AMZN Q2 2026 Earnings Analysis
Amazon reported Q2 revenue of $200.6B (+20% YoY) and operating income of $27.5B (+43% YoY), driven primarily by AWS acceleration to 36.7% growth with $496B backlog; 2026 CapEx raised to $220B.
Key Metrics
Puntos clave
- AWS revenue accelerated to 36.7% with backlog hitting $496 billion
- Operating income surged 43% YoY to $27.5B on efficiency gains
- 2026 CapEx raised to $220B; AI revenue run rate exceeded $25B
Audio not available for this episode.
Transcript is available below.
Escuchar en
Disponible en
Transcript
// Full episode scriptWelcome to Beta Finch, your AI-powered earnings breakdown, where we take the earnings call transcripts you don't have time to read and turn them into a conversation you can actually enjoy. I'm Alex.
And I'm Jordan. And Alex, before we dive into Amazon's Q2 numbers, which are honestly wild—
They are, but first — a quick disclaimer. This podcast is AI-generated content for educational and entertainment purposes only. Nothing we discuss should be considered investment advice. Always do your own research and consult a qualified financial advisor before making any investment decisions.
Good, glad that's out of the way, because we've got a lot to unpack.
Let's start with the headline numbers. Amazon posted $200.6 billion in revenue for Q2 2026, up 20% year-over-year. Operating income came in at $27.5 billion, up 43%. But the real story, as usual these days, is AWS.
AWS grew 36.7% year-over-year — and that's the fifth straight quarter of acceleration. Fastest growth in 18 quarters, back when the business was less than half its current size. They added $4.6 billion in revenue quarter-over-quarter, about 80% more than their biggest jump ever.
And the run rate — AWS is now annualizing at $169 billion. Andy Jassy pointed out that would rank 24th on the Fortune 500 as a standalone company. That's not a division anymore, that's a Fortune 25 company.
The backlog is the number that jumped out at me though — $496 billion, growing triple digits year-over-year. That's demand they haven't even fulfilled yet.
Right, and that ties into the CapEx story. Amazon bumped its 2026 cash CapEx guidance from about $200 billion to $220 billion, mostly because of higher memory chip costs. Jassy spent a good chunk of the call explaining the math behind that — data centers take about two years to build but then monetize for 30-plus years, while servers break even in under three years and run five to six years. It's basically Amazon saying "trust the ROIC math," even while free cash flow takes a hit in the near term.
And CFO Brian Olsavsky backed that up on margins — AWS operating income was $16.6 billion, a 39% operating margin. That's up 650 basis points year-over-year, or 520 if you strip out a one-time derivative accounting gain on energy contracts. An analyst on the call, Doug Anmuth, basically asked "shouldn't AI workloads be dragging margins down?" and the answer was: not so far, thanks to efficiency gains and capacity optimization.
Speaking of AI specifically — two numbers stood out. Amazon's chips business, that's Trainium and Graviton, now has an annual revenue run rate over $25 billion, growing triple digits. And separately, AI revenue overall also crossed $25 billion run rate. Both Anthropic and OpenAI have made multi-year, multi-gigawatt commitments to Trainium.
There was also a really interesting exchange about whether Amazon needs its own frontier model. Jassy's answer was basically "no, but we're building one anyway" — for cost control, prioritization, and speed, not because Bedrock needs it to survive. His view is there won't be one model to rule them all, there'll be half a dozen comparably good frontier models within a few years, and Amazon wants to be one of them while still hosting all the others in Bedrock.
There was also news that Amazon is exploring selling Trainium chips directly to third-party data centers, separate from AWS cloud. Jassy said there's "a real chance" that happens.
Let's talk Stores for a second, because it's easy to get AWS tunnel vision on these calls. Worldwide paid units grew 17%, North America revenue hit $116.2 billion, up 16%. Grocery is a big theme — Amazon says it's now the second-largest grocer in the U.S., with monthly active perishables customers up more than 50% since the start of the year.
And Amazon Now, their 30-minutes-or-less delivery service, expanded to 250-plus cities across 9 countries, with gross sales up over 80% quarter-over-quarter.
On the advertising side, Amazon Ads pulled in $19.8 billion, up 26% year-over-year, with sports content — NBA, Thursday Night Football, NASCAR — driving inventory sell-outs and higher engagement from multi-sport advertisers.
Now, guidance. Q3 net sales are guided between $197 and $202 billion, and operating income between $22.5 and $26.5 billion. That's a deceleration from Q2, but there's a clean explanation — Prime Day shifted into Q2 this year instead of Q3, which distorts the year-over-year comparison by about 400 basis points. There's also an 80 basis point FX headwind baked in.
So strip out the calendar noise and the underlying business trends look consistent with the first half of the year.
One more thing worth flagging from Q&A — an analyst asked about the backlog being two and a half times what it was a year ago, and whether that changes the capacity outlook beyond 2027. Jassy's answer was essentially: demand for 2028 is already "striking," and enterprises are still in the very early innings of running inference at scale. He described the AI adoption curve as "barbelled" right now — AI labs and a few breakout consumer apps on one end, cost-avoidance use cases on the other, and the massive middle of enterprise production workloads still mostly untapped.
That's the framing that matters for investors, I think — Amazon isn't saying "the AI boom is here." They're saying the big wave, the enterprise production middle, hasn't even arrived yet, and that's what's driving the CapEx.
So what does this mean going forward? AWS is accelerating, not decelerating, which is rare for a business this size. Margins have held up better than skeptics expected even with heavy AI investment. The CapEx number keeps climbing — now $220 billion for the year — which means free cash flow headwinds are likely to persist near-term, exactly as management flagged. And the Stores and Ads businesses are quietly compounding alongside all the AI headlines.
Before we wrap — everything discussed is AI-generated analysis for educational purposes. Past performance doesn't guarantee future results. Please do your own due diligence.
Big quarter, lots of moving pieces, and a company clearly betting that spending big now pays off for years. We'll be back next quarter to see how the CapEx-to-revenue math is tracking.
Thanks for listening to Beta Finch — see you next time.