For years, Amazon has asked Wall Street to be patient. The spending is heavy, the timeline is long, and the payoff is coming. Trust the process.
Andy Jassy’s annual shareholder letter, published April 9, lands a bit differently. This time, he’s not really asking for patience—more like handing out receipts. In the letter, Jassy put hard numbers around Amazon’s AI and chip businesses for the first time, and the market responded by pushing the stock higher. In a way, it’s the oldest trick in finance, just with better paperwork: clarify the trajectory, and suddenly people feel like they can model it.
The biggest headline for investors is AWS AI revenue. Jassy said the AWS AI revenue run rate is now above $15 billion as of Q1 2026, and—importantly—it’s the first time Amazon has disclosed this figure. That matters because it turns a promise into something you can point at. Investors who’ve been waiting for proof now have a concrete number to chew on.
Zoom out and the scale gets even more telling. AWS’s total annualized revenue run rate stands at approximately $142 billion, and that implies AI services already represent roughly 10% of the entire AWS business. Jassy also described demand as “ascending rapidly.” It’s one of those phrases that sounds generic until you pair it with the infrastructure reality Amazon keeps running into: demand is outpacing supply. AWS added 3.9 gigawatts of new power capacity in 2025 and expects to double its total power capacity by the end of 2027, but some customers are still unable to get the compute they want. The whole thing feels familiar—like hearing the hum of a crowded data center at night and realizing it’s not just lights, it’s capacity constraints.
Then there’s the chip business, where Jassy’s language is blunt. He disclosed that Amazon’s chip business, including Graviton, Trainium, and Nitro, now has an annual revenue run rate above $20 billion, growing at triple-digit rates year-over-year. Jassy said that chip business doubled from Q4 2025, and he wrote: “Our chips business is on fire,” adding that it “changes the economics for AWS, and will be much larger than most think.” The demand signals are just as aggressive: Trainium2 has largely sold out; Trainium3, started shipping in early 2026, is nearly fully subscribed; and Trainium4, still about 18 months from broad availability, has already been significantly reserved. Two large customers asked to buy all of Amazon’s available Graviton chip capacity for 2026, and Graviton is now used by 98% of the top 1,000 EC2 customers.
The letter also tries to shut down a question investors have been circling for a while: is Amazon’s huge spending on AI and chips just a gamble? Jassy said the company plans to spend approximately $200 billion in capital expenditures in 2026, mostly tied to AI, chips, and infrastructure, and he argued it’s not “on a hunch.” Existing commitments, he said, cover “a substantial portion” of the capex spend, and Amazon expects to monetize most of it in 2027 and 2028. One commitment alone is a deal with OpenAI worth more than $100 billion. He also made the case that if Amazon sold its chips externally the way Nvidia does, the business would be running at roughly $50 billion in annual revenue.
So what does it mean for Amazon stock? Analysts at Jefferies said the letter made them more constructive on Amazon, reiterating a buy rating and $300 price target. They noted that Trainium and Graviton are becoming a “structural advantage” for AWS. That’s the point: even if it doesn’t instantly change the income statement, clarity can change sentiment. Amazon may still be in the middle of a heavy spending cycle, sure—but Jassy just gave investors a stronger picture of where AI is already monetizing. And once you see the numbers laid out like this… you get the sense the payoff might be closer than “eventually.” Or maybe that’s just what the market wants to believe right now, and we’ll see.