economics. STRUCTURAL margin advantage that compounds yearly. The Graviton/Trainium flywheel is the most underappreciated competitive moat in cloud.
AWS self-serve. Advertising self-serve. Same-day delivery to ~100M customers. Prime convenience lock-in. All green flags.
Platform verdict: Best of breed. The purest "scale in platform" at mega-cap scale.
| Q1 FY25 | Q2 FY25 | Q3 FY25 | Q4 FY25 | |
|---|---|---|---|---|
| Revenue ($B) | $155.7 | $167.7 | $180.2 | $213.4 |
| YoY % | 8.7% | 13.3% | 13.4% | 13.6% ^^ |
| Beat vs Guide | +1.60% | +3.84% | +1.93% | +1.85% |
| QoQ Add ($B) | -- | +$12.0 | +$12.5 | +$33.2 |
FY25 Total: $716.9B (+12.4% YoY). 8.7% -> 13.3% -> 13.4% -> 13.6% = clear re-acceleration ^^
| Segment | Revenue | YoY % | Op Margin | Signal |
|---|---|---|---|---|
| AWS | $35.6B | +24% ^^!! | 35.0% | Fastest growth in 13 quarters |
| Advertising | $21.3B | +23% !! | High | 3rd largest digital ad platform |
| North America | $127.1B | +10% | 9.0% (record) ^^ | Logistics flywheel |
| International | $50.7B | +17% | 2.1%* | Italy charge depressed |
| Subscriptions | $13.1B | +14% | -- | Prime stable |
| 3P Seller | $52.8B | +11% | -- | 61% of WW paid units |
*International ex-charges expanded YoY.
| Q1 FY25 | Q2 FY25 | Q3 FY25 | Q4 FY25 | |
|---|---|---|---|---|
| AWS Revenue ($B) | $29.3 | $30.9 | $33.0 | $35.6 |
| AWS YoY % | ~17% | ~19% | ~20% | 24% ^^!! |
| QoQ Add ($B) | +$0.5 | +$1.6 | +$2.1 | +$2.6 ^^ |
| AWS Op Margin | 39.5% | 32.9% | 34.6% | 35.0% |
| Metric | FY24 | FY25 | YoY |
|---|---|---|---|
| Net Income | $59.2B | $77.7B | +31.3% !! |
| EPS | $5.53 | $7.17 | +29.7% !! |
| Operating CF | $115.9B | $139.5B | +20% !! |
| CapEx (net) | $77.7B | $128.3B | +65% |
| FCF (TTM) | $38.2B | $11.2B | -71% |
| SBC % Rev (Q4) | 2.7% | 2.1% | -60bps ^^ |
NA Op Margin 9.0% = RECORD ^^. FCF compression from deliberate capex, NOT operational deterioration.
Cash + Securities: $123.0B | LT Debt: 65.6B|NetCash : **57.4B** (fortress)
Q1 FY26 guide: Revenue $176.0B mid (+13% YoY). Op Income 19.0Bmid.FY26CapEx: 200B.
"As fast as we install this AI capacity, we are monetizing it. It's just a very unusual opportunity." — Jassy
"Trainium is a multibillion dollar annualized run rate business at this point, and it's fully subscribed." — Jassy
"I'm not sure folks realize how strong a chips company we've become over the last 10 years." — Jassy
"The lion's share of [AI] demand is still yet to come in the middle of that barbell." — Jassy
"Companies that have the excellence in infrastructure and the components that give better price performance are going to have advantaged financials." — Jassy
1. AWS QoQ acceleration is the real story. Forget the 24% YoY headline. QoQ adds: $0.5B -> $1.6B -> $2.1B -> $2.6B. The rate of acceleration is itself accelerating. AWS added $7.0B YoY in Q4 alone on $142B ARR. Mathematical signature of a new S-curve — AI layering onto cloud migration.
2. $244B backlog growing 40% vs revenue 24% = 16pp bullish divergence. Demand building faster than revenue recognition. Capacity-constrained growth. Backlog = 1.7x annual AWS revenue. Signed contracts, not speculation.
3. Custom silicon flywheel is the most underappreciated moat. Trainium under Bedrock: customers get 30-40% better price-performance (win), Amazon gets better unit economics (win). EXACT same pattern as early logistics investment. Azure and GCP buy GPUs at market price. Amazon makes its own. 3-5 year compounding structural margin advantage.
4. $200B capex looks scary until you see the cash structure. Operating CF 139.5B.DeltaaboveOCF= 60B. Cash + securities = $123B. Net cash $57B. Funded for 2+ years without stress. Operating income growing +17% YoY expands the funding base.
5. No FCF floor = genuine concern. Two analysts asked (Mahaney, Anmuth). Two deflections. No circuit breaker if AI demand disappoints. Trust but verify — watch backlog conversion rate every quarter.
6. Agentic shopping ad disruption is sleeper risk. Morton (MoffettNathanson) asked about funnel compression. Jassy pivoted to Rufus advantages (300M users, 60% higher purchase completion). Compelling BUT did not quantify impact on $69B advertising. 5% compression = $3.5B. Monitoring.
7. Connect $1B ARR, 30%+ growth = hidden building block. Enterprise contact center platform, 20M+ daily interactions. Deepens AWS enterprise stickiness, barely mentioned in coverage.
| Metric | Prior Belief | Actual | Verdict |
|---|---|---|---|
| Revenue | ~$210B | $213.4B | Beat — 7th consecutive |
| AWS growth | 21-22% | 24% | Beat ^^ — fastest 13Q |
| AWS margin | 33-34% | 35.0% | Beat — stable |
| Custom chips | Growing | >$10B, triple-digit | Beat !!! |
| Advertising | 20-22% | 23% | Beat |
| NA margin | 8-8.5% | 9.0% record | Beat ^^ |
| FY26 capex | $150-180B | ~$200B | Above |
P/E 28.4x | P/OCF 15.8x (cheap vs MSFT 28x, GOOG 20x) | PEG 0.93 on 30% EPS growth
**SOTP ~3, 200B * *(alignedwithAtlas).AWSaloneat14xrevenue= 2.0T = ~90% of market cap. Advertising, silicon, retail essentially free at current price.
| Engine | Growth | Tier |
|---|---|---|
| Custom chips | Triple-digit on $10B | Tier 1 !!! |
| Bedrock AI | +60% QoQ | Tier 1 !!! |
| AWS | 24% on $142B | Tier 2 |
| Advertising | 23% on $69B | Tier 2 |
| Retail/Subs | 10-14% | Tier 3 |
Consolidated 13.6% = Tier 3 by numbers. Quality of growth = Tier 2 caliber.
Strong hold / Watchlist add for new positions.
Amazon delivered the most compelling cloud quarter I've seen from ANY company at scale. AWS re-accelerating to 24% on $142B ARR while building a custom silicon platform that delivers 30-40% better price-performance — that is scale in platform at its finest. $244B backlog growing 40% = not a one-quarter blip; multi-year AI infrastructure build.
Platform assessment: best of breed on every dimension. Building blocks, self-serve GTM, crowdsourced intelligence, cloud-native, convenience over price.
Why watchlist? My portfolio is for smaller, faster-growing companies where platform optionality is still being priced in. Amazon's excellence is well-known. At $2.2T, the return profile fits large-cap allocation, not concentrated hypergrowth.
For existing holders: ADD on weakness. SOTP ~40-50% upside. PEG 0.93. AWS + silicon + ads = three compounding engines driving 25-30% EPS growth for 2-3 years.
Watch: Jassy's non-answer on FCF guardrails. Trust but verify. If $244B backlog decelerates while capex stays $200B, thesis changes. Right now? Demand signals unambiguous.
Watchlist. Strong conviction for holders.
long AMZN 0% (watchlist — not held in concentrated hypergrowth portfolio)
-muji