Amazon Web Services just had its strongest quarter in nearly four years, and the AI boom is the reason. According to TechCrunch AI, AWS net sales jumped 28% year-over-year to $37.6 billion in Q1, the fastest growth rate the cloud unit has clocked in 15 quarters. CEO Andy Jassy made the call on the earnings stage Wednesday, framing AWS as the backbone supplier for the entire AI industry.
This is the picks-and-shovels thesis playing out in real numbers. While the spotlight stays on model labs and chatbots, the companies renting out compute are the ones cashing the checks. Amazon joined the list of tech giants that beat Wall Street’s Q1 expectations, and AWS is the engine doing most of the lifting.
The growth numbers Jassy wants you to notice
Jassy didn’t bury the comparison. He stacked the AI wave directly against AWS’s original launch trajectory:
- Three years after AWS launched, the business hit a $58 million revenue run rate.
- Three years into the current AI wave, AWS’s AI revenue run rate is over $15 billion.
- That’s roughly 260 times larger than the early-AWS curve.
“It’s very unusual for business to grow this fast on a base this large,” Jassy said during the call, as quoted by TechCrunch AI. “The last time we saw growth at this clip, AWS was roughly half the size. We’ve never seen a technology grow as rapidly as AI.”
What stands out here is the scale of the base. Hyperscalers usually slow down as they get bigger. AWS is doing the opposite, and Jassy is leaning hard into the message that customers keep picking Amazon for AI workloads.
The capex bill is enormous
The other half of the story is what it costs to keep that growth going. Amazon is pouring cash into land, power, buildings, chips, servers, and networking gear, and Jassy was direct that the spending will keep climbing in the near term.
“The faster AWS grows, the more short-term capex we’ll spend,” he said. “AWS has to lay out cash for land, power, buildings, chips, servers, and networking gear, in advance of when we can monetize it.”
The pull on free cash flow is severe. Per TechCrunch AI:
- Free cash flow dropped to $1.2 billion for the trailing twelve months.
- That’s a 95% drop from the $25.9 billion Amazon reported in Q1 2025.
- The driver: a $59.3 billion year-over-year increase in property and equipment purchases, much of it tied to AI infrastructure.
Jassy positioned the spend as a long-life investment. Data centers last 30-plus years. Chips, servers, and networking gear have a five-to-six-year useful life. He pointed back to the original AWS buildout cycle as a template, saying “we’ve been through this cycle with the first big AWS growth wave, and like the results.”
Why this matters for the AI industry
A few signals for practitioners and operators to track:
- Compute supply is still the bottleneck. AWS wouldn’t be torching free cash flow on infrastructure if demand were softening. Expect tight GPU and capacity availability to continue.
- The cost curve isn’t bending yet. Amazon’s willingness to absorb a 95% free cash flow hit signals confidence that AI workloads will keep filling capacity faster than it gets built.
- Hyperscaler concentration deepens. When AWS, Microsoft, and Google all spend at this clip, smaller cloud providers and neoclouds face a tougher fight on scale economics.
- Investor patience is the new variable. Jassy’s comment that “the early years, free cash flow is challenged” is a preview of how every hyperscaler will frame margin pressure for the next several quarters.
Amazon’s overall sales rose 17% to $181.5 billion, with North America up 12% and international up 19%. The retail business is healthy, but AWS is doing the heavy lifting on the AI narrative.
The takeaway: the AI infrastructure arms race is still in its capital-intensive build phase, and the hyperscalers that can stomach the cash burn are pulling further ahead. Full earnings details and Jassy’s full commentary are at the original TechCrunch AI report.