AWS is printing money, but Amazon can’t stop spending it

Amazon reported its Q1 2026 earnings yesterday, and the headline numbers are what you’d expect from a company that’s been on a roll: AWS revenue came in at $28.7 billion, beating analyst estimates by about a billion. That’s 22% year-over-year growth, which is solid by any standard, though not quite the torrid pace we saw during the pandemic boom.

What caught my eye wasn’t the revenue number itself, but what CEO Andy Jassy said on the earnings call. He basically told investors to get comfortable with big capital expenditure numbers for the foreseeable future. Amazon’s CapEx hit $18.4 billion this quarter, and most of that is going into AWS infrastructure — data centers, networking gear, and the GPUs needed to run all those AI workloads everyone’s suddenly obsessed with.

This is where things get interesting. AWS is clearly the cash cow here, generating operating income of $9.8 billion, which is more than the entire company’s operating profit in some previous quarters. But the spending is also accelerating faster than I expected. Jassy specifically called out “generative AI” as a primary driver, which is corporate-speak for “we need to buy every Nvidia H100 we can get our hands on.”

The reality is that Amazon is in a weird position. It has to keep pouring money into infrastructure to support the AI gold rush, but there’s no guarantee that all those startups building on AWS will still be around in two years. Remember the crypto boom? Same thing happened with mining rigs. The difference this time is that the hyperscalers are building capacity for enterprise workloads too, which tend to be stickier.

Microsoft and Google are in the same boat, by the way. All three are spending like drunken sailors on AI infrastructure, and none of them can afford to slow down because the market is moving too fast. But Amazon has an advantage that its competitors don’t: AWS is already profitable at massive scale. Azure and Google Cloud are still playing catch-up on margins.

What I’m watching now is whether AWS can maintain its pricing discipline while all this capacity comes online. The temptation to cut prices to fill all those new servers will be enormous, and we’ve seen Amazon do that before. But if they can keep margins intact while spending like this, that’s a signal that the AI boom is the real deal, not just another hype cycle.

For AWS customers, this spending spree means more capacity and potentially lower prices down the road. But it also means Amazon is betting big on you sticking around. If you’re running workloads on AWS, you’re essentially piggybacking on the company’s bet that AI will be the next big platform shift. That’s either reassuring or terrifying, depending on your risk tolerance.