Amazon’s Jassy says AI revenue run rate is over $15 Bln, hints at future chip sale

Amazon CEO Andy Jassy revealed during the company's quarterly earnings call on April 30, 2024, that the annual revenue run rate for its artificial intelligence services has surpassed $15 billion, a significant milestone that underscores the rapid growth of its cloud and AI divisions. The announcement, made from Amazon's headquarters in Seattle, was framed within the context of intense competition in the generative AI market, where Amazon Web Services (AWS) is vying for dominance against rivals like Microsoft Azure and Google Cloud. The $15 billion run rate represents the projected annual revenue based on current performance, highlighting how quickly AI has become a major revenue pillar for AWS. This growth is primarily driven by demand for Amazon's Bedrock service, which allows developers to build applications using various foundation models, and its custom AI chips, Trainium and Inferentia. Jassy emphasized that while AWS has historically relied on Nvidia's GPUs, its own silicon is becoming increasingly competitive, offering cost-effective alternatives for training and running AI models. In a notable forward-looking statement, Jassy hinted that Amazon might eventually sell its proprietary AI chips to other cloud providers or enterprises directly. This strategic move would mark a significant shift, positioning Amazon not just as a cloud service consumer of chips but as a direct competitor in the semiconductor market against Nvidia and AMD. The suggestion indicates Amazon's confidence in its chip technology and its ambition to capture a larger share of the entire AI infrastructure value chain, from silicon to software services. The announcement reflects a broader industry trend where major tech companies are vertically integrating to control their AI destiny. Amazon's massive investment in custom silicon aims to reduce dependency, lower costs for customers, and create a more differentiated and sticky ecosystem within AWS. This financial disclosure and strategic hint are seen as a direct response to similar AI revenue boasts from competitors, signaling that the battle for cloud and AI supremacy will be fought on both the software and hardware fronts in the coming years.