TL;DR
TL;DR
At $274.48 (settled close, 2026-08-07) — −3.4% off its $284.02 52-week high (2026-08-03), +23.0% above the level a year ago ($223.13 on 2025-08-07 — a real re-rating), and +21.1% above the pre-Q2-print close ($226.65 on 2026-07-29) — Amazon trades at a reconciled ~3.82× funded EV/TTM-revenue that looks cheap only until it is split by segment. The whole debate is one question: does the ~$200B/yr AI/AWS capex convert to free cash flow — now negative and getting deeper — while AWS holds the +36.8%/39.4% it just printed, before the market re-anchors AWS off its premium-to-peers cloud multiple, against a dated antitrust/DMA tail.
- Bull. Revenue is re-accelerating (Q2 +19.6%, up from Q1's +16.6%) with both margins expanding (gross +44bps to 52.26%, operating +226bps to 13.69%) and operating income +43.2%; AWS reaccelerated to $42,232M, +36.8% ("fastest in 18 quarters") at a 39.4% segment margin that expanded (32.9%→37.7%→39.4%) — the opposite of Q1's YoY dip — and is ~61% of consolidated operating income on ~21% of revenue; Advertising is a $76,072M-TTM, +26% high-margin line the blend buries. You pay a market cloud multiple for the #1-share hyperscaler that just beat its guide — but the bull is explicitly quality/growth, not "cheap," because the SOTP is fair, not a bargain.
- Bear. Concede all of that — the bear needs none of it to be false. The buildout consumes all the cash and then some: Q2 capex $54,208M exceeded operating cash flow $45,387M, so Q2 FCF was −$8,821M (a second consecutive negative quarter) and TTM FCF is −$11,625M (gross) / −$7.6B (Amazon-net) — deeper than a quarter ago; D&A $75,200M is only 43.5% of gross capex, so less than half the spend has hit the P&L. Funded debt roughly doubled to $133,320M and Amazon is now −$10,332M funded net debt (−$120,103M lease-inclusive with $109,771M of leases); capital return is $0 buyback, $0 dividend — no floor. Two-thirds of Q2 pre-tax income is the $53.4 billion non-operating Anthropic mark (NI exceeds OI by $35,186M after tax), not operations.
- Priced-in break-even. At the funded EV, marking the $627,276M non-AWS business at Walmart's 1.32× implies AWS is already priced at ~14.4× revenue — above MSFT (11.34×) and GOOGL (9.45×). To justify that at a peer multiple, AWS forward revenue must reach $188.3B–$225.9B (≈+27% to +52%) off the $148,404M TTM base — the just-printed +36.8% clears the MSFT threshold but not GOOGL's, so on next-twelve-month revenue the AWS EV is 10.5× — a peer NTM multiple, no margin of safety. The multiple is "paid for" iff AWS holds ≈+27–37% growth at a held ~39% margin and capex/revenue normalizes below 20.8% so trailing FCF re-turns positive — against the uncompensated EU-DMA / FTC-antitrust tail. If any leg fails the downside band is −17.6% to −38.1% (bear SOTP). No buy/sell/hold — research tooling, not advice.
Watch next — a cold reader can act on these:
| Window | Event | If it lands bullish | If it lands bearish |
|---|---|---|---|
| ~late Oct 2026 (est. 2026-10-22, Wall Street Horizon, confirmed / ~2026-10-29 MarketChameleon, unconfirmed; accessed 2026-08-09; Amazon IR not yet posted; cadence from the 2025-10-30 Q3'25 print) | Q3 FY2026 earnings print — the load-bearing forward catalyst (the Q2 print resolved 2026-07-30, +15.3%) | Net sales at/above the $197–202B guide; AWS growth ~mid-30s% with margin ~39%; OI at/above the $22.5B low end; capex "digesting" so TTM FCF troughs toward positive | Revenue/OI at the low end or a miss; AWS decel or margin rollover as depreciation lands; capex guide raised past the ~$173B TTM run-rate with TTM FCF still negative; the $53.4B Anthropic non-op gain does not recur |
| ~Feb 2027 (Q4 FY2026 print) | Trailing-FCF conversion — definitive read on capex/revenue vs the 20.8% break-even | TTM FCF re-turns positive as capex/revenue normalizes below 20.8% — the "capex converts to ROIC" thesis is confirmed | TTM FCF stays negative with the ~$200B FY2026 capex figure held/raised — the spend is not converting; no capital-return cushion underneath |
| Final decision ~late Oct 2026 (written reps due Sept 2026; preliminary 2026-06-25; EC IP/26/1444, The Register, accessed 2026-08-09) | EU DMA — AWS cloud-gatekeeper designation (first cloud designation; fines to 10% of global turnover) | AWS declined/narrowed, or obligations prove immaterial to AWS economics | AWS itself designated → interoperability / anti-lock-in duties on the segment that is ~61% of operating income |
| Bench trial 2027-03-29 (pretrial 2027-03-15; MLex/Law360, accessed 2026-08-09) | FTC v. Amazon marketplace-monopoly antitrust trial (FTC + 18 state AGs + PR, Judge Chun, W.D. Wash.) | A defense verdict or conduct-only remedy leaves marketplace/advertising/AWS economics intact — the tail is removed | A liability finding → structural remedy on the retail/ads economics the SOTP marks at Walmart parity — the −38.1% (Bear-3) case |