Amazon Reveals AWS AI Business Surpasses $15 Billion Annualized Revenue Rate

Andy Jassy handed investors a scoreboard they had been waiting years to see. In his April 9 shareholder letter, the Amazon chief executive revealed that AWS's artificial intelligence business generated an annualized revenue run rate of more than $15 billion in the first quarter of 2026, marking the first time the company put a concrete financial figure on a business unit it has backed with billions of dollars in investment.

The $15 billion figure, based on first-quarter performance, represents roughly 10% of Amazon Web Services' $142 billion revenue run-rate and converts what many investors had treated as a long-term strategic bet into a measurable line of business. Amazon shares rose 2% in premarket trading following the disclosure.

Jassy did not stop at the AI services figure. He also revealed that Amazon's custom silicon business, which produces Graviton CPUs, Trainium AI training chips, and Nitro networking components, carries an annualized revenue run rate exceeding $20 billion. That figure elevates Amazon's in-house chip operation from an internal cost advantage into a revenue-generating unit with clear external ambitions. "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," Jassy wrote, a move that would put Amazon in direct competition with Nvidia and AMD in the high-performance AI hardware market.

Both disclosures served a precise purpose: defending the $200 billion capital expenditure plan Amazon projected for 2026, a figure that had rattled investors and stoked fears of an industry spending bubble. Amazon's free cash flow dropped from $38 billion to $11 billion last year, driven by a $50.7 billion increase in capital spending, primarily on AI infrastructure. "We're not investing on a hunch," Jassy said, noting that a substantial portion of the 2026 AWS capex already carries customer commitments set to be monetized in 2027 and 2028. Some customers had requested to secure all available capacity for certain AI chips, underscoring the intensity of competition for compute resources.

The $15 billion figure also sets a new benchmark in the hyperscaler arms race. Before the disclosure, Microsoft's AI revenue run rate of $13 billion was the highest publicly disclosed figure among the three major hyperscalers. Amazon's number now exceeds that mark and reframes the competition from investment posture to measurable revenue output.

Growth could be faster still. Jassy noted that AWS's cloud business as a whole would be growing even faster without the capacity constraints currently affecting the technology industry, a reference to power availability, datacenter construction timelines, and chip supply chains that continue to limit how quickly demand converts into recognized revenue.

The numbers Jassy disclosed signal that the hyperscaler AI arms race has moved visibly from the investment phase into productization. Whether Amazon can sustain the margins that would justify its infrastructure bet is the next question the numbers will have to answer.