AWS revenue growth has accelerated for five straight quarters to 36.7% YoY with a $496B backlog and record 39.4% margins — yet Amazon raised its 2026 capital expenditure plan to a staggering $220B, flipping trailing free cash flow negative (-$7.6B) and doubling long-term debt to $128.9B. Six analyst lenses, three scenarios, four time horizons.
Gray line = Amazon’s actual price trajectory into Oct 13, 2026 ($196.00 52-week low on Feb 17, 2026 → $287.20 all-time high on Aug 3, 2026 post-Q2 earnings → $258.58 today); colored paths = synthesized scenario midpoints forward, probability-weighted (base 50% · bull 25% · bear 25%). Mid-year marks. Wall Street 12-month consensus ≈ $330 (range $175–$425, “Buy” from 58 of 60 analysts).
Those probabilities are a judgment call — so make them yours. Drag to set how likely the bear (AI capex overbuild) and bull ($1T AWS monetization) cases are; the blended target below, the dotted line on the chart, and the prob-weighted row of the scenario cards all update live.
Each desk below evaluates Amazon through an independent institutional framework before reconciling into our scenario cone. Notice where they collide: the P&L has never looked better, while the cash flow statement is absorbing the largest capital buildout in corporate history.
AWS revenue accelerated for five consecutive quarters (17% → 20% → 24% → 28% → 36.7% in Q2’26) off a $169B annualized run-rate, while AWS operating margin hit 39.4%. GenAI and custom silicon (Trainium2/Inferentia) each crossed $25B annualized run-rates. Demand is supply-constrained with a $496B backlog.
Operating cash flow rose 33% to $161.4B TTM, but net capex exploded to $169.0B TTM (and ~$220B guided for FY26), flipping TTM free cash flow to -$7.6B and doubling long-term debt from $65.6B to $128.9B in six months. Also strips out the $53.4B non-cash Anthropic revaluation from Q2 GAAP earnings ($5.75 GAAP vs $1.97 clean EPS).
Hyperscalers are trapped in a prisoner’s dilemma spending $220B/yr on 4–5 year depreciating silicon and inflated HBM memory. As $150B+ of annual AI capex enters the depreciation schedule in 2027–2028, AWS margins will compress 800+ bps just as open-weights models commoditize inference pricing and FTC remedies target 3P retail take rates.
Amazon owns three interlocking toll roads: regionalized North America fulfillment (driving retail EBIT to $9.1B in Q2), a $19.8B/quarter (+26% YoY) high-margin Advertising layer at the point of purchase, and AWS custom silicon offering 30–40% better price-performance than merchant GPUs. Competitors cannot replicate the physical + cloud footprint.
At $258.58, AMZN trades at 16.8× EV/EBITDA — 57% below its 10-year median of 27.2× — and 24.6× FY27E clean EPS despite 95th-percentile upward estimate revisions. Shares are coiling between the 50-day SMA ($256.93) and 200-day SMA ($241.51) after digesting the Aug 3 peak ($287.20), though low accrual scores (22/100) cap breakouts until Q3 FCF clarity.
Amazon’s $20B capex hike (from $200B to $220B) came almost entirely from HBM/DRAM memory price inflation, while AWS growth is gated by utility grid interconnects rather than customer demand. Currently monetizing at ~$8 per incremental watt; scaling higher-margin AI inference workloads toward $10+/watt unlocks massive operating leverage once power shells come online.
What the sell-side expects over the next year (as of Oct 13, 2026). Bars are sorted low to high across selected major desks; the dashed line marks today’s $258.58 — 58 of 60 analysts rate AMZN a Buy, with only D.A. Davidson sitting below spot on capex/FCF caution.
Published 12-month sell-side targets as of October 2026 (including October updates from Tigress Financial [$385], TD Cowen [$350], and Rosenblatt [$360], and post-Q2 resets from Morgan Stanley [$335 base / $500 bull], JPMorgan [$365], Wells Fargo [$329], and D.A. Davidson [$250 Neutral]). Across 60–74 tracked analysts, the mean/median consensus is $329–$331 (+28% upside).
Synthesized scenario midpoints (mid-year). Returns shown vs. today’s $258.58. These are illustrative frameworks, not predictions — five-year outcomes hinge on whether $220B/yr of capex converts into $100B+ of annual free cash flow by 2029–2031.
Where the money actually goes — and why the market is debating AMZN. Watch the clay bar (capex) quadruple from $52.7B in 2023 to ~$220.0B in 2026E, flipping the olive bar (free cash flow) negative while the slate bar (long-term debt) doubles to $128.9B.
Sources: Amazon FY23–FY25 Form 10-K, Q2 2026 Form 10-Q (ended June 30, 2026), and management FY26 capex guidance. Capex consumed 62% of operating cash flow in 2023 ($52.7B vs $85.0B OCF), rose to 94% in 2025 ($131.8B vs $139.5B OCF), and exceeds 100% of OCF in trailing Q2’26 ($169.0B net capex vs $161.4B OCF = -$7.6B TTM FCF) and FY26E (~$220B guided capex vs ~$198B projected OCF = -$22B FCF). Long-term debt jumped from $65.6B at year-end 2025 to $128.9B in Q2’26 (backed by $123.0B in cash & marketable securities).
Our scenario targets multiply clean operating EPS by an exit multiple. Note the dashed outline in 2026E: GAAP EPS ($12.89 consensus) is inflated by a one-time $53.4B (~$4.95/share) non-operating accounting gain from Amazon’s Anthropic stake revaluation in Q2’26; the solid bars track clean operating EPS.
*In Q2 2026, Amazon booked a $53.4B pre-tax non-operating valuation gain on its Anthropic stake, lifting Q2 GAAP diluted EPS to $5.75 (vs. $1.97 clean operating EPS) and full-year 2026 GAAP EPS consensus to $12.89. Serious institutional models value AMZN off clean operating EPS ($8.31 in 2026E and $10.50 in 2027E, compounding toward ~$22.50 by 2031E in the base case). At 26× $22.50, the base case yields $585.
Q2 FY26 year-over-year growth across Amazon’s retail, cloud, advertising, and AI silicon lines — showing why management is willing to spend $220B to avoid leaving cloud demand on the table.
At an $800B+ revenue scale, posting +20% consolidated top-line growth and +43% operating income growth is extraordinary. AWS (+36.7% YoY) and Advertising (+26% YoY) together generate over 75% of Amazon’s operating profit, while GenAI services and custom Trainium/Inferentia silicon each passed $25B annualized revenue run-rates in Q2 2026.
The entire valuation argument compresses into one disagreement: is the $220B capex bill locking in a high-margin cloud monopoly, or destroying free cash flow in an AI arms race?
Eight institutional risks mapped across a 3–5 year horizon. The hottest risk sits in the upper-right corner: multi-year FCF compression if $220B/yr capex persists without commensurate $/watt pricing power.
What breaks: Capex stays at $220B+/yr into 2027–2028 while OCF lags, keeping free cash flow negative and forcing continuous debt issuance above $150B.
What breaks: Simultaneous hyperscaler buildouts (AWS, Azure, GCP, Oracle) flood the market with GPU/ASIC capacity, collapsing $/watt monetization below $8.
What breaks: Utility transformer and transmission bottlenecks leave billions of dollars of purchased AI servers sitting unenergized, delaying $496B backlog conversion.
What breaks: Surging HBM3e/HBM4 memory prices — which already added $20B to FY26 capex — erode hardware ROIC without adding incremental compute units.
What breaks: Putting $350B+ of 2025–2026 capex onto a 5-year server depreciation schedule compresses AWS EBIT margins from 39.4% back toward 30%.
What breaks: An adverse federal court remedy forces Amazon to decouple Prime shipping from 3P Buy Box placement, structurally lowering retail take rates.
What breaks: Launching 3,200+ low-Earth-orbit satellites requires $15B–$20B of capital with slow subscriber payback against SpaceX Starlink’s lead.
What breaks: Temu, Shein, and TikTok Shop force permanent fee discounts on Amazon Haul and lower-ticket Apparel/Home third-party merchants.
What breaks: Shifting Prime Day into June boosted Q2’26 (+20% YoY) at the expense of Q3’26 (+9–12% guide), creating optical deceleration headlines on Oct 29.
Hover the dotted terms in the metrics strip, or scan the desk’s working definitions here.