Skip to content
orbyd

Dossier · AMZN · Dormant

AMZN

MEDIUM a1Compounder

Last analysed · · source: theme_discovery

Current thesis

AWS reacceleration (+28% YoY Q1, 15-qtr high) + Trainium silicon royalty is fundamentally ACCELERATING, but the tape won''t confirm — AMZN sold off on a blowout print over ~$200B capex and has chopped near its 50-day MA for six weeks. New wildcard: Anthropic hired IPO banks (2026-06-03), setting up a possible mark-up of Amazon''s stake. No fresh momentum trigger; next AWS binary is the ~2026-07-30 Q2 print.

Invalidation trigger

Weekly close below the 200-day MA (~$229), OR Q2 print (~2026-07-30) shows AWS net-sales growth decelerating below +25% YoY — either one kills the reacceleration leg and turns this into a value-trap mega-cap.

Thesis status

Open commitment scored if the trigger above fires How this is scored →

Current Thesis

The leg we'd be buying is unchanged from last refresh: AWS reacceleration + Trainium custom-silicon royalty + an emerging Anthropic-stake mark-up wildcard. Q1 2026 (reported 2026-04-29) printed AWS net sales of $37.6B, +28% YoY — the fastest in 15 quarters — on a ~$150B run-rate, with AWS AI revenue at a >$15B run-rate and record 13.1% company operating margin. The fundamental narrative is genuinely ACCELERATING. The problem is the same one we flagged on 2026-05-24: the tape will not confirm. AMZN sold off on a blowout print, the market fixated on ~$200B FY26 capex + Jassy's memory-cost warning, and the stock has been chopping around its 50-day MA for six weeks rather than breaking out. We'd be buying a story, not a fresh momentum trigger. New since last week: Anthropic hired Morgan Stanley/Goldman to lead its IPO (2026-06-03), which sets up a potential value-crystallization event for Amazon's multi-billion Anthropic stake — but also feeds the "IPO liquidity drain" macro worry, and rate-hike fears are back (2026-06-03).

Bull Case

  • AWS +28% YoY, 15-quarter high (Q1, 2026-04-29) — the reacceleration bears said wouldn't come, on a $150B run-rate base. Size AND acceleration is the rarest mega-cap combo.
  • AI run-rate >$15B and compounding (Q1 call, 2026-04-29) via Bedrock + Trainium; management says a substantial chunk of capex is pre-committed by customers.
  • Enterprise AWS momentum is still landing deals: Autodesk expanded onto AWS with a joint AI-development + marketplace agreement (2026-06-03) — concrete evidence the pipeline the +28% rests on is still filling.
  • Trainium = picks-and-shovels royalty. Anthropic's >$100B AWS/Trainium commitment (2026-04-20) underwrites up to 5GW, ~1GW online by year-end. Anthropic's ecosystem push — Claude Partner Network Services Track (2026-06-03), Project Glasswing expansion (2026-06-02) — drives more Claude-on-Bedrock consumption straight into AWS.
  • Anthropic IPO = stake mark-up optionality. Banks hired 2026-06-03; an Anthropic public listing would crystallize Amazon's equity stake at a marked valuation, a non-AWS catalyst the market isn't pricing.
  • Scale milestone: Amazon ended Walmart's 13-year run as the #1 S&P 500 / Fortune 500 revenue name for 2025 (2026-06-03) — narrative reinforcement of platform dominance.

Bear Case

  • Capex is the swing factor and the tape hates it. ~$200B FY26 capex, $43.2B in Q1 alone; the 2026-04-29 down-move on a beat is the warning shot — every quarter capex outruns AWS monetization, the multiple compresses.
  • Jassy's memory-cost warning (Q1 call) — component/memory prices "skyrocketing," capacity-constrained — directly threatens the AWS-margin and ROIC story bulls underwrite.
  • Macro turned against multiples: S&P 500 / Nasdaq 100 fell from records on hotter data reviving Fed rate-hike bets (2026-06-03). A $2.7T name with a premium multiple is pure duration risk into that.
  • IPO liquidity drain: SpaceX/OpenAI/Anthropic listings could pull capital out of mega-cap holders (2026-06-02); AI-capex sentiment is already fragile (Burry shorting NVDA, 2026-06-02).
  • Mega-cap velocity ceiling. Even a perfect narrative doesn't produce the parabolic leg this playbook hunts. AWS-reaccel is a grind-higher trade, not a squeeze.

Setup & Price Structure

  • No fresh price context delivered this refresh. Last known print (2026-05-22) ~$267.57, range ~$255–268. Treat levels below as last-known, not live — re-pull before sizing.
  • Chopping around the 50-day MA (~$256), well above the 200-day MA (~$229). Uptrend intact but stalled — not extended, not blown-off, and critically not breaking out.
  • Post-earnings character remains the tell: beat on every line and faded. Six weeks of sideways consolidation near the 50-day is digestion, not accumulation thrust. This is a MATURING tape on an ACCELERATING fundamental — the two haven't reconciled.
  • Entry discipline: This is not a clean momentum trigger. A LOW probe is the most this setup justifies until either (a) a weekly close back above the post-earnings high reasserts trend, or (b) the Q2 print re-fires the AWS reaccel. No chase here.

Catalyst Calendar (next 30 days)

  • ~2026-06-10 (est.) — May CPI: hot data feeds the rate-hike narrative that hit mega-caps on 2026-06-03; directly moves the multiple.
  • ~2026-06-17 (est.) — FOMC decision: macro, not AMZN-specific, but the dominant near-term driver of mega-cap risk appetite.
  • Anthropic IPO process — ongoing: banks hired 2026-06-03; roadshow/pricing timing unconfirmed (likely H2 2026), watch for a dated S-1/roadshow that would mark Amazon's stake.
  • ~2026-07-08–16 (est.) — Prime Day: retail-margin read-through, just outside the 30-day window.
  • ~2026-07-30 (est.) — Q2 2026 earnings: the real AWS-reaccel binary. Outside 30d but it's the entry-forcing event. No AMZN-specific binary lands inside the next 30 days — this is a watch, not an entry window.

What Would Change Our Mind

  • Upgrade to HIGH/buy: weekly close above the post-earnings high (reclaim ~$268+ with the 50-day flipping to support) on AWS/AI-flow confirmation → momentum finally aligns with fundamentals.
  • Confirm thesis: Q2 (2026-07-30) AWS growth holds ≥+28% YoY or accelerates, with capex guide flat-to-down → multiple re-rates.
  • Invalidate / exit: weekly close below the 200-day MA (~$229), OR Q2 AWS growth decelerates below +25% YoY → the reacceleration leg is dead, this becomes a value-trap mega-cap.
  • Stand aside: sustained macro tightening (Fed actually hikes) compresses every premium multiple regardless of AWS fundamentals.

Correlation Notes

  • Mag-7 / AI-capex beta: AMZN moves with NVDA/MSFT/GOOGL on any AI-capex sentiment swing. The Burry-vs-Eisman NVDA short debate (2026-06-02) is the sentiment barometer — an NVDA capex-doubt unwind hits AMZN regardless of its own print.
  • Rate-sensitivity: behaves as long-duration mega-cap; the 2026-06-03 rate-hike-fear selloff is the template for macro drawdown risk.
  • Anthropic linkage (two-way): Amazon is long Anthropic via both equity stake and the $100B compute deal — an Anthropic IPO is upside optionality, but Anthropic/OpenAI/SpaceX listings collectively threaten to drain market liquidity (2026-06-02).
  • Hyperscaler cluster: trades alongside MSFT (Azure) and GOOGL (GCP); a single hyperscaler's cloud-growth miss re-rates the whole cluster's AWS-reaccel premium.