AAMZN
    Stock Analysis

    Why I Decided To Make Amazon One Of My Biggest Investments

    Why I Decided To Make Amazon One Of My Biggest Investments
    • Purchased AMZN despite failing the firm’s TOLL+M framework (63/100) to gain largest global cloud exposure and diversify a portfolio concentrated in toll-like, supplier-side assets
    • AWS 2Q26: revenue +36.7% YoY, added ~$4.6bn QoQ, $169bn annualized run rate; AI revenue and chips businesses each >$25bn ARR; operating income $16.6bn at 39.4% margin, +650bps YoY
    • AWS backlog $496bn and remaining performance obligations 2.5x 3Q25, with much AI capacity contracted for >=5 years - supports long-duration, predictable revenue streams
    • 2026 EPS includes an estimated ~$3.44 per share boost from a ~$50.5bn unrealized Anthropic markup and tax items, so the consensus 2027 EPS "decline" is largely an accounting and depreciation timing effect
    • Consensus CapEx ramp to $217bn (2026), $269bn (2027), $292bn (2028) pushes FCF negative through 2027 then sharply positive in 2028-29; key risk remains sustained input inflation that raises Amazon’s costs
    Leo Nelissen
    Sep 21, 202612:27 PM110

    Introduction

    I bought Amazon (AMZN).

    That sentence took me years longer to write than it should have, and the reason is highly uncomfortable, as Amazon fails my own framework. I'm talking about the TOLL+M framework, of course.

    It fails it badly. It also pays no dividend, and it is about to spend more on CapEx in a single year than the entire market value of most companies in the S&P 500.

    Also, it sits on the wrong side of almost every trade I have built this portfolio around.

    Yet, I bought it anyway, and that's not because Amazon has something called "Amazon Leo," which I think sounds really cool.

    In this article, I'll walk through why Amazon is not a TOLL stock and why that turned out to be the entire point. I'll go through what 2Q26 actually showed once you strip out the accounting noise, and there is a staggering amount of it.

    I'll show you why the 2027 consensus earnings "decline" everyone is quoting is not really an earnings decline. I'll explain what happens to Amazon's cash generation between now and 2029. And I'll be honest about the three things that would make me wrong, including one where my own macro thesis works against my own position.

    So, let's dive in!

    Keep Reading

    Create a free account to access more content.

    Sign InCreate Free Account

    Sign in to leave a comment and join the discussion.

    Sign Up Free