Amazon just had its best single day in over a decade — the stock jumped roughly 15% on July 31, and the headline number driving it was a net income of $62.6 billion, nearly triple what the company posted a year earlier. If you only read the headline, you'd think Amazon's core business just had a historic breakout quarter.
It didn't, exactly. And the gap between the headline and the reality is worth understanding before you decide what this rally actually means.
The $53.4 Billion Number Nobody's Framing Correctly
Buried in Amazon's Q2 2026 report is a detail that explains most of that tripled profit: $53.4 billion of pre-tax income came from a mark-to-market gain on Amazon's stake in Anthropic. That's an accounting revaluation of an investment, not revenue Amazon earned by selling anything. Net income for the same quarter last year was $18.2 billion — strip out the Anthropic markup and this quarter's operating reality looks a lot closer to that baseline than to $62.6 billion.
This isn't a one-off oddity, either. It's the same mechanic that's been distorting Big Tech earnings all year — I flagged a similar accounting quirk in
Meta's Free Cash Flow Just Fell 91%. The Real Story Is an Accounting Trick, and the pattern here rhymes: a genuinely strong quarter gets a headline number inflated (or in Meta's case, deflated) by something that has nothing to do with operations.
What Actually Happened at AWS — and It's Genuinely Good
Here's the part that doesn't need an asterisk. AWS revenue hit $42.2 billion, growing 37% year-over-year — the fastest pace in 18 quarters — and AWS operating margin expanded from 32.9% to 39.4%, a 650 basis point jump that pushed AWS operating income to $16.6 billion, up 64% from a year ago.
That margin expansion matters because it directly answers a question I raised in
Amazon Jumped 14%. Alphabet Fell 6% on the Same Capex Story: does massive AI infrastructure spending actually pay off, or does it just compress margins while the market waits for a payoff that never arrives? For AWS, at least this quarter, the answer leaned toward "it's paying off" — AI services and Amazon's custom chip business each crossed $25 billion in annualized revenue run rate, both growing at triple-digit rates. AWS's backlog also climbed to roughly $496 billion.
The Catch: Capex Is Still Climbing Faster Than Comfort
None of this comes free. Amazon raised its 2026 capex guidance by $20 billion to roughly $220 billion, and trailing-twelve-month property and equipment purchases are already up 64% year-over-year to $169 billion. Management attributed part of the increase to input inflation — pricier memory and components — rather than pure capacity expansion, which is its own yellow flag given how tight AI hardware supply has been all year.
So you have two things happening at once: a business (AWS) that's proving the AI infrastructure bet can widen margins, and a spending trajectory that keeps accelerating with no clear ceiling in sight. Both can be true, and the market rewarded the first while mostly ignoring the second on earnings day.
Why the Mark-to-Market Gain Actually Matters for Your Read on the Stock
It's tempting to wave off the Anthropic gain as noise since it's non-cash and non-operating. But it's not irrelevant — it tells you two things. First, Amazon's early, unglamorous bet on Anthropic (originally a $2.7 billion-plus commitment) is now worth enough that a single quarter's revaluation can swing net income by tens of billions of dollars. Second, and more important for risk management: that same mechanism runs in reverse. If Anthropic's private valuation gets marked down in a future quarter — the same way public AI-adjacent stocks have swung wildly this year — Amazon's headline net income could just as easily post a shock decline that has nothing to do with retail, ads, or AWS actually deteriorating.
Seeking Alpha's read on the quarter makes a similar point: the AWS margin story is real and durable, but investors conflating that with the net income headline are pricing in more certainty than the numbers actually support.
So Is the 15% Pop Justified?
Partially. The AWS numbers — 37% growth, 39.4% margin, a $496 billion backlog — are the kind of results that deserve a real re-rating, and they're the part of this report that should actually move a long-term thesis. The Anthropic gain, on the other hand, is a paper event that happens to have landed in the same quarter and made the headline number look far more dramatic than the underlying business improvement.
My honest read: the stock move is roughly half genuine operating strength and half a market that doesn't always slow down to separate operating income from investment marks on earnings day. If AWS keeps expanding margins at this rate for another quarter or two while capex growth stays contained, the re-rating holds up on fundamentals alone — you won't need the Anthropic gain to justify the multiple. If AWS margin expansion stalls, or if capex guidance gets raised again without a clear demand story attached, that's the signal the market got ahead of itself.
What would change my mind here is simple: watch AWS operating margin next quarter. A repeat of 35%+ margins on continued 30%+ growth is a real signal. A pullback toward last year's 33% level, paired with capex still climbing, would suggest the AI infrastructure bet is starting to cost more than it's returning — right as the market has already priced in the good outcome.
This is not financial advice — always do your own research before making investment decisions.
The One Thing to Actually Check
Before you read anything else into this rally, open Amazon's next 10-Q and go straight to the "other income" line, not the net income headline. If that line stays volatile while AWS operating margin keeps climbing independently, you're looking at a genuinely strong business wrapped in a noisy accounting number — and that's a very different stock story than a one-quarter fluke.
댓글
댓글 쓰기