Amazon

Amazon’s Q1 2026 print, reported April 29, broke records on almost every line and exposed one ratio that matters more than the rest. Net sales of $181.5 billion (+17% YoY) beat the Street’s $177 billion consensus, AWS grew 28% to $37.6 billion — its fastest pace in 15 quarters, on a $150 billion annualized run-rate — and total operating income rose 30% to $23.9 billion at a record 13.1% margin. Net income jumped to $30.3 billion ($2.78 per diluted share), but $16.8 billion of that came from a pre-tax mark-to-market gain on Amazon’s Anthropic stake, not operations. The catch: trailing-twelve-month free cash flow collapsed to roughly $1.2 billion (from $25.9 billion) as Q1 capex hit $43.2 billion.

What makes Amazon distinctive in mid-2026 is the AWS-Anthropic vertical integration. The expanded deal commits Anthropic to spend more than $100 billion on AWS over ten years and secures up to 5 gigawatts of Trainium capacity for training and serving Claude, while Amazon invests up to an additional $25 billion in the lab. Amazon’s in-house silicon — Graviton, Trainium and Nitro — now exceeds a $20 billion annualized run-rate growing at triple-digit rates. Advertising rose 24% to $17.2 billion. CEO Andy Jassy reaffirmed roughly $200 billion of calendar-2026 capex, the largest single-year capital commitment in Amazon’s history.

The Q1 2026 13F cycle reshuffled Amazon’s institutional base sharply. The headline: Berkshire Hathaway fully exited AMZN — a position originally initiated under Todd Combs in 2019 — in its first post-Buffett filing. At the same time, AMZN remained a consensus megacap, held by six of the seven most-watched funds. Bill Ackman’s Pershing Square is now a meaningful holder (a notable reversal), Third Point ran AMZN to a top portfolio weight, Bridgewater kept adding, and Seth Klarman’s Baupost lifted its stake from roughly 2.1 million to 3.1 million shares, with Tiger Global holding around 10 million shares.

Insider activity has been muted but coherent. Andy Jassy has made no notable open-market sales recently, and his last scheduled 10b5-1 execution was in late 2025. The most-watched insider event remains Jeff Bezos’ 10b5-1 plan disclosed in late 2024, which authorized up to 25 million shares of sales through 2026; Bezos has executed the bulk of that authorization, with the program winding down this year. The market has absorbed the distribution without sustained downward pressure — a signal of demand depth at the megacap-AI-exposure tier, even with Berkshire stepping away.

Amazon trades in the low-to-mid $250s in early June 2026, roughly 13% below the average analyst price target of around $313-319. Wall Street is overwhelmingly constructive: of roughly 45-66 covering firms the consensus is a Strong Buy, with targets spanning a low near $207 to a high around $370. On headline GAAP earnings the multiple looks full, but stripping the non-operating Anthropic gain and valuing AWS on an EV/cloud-revenue basis, Amazon screens cheaper than Microsoft’s Azure-implied multiple. The divergence resolves the day Amazon breaks out AI-specific revenue rather than burying it in the AWS line. The case turns asymmetric to the upside if AWS holds 28%+ into Q3 with stable ~38% segment margins, since the inference-layer take rate runs materially above legacy IaaS.

1. July 30, 2026 (after close): Q2 2026 earnings — Amazon guided net sales of $194-199 billion (+16-19%) and operating income of $20-24 billion; the key test is whether AWS sustains growth above 25% with segment margin holding near 38%, the first sign Trainium cost displacement is flowing into the income statement
2. Second half 2026: Project Rainier and Trainium ramp — full-scale revenue contribution from the Anthropic-dedicated Trainium cluster (including the Indiana site) and progress on the up-to-5 GW capacity build, the most concrete Trainium-vs-NVIDIA economics the market can observe
3. Through 2026: Tariff pass-through and Bezos 10b5-1 wind-down — watch whether seller price hikes pressure retail volumes, and whether the conclusion of the Bezos selling plan this year removes a structural overhang on share supply

Amazon is the cleanest AWS-AI compounder I track if you can look through the quarterly free-cash-flow chaos. AWS reaccelerating to 28% on a $150 billion run-rate, a record 13.1% total operating margin, expanding retail profitability, and the Anthropic-Trainium lock-in are arguably the strongest combined operating signals across the hyperscalers right now. The Berkshire exit caught headlines, but the rest of the smart-money tape — Pershing, Third Point, Baupost adding — tells the more interesting story. My add-trigger is any quarter where AWS grows above 25% AND North America margin holds or expands, signalling the ~$200 billion capex is funding revenue rather than depreciation drag. This is analysis, not a buy call: I would not chase AMZN on the Anthropic-inflated EPS, and the thesis weakens if AWS deceleration drops below 22% before fiscal 2027, when the free-cash-flow compression turns structural rather than cyclical.