AWS AI $15B Run Rate

Amazon just put a real number on a business most cloud companies have described in foggy language: Amazon Web Services said its artificial intelligence services are running at more than $15 billion a year in the first quarter of 2026. That is the first time Amazon has publicly broken out a direct revenue figure for this part of the business. (reuters.com) (aboutamazon.com) A revenue run rate is a speedometer, not an odometer. It takes the current quarter’s pace and stretches it over 12 months, so Amazon is saying customers are already buying these tools at a pace above $15 billion a year, not that it has booked exactly $15 billion in cash so far. (reuters.com) (aboutamazon.com) Amazon Web Services is the part of Amazon that rents computing power the way a utility rents electricity. Companies use it to store data, run software, and now train and serve artificial intelligence models without building their own data centers. (aboutamazon.com) (amazon.com) That matters because investors have spent two years hearing that artificial intelligence would require giant spending on chips, power, and buildings before the money showed up. Andy Jassy used the new figure in his shareholder letter while defending roughly $200 billion in capital spending for 2026, much of it tied to artificial intelligence infrastructure. (cnbc.com) (aboutamazon.com) Amazon paired the artificial intelligence number with another one: its chips business is now running above $20 billion a year. Those chips include Graviton for general cloud workloads and Trainium for artificial intelligence training, which Amazon uses to lower its dependence on Nvidia’s processors and to sell cheaper computing to customers. (reuters.com) (aboutamazon.com) Jassy’s argument is that artificial intelligence does not sit in one neat box on a price list. A company that buys an artificial intelligence model from Amazon also rents storage, networking, databases, and security around it, so the $15 billion figure captures only part of the spending wave moving into Amazon Web Services. (aboutamazon.com) (constellationr.com) The comparison Amazon chose was revealing. Jassy said Amazon Web Services had a $58 million revenue run rate three years after it launched commercially, while its artificial intelligence business is above $15 billion three years into the current wave, which he said is nearly 260 times larger at the same stage. (aboutamazon.com) This also tells you where big companies are actually spending. Reuters reported that the disclosure suggests large customers are shifting meaningful budgets into cloud artificial intelligence services now, not waiting for some later payoff, and Amazon said demand is still high enough that capacity limits are holding back growth. (reuters.com) (qz.com) Amazon is not alone in chasing this market, but it is the first clear sign from one of the biggest cloud providers that enterprise artificial intelligence is already a large business rather than a science project. Once a service inside Amazon Web Services is running at a $15 billion annual pace, the question stops being whether companies will pay for cloud artificial intelligence and shifts to which provider captures the next $15 billion. (reuters.com) (cnbc.com)