AWS AI Revenue Proves the AI Spending Boom Is Finally Paying Off

• AWS AI revenue tops US$15 billion in Q1 2026, presents real returns.
• Amazon’s custom chip business doubles to US$20bn.
• Capacity constraints mean demand is still outrunning supply

For the better part of two years, the loudest criticism of AI investment has been the same: show me the revenue. Amazon CEO Andy Jassy had a straightforward answer this week. In his annual shareholder letter, he put AWS AI revenue at over US$15 billion in Q1 2026, noting it is “ascending rapidly,” and drawing a pointed comparison to AWS’s own early days, when the cloud business had a US$58 million revenue run rate three years after launch.

Three years into the AI wave, that figure is 260 times larger. The disclosure helped alleviate some investor concerns about the billions Amazon is deploying into AI-related initiatives. Following the release of the letter, Amazon shares rose 2% in pre-market trading.

The numbers landed at a moment when markets have been pressing hyperscalers hard on whether AI infrastructure spending will produce durable returns or simply represent the largest capital mis-allocation in tech history.

Jassy’s response to that was direct. “We’re not investing approximately US$200 billion in CAPEX in 2026 on a hunch,” he wrote.

The chip business nobody saw coming

The AWS AI revenue figure is significant. These custom chip numbers are arguably more so. Amazon’s chips business, which includes Graviton processors, Trainium AI chips and Nitro networking cards, now has an annualised revenue run rate of over US$20 billion, doubling from the US$10 billion disclosed with fourth-quarter results.

Jassy drew a direct line between this trajectory and what happened in CPU compute. Graviton, which offers up to 40% better price-performance than x86 processors, is now used by 98% of the top 1,000 EC2 customers. Trainium2 had about 30% better price-performance than comparable GPUs and has largely sold out. Trainium3, which started shipping at the start of 2026, is 30 to 40% more price-performant than Trainium2 and is nearly fully subscribed.

The demand picture is striking enough that two large AWS customers asked whether they could buy all of its custom Graviton capacity for 2026. Amazon declined – it still needs to provide CPU capacity to other customers – but Jassy cited it as a signal of where appetite sits.

Jassy also raised the prospect of selling chips externally. “There’s so much demand for our chips that it’s quite possible we’ll sell racks of them to third parties in the future,” he said. That would put Amazon in direct competition with Nvidia and AMD in a segment the company has so far only served internally.

What this means for the Asia Pacific

The AWS AI revenue milestone does not exist in isolation for the region. AWS has committed SG$12 billion, approximately US$9 billion, to Singapore between 2024 and 2028. Across Southeast Asia, Malaysia is gaining traction not only as an offshoot of Singapore but increasingly for local workloads as the government digitises the economy, while Indonesia’s cloud computing market has been growing at a CAGR of 48% over the last five years.

That regional build-out sits inside a broader constraint that Jassy acknowledged plainly. AI revenue “could be growing even faster,” but for capacity constraints currently facing the tech industry. AWS added 3.9 gigawatts of new power capacity in 2025 and plans to double total power capacity by the end of 2027.

Access to grid power remains the biggest constraint in the Asia Pacific, a challenge set to intensify in 2026 as data centre workloads change to more intensive AI-driven demand. For enterprises in the region building AI workloads on AWS, that capacity ceiling is not an abstract concern – it shapes availability, pricing and lead times on the services they are planning around.

The CAPEX argument

Amazon reiterated its plans to spend roughly US$200 billion in capital expenditure in 2026, with much of it expected to be monetised in 2027 and 2028, and already supported by customer commitments, including more than $100 billion from OpenAI.

The AI investment cycle has produced genuine scepticism among analysts and investors about whether the returns will match the spending. Brian Mulberry, chief market strategist at Zacks Investment Management, said: “The AI run-rate is a strong validation that AWS is successfully turning the AI boom into real, high-growth revenue. It’s still early days per Jassy, but the momentum positions AWS as a leader in AI infrastructure.”

Whether other hyperscalers can point to comparable proof points when they report will determine how long that scepticism lingers. For now, Amazon has put a number on what the AI time looks like as a revenue line. The rest of the industry is being asked to follow.

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. This comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.