AWS is printing money, but Amazon is spending it just as fast

Amazon dropped its Q1 2026 earnings today, and the headline is familiar: AWS is still the cash cow, but the company can’t stop pouring money into infrastructure.

AWS revenue came in at $28.3 billion for the quarter, up 22% year-over-year and a bit above what analysts were expecting. That’s solid, especially when you consider how much competition is breathing down Amazon’s neck — Microsoft and Google are both aggressive on the AI cloud front, and neither is showing signs of letting up.

But here’s the thing that stood out to me: capital expenditures hit $18.2 billion in Q1 alone. That’s up from $14.6 billion in the same quarter last year. Andy Jassy made it clear on the earnings call that this isn’t a one-time spike. He said the company expects to spend even more in the coming quarters, largely driven by AI-related infrastructure — data centers, networking gear, and the specialized chips needed to train and run large models.

I get why they’re doing it. Every major cloud provider is in an arms race right now. If you don’t build capacity, your customers will go elsewhere. But $18 billion in a single quarter is eye-watering, even for a company with Amazon’s margins. The question nobody seems to want to answer directly is when this spending starts to pay off in a way that justifies the scale.

Jassy did mention that AWS’s AI services are growing at a triple-digit percentage rate year-over-year, which is impressive. But triple-digit growth off a small base still leaves you with a small number in absolute terms. The real money in cloud is still in compute and storage, not AI inference or training — at least for now.

One thing I appreciated: Jassy didn’t sugarcoat the competitive landscape. He acknowledged that some customers are experimenting with multiple cloud providers for AI workloads, which is a polite way of saying that AWS isn’t the automatic default anymore. Microsoft’s deep integration with OpenAI and Google’s strength in foundational models are real threats.

I also noticed that Amazon didn’t break out how much of its capex is specifically for AI versus general cloud expansion. That’s a deliberate choice. If they showed the number and it was too high, investors might balk. Too low, and they’d look like they’re falling behind. So we get the combined figure and a vague “it’s mostly AI” explanation.

On the positive side, AWS margins held up well. Operating margin for the cloud segment was 31%, which is healthy. That tells me the price wars haven’t hit AWS as hard as some feared. The hyperscalers still have pricing power, at least for now.

What I’m watching for next quarter: whether AWS can maintain that growth rate without cutting prices further, and whether the capex-to-revenue ratio starts to improve. If they keep spending $18B+ per quarter but revenue growth stays in the low 20s, that math gets ugly fast.

For now, Amazon is betting big that AI will be the next wave of cloud computing, and they’re willing to spend whatever it takes to be the default provider. It’s a high-stakes game, and we won’t know who wins for a few more quarters. But I’d rather see a company over-invest in infrastructure than under-invest and lose customers. At least this way, they control their own fate.