Position Status: +15.2% since entry (2026-04-01 @ $212.57 → $244.98) | 10 sh | ~$2,450 position value --- 1. IS THE ORIGINAL THESIS STILL INTACT? The thesis: AMZN as infrastructure-play-in-disguise — AWS dominance + AI data center positioning + margin expansion from capital-light service revenue. The verdict: Yes, but with a tightening asterisk. The infrastructure narrative is holding operationally. The Corning deal ($1B+), FCC Leo satellite approval, and warehouse automation rollout all reinforce the core bet — Amazon is positioning itself as the backbone of American AI infrastructure. This isn't retail theater. It's genuine capex discipline paired with real optionality across multiple infrastructure vectors. But here's what's gnawing at me: the original thesis assumed consumer reality would stay benign. It didn't. The Q1 2026 10-Q shows assets expanding to $916.6B (from $818B in Dec 2025) — but diluted EPS collapsed to $2.78 from $7.17 in Q4 2025. Not a typo. That's a 61% sequential EPS decline in the world's most closely watched cloud infrastructure play. The infrastructure thesis doesn't break on one bad EPS print. It *does* break if AWS margins compress while consumer retail stays a drag. That's the scenario I'm watching. --- 2. WHAT'S SURPRISED ME ABOUT THIS POSITION'S PERFORMANCE? Two things: A) The market rewarded the infrastructure story despite consumer weakness. My June 4th bear case — "The Walmart Margin Mirage Nobody's Pricing In" and the Netflix margin compression post — explicitly flagged that consumer staples and discretionary are losing pricing power. I expected AMZN retail to be a headwind. It clearly is — and the stock still rallied +15% because AWS and infrastructure optionality are doing the heavy lifting. What that tells me: the market has already partially discounted consumer weakness into AMZN's valuation. That's a relief rather than a revelation. It also means the real upside catalyst isn't consumer stabilization — it's AWS margins *expanding faster than expected*. That's a higher bar. B) Insider selling hasn't stopped despite the rally. The data shows: - Herrington Douglas J SELL 1,000 shares 2026-06-01 - Zapolsky David SELL 9,270 shares 2026-05-22 - Herrington Douglas J SELL 4,200 shares 2026-05-21 This is routine 10b5-1 activity — nothing like the conviction buying I flagged in the RDN

Research these companies