On April 29, 2026, Amazon reported first-quarter results: net sales of $181.5 billion, up 17% year over year, with AWS revenue of $37.6 billion, up 28% — the segment’s fastest growth in fifteen quarters. Buried in the same release: trailing-twelve-month free cash flow of just $1.2 billion, down 95% from $25.9 billion a year earlier. Surging AI demand on one side, burning cash on the other — that is Amazon’s actual position in 2026.
The Q1 Numbers
- Total net sales of $181.5 billion (up 17%), with North America up 12% and rest-of-world up 19%
- Operating income of $23.9 billion at a 13.1% operating margin
- Net income of $30.3 billion and EPS of $2.78 — inflated by a $16.8 billion pre-tax gain on Amazon’s Anthropic stake; as one analyst put it, strip that out and the quarter looks “good” rather than “great”
- Q2 revenue guidance of $194–199 billion
Shares fell more than 3% right after the release, then turned to a gain of roughly 5% after hours once the earnings call wrapped.
The AI Engine Behind AWS
A 28% growth rate on a $37.6 billion quarterly base is the fastest since 2022. CEO Andy Jassy did not mince words on the call: “We’ve never seen a technology grow as rapidly as AI.” He added that “it’s very unusual for business to grow this fast on a base this large” — the last time AWS grew at this clip, it was roughly half the size.
Two figures explain the structural shift. First, AWS’s AI annualized revenue run rate now exceeds $15 billion — compare that to the $58 million run rate AWS itself posted three years after launching, a factor of roughly 260. Second, Trainium, Amazon’s homegrown silicon, has passed a $20 billion annual run rate; viewed standalone, the chip business would carry about a $50 billion run rate, which Jassy called “one of the top three data center chip businesses in the world.” He also signaled that full Trainium racks could be sold to external customers within the next couple of years, putting Amazon in direct competition with Nvidia. Graviton, the cheaper Arm-based line, is already used by 98% of Amazon’s top 1,000 EC2 customers with up to 40% better price-performance. As for OpenAI models landing on Bedrock, Jassy described the demand as “unprecedented.” AI is showing up inside the retail machine too: the Rufus shopping assistant grew monthly users 115% with engagement up 400% year over year, and an internal team used agentic coding tools to rebuild a Bedrock service with five people in 65 days — a project scoped at roughly 40 people for a year.
Capex Is Eating the Cash Flow
First-quarter capital expenditure hit $43.2 billion, against a full-year plan of roughly $200 billion set in February, with most of it aimed at AI infrastructure. Purchases of property and equipment ran $59.3 billion higher than a year ago, and that alone crushed trailing free cash flow to $1.2 billion. Amazon guided second-quarter operating income to $20–24 billion, slightly below the consensus midpoint — a reminder that the spend-and-build phase has years left to run.
Jassy’s Defense
Jassy did not dodge the pressure: “The faster AWS grows, the more short-term capex we’ll spend.” His argument is a timing arbitrage between asset life and revenue. Data centers last 30-plus years; chips, servers, and networking gear last five to six. Cash goes out first, revenue follows — “the early years, free cash flow is challenged” when capex growth outpaces revenue. He points to the first AWS buildout wave as proof the curve works: “We’ve been through this cycle with the first big AWS growth wave, and like the results.”
What It Means for Builders
First, AI capacity supply is expanding fast: a product built on Bedrock or Trainium should worry less about compute availability over the next two years and more about price competition and switching costs. Second, if Trainium racks really go on sale externally, the non-Nvidia camp gains a giant supplier — the same silicon boom is already pushing Cerebras back to its IPO filing, and developers’ chip menu keeps getting longer. Third, don’t read $1.2 billion of free cash flow as an emergency: Amazon deliberately separates paper profits (including the Anthropic gain) from cash flow in its narrative. The real test is whether AI revenue growth catches up with capex before investor patience runs out.
Sources
- Amazon’s cloud business is surging — and so is its capital spending — TechCrunch
- Amazon earnings recap: AWS comes through big as Q1 tops expectations — Business Insider
- Amazon (AMZN) Earnings Call: AI, AWS and CapEx Surge — The Globe and Mail
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.