Amazon backlog hits $465B

Amazon’s latest quarter was good on the surface. Revenue grew 17%. AWS accelerated again. Profit jumped. But the number that changed the conversation was not sales or earnings. It was backlog — the pile of future AWS revenue Amazon says is already under contract, plus a fresh Anthropic commitment that landed after the quarter closed. ### What is this $465 billion number? It is not a line item Amazon printed as “$465 billion” in the release. The company said AWS backlog was $364 billion at the end of Q1 2026. (ir.aboutamazon.com) Management and transcript coverage also made clear that this figure did not include a newly announced Anthropic commitment worth more than $100 billion over 10 years. Put those together and investors get to an implied figure a little above $465 billion. Basically, the headline number is a stitched-together view of contracted demand, not a single reported metric. ### Why did investors care so much? Because backlog is the cleanest answer to the capex question. Amazon is spending hard on AI infrastructure — land, power, buildings, networking, servers, and chips — long before all that capacity shows up as revenue. In the quarter, free cash flow over the trailing 12 months fell to $1.2 billion, driven mainly by a $59.3 billion year-over-year increase in property and equipment purchases. If you think demand is fuzzy, that spending looks scary. If you think demand is already contracted, the spending looks more like pulling supply forward. (theglobeandmail.com) ### How strong was AWS, really? Pretty strong by Amazon standards. AWS revenue rose 28% year over year to $37.6 billion, its fastest growth in 15 quarters, and segment operating income reached $14.2 billion. One detail bulls keep circling is that AWS added about $2 billion sequentially from Q4 to Q1 — a big move for a business already running at roughly $150 billion annualized revenue. That suggests AI demand is not just showing up in bookings; it is already landing in reported sales. (ir.aboutamazon.com) ### Where do Amazon’s own chips fit in? Right in the middle of the story. Andy Jassy said Amazon’s chips business topped a $20 billion revenue run rate and was growing triple digits year over year. That matters because Amazon is not just renting out generic cloud capacity. It is trying to own more of the stack — custom silicon like Trainium and Graviton, plus the cloud services wrapped around them. The upside is better economics and tighter customer lock-in. The catch is execution — custom chips only help if customers actually build around them at scale. (ir.aboutamazon.com) ### Why does Anthropic matter so much? Anthropic is not just another customer. It is a proof point. Amazon deepened its investment relationship with the AI company, and Anthropic in return committed to spend more than $100 billion on AWS over the next decade. That is the kind of agreement investors were looking for — something concrete enough to justify years of infrastructure build-out. It also helps explain why Amazon is comfortable spending ahead of revenue recognition. The customer is already there. (ir.aboutamazon.com) ### Is this only an AWS story? Not quite. Amazon’s ad business also kept growing — Jassy said advertising topped $70 billion in trailing-12-month revenue — and the retail side showed unusually strong unit growth. Those businesses matter because they help fund the build cycle. AWS is the center of the AI thesis, but Amazon’s broader machine is what lets the company spend at hyperscale without looking like a single-bet story. ### So what’s the real takeaway? (fool.com) The backlog number does not mean Amazon has solved every AI economics question. Revenue still has to arrive. Margins still have to hold. But the quarter made one thing clearer: this is not Amazon spending wildly and hoping demand appears later. It is Amazon spending against a very large book of contracted cloud demand — with Anthropic making that book look even bigger. (ir.aboutamazon.com)