Generative AI is a “really unusually large, maybe once-in-a-lifetime type of opportunity,” Amazon CEO Andy Jassy told investors
With hyperscalers planning massive investments in AI infrastructure, Amazon is working to further capitalize on what CEO Andy Jassy called a “really unusually large, maybe once-in-a-lifetime type of opportunity.” In a separate call with investors, Jassy pegged Q4 capex at $26.3 billion and said, “I think that is reasonably representative of what you expect [of] an annualized capex rate in 2025…The vast majority of that capex spend is on AI for AWS,” according to reporting from CNBC.
Discussing how Amazon Web Services (AWS) is delivering AI-enabled services, along with the underlying AI infrastructure down to homegrown custom silicon for model training and inference, Jassy said, “AWS’s AI business is a multi-billion dollar revenue run rate business that continues to grow at a triple-digit year-over-year percentage, and is growing more than three times faster at this stage of its evolution as AWS itself grew. And we felt like AWS grew pretty quickly.”
AWS reported net sales of $28.8 billion in Q4, up 19% year-over-year, and operating income of $10.6 billion, up from $7.2 billion in Q4 2023. For the full-year 2024, AWS saw net sales of $107.6 billion, up 19% year-over-year, and operating income of $39.8 billion, up from $24.6 billion in 2023. Full financials are avialable here.
Digging into its work with customers to support AI, Jassy essentially ran through AWS’s approach to the full stack of AI infrastructure “with each layer being a giant opportunity and each is progressing rapidly.” The “bottom layer” he said support AI model building with access to NVIDIA H200 GPUs; Jassy also called out the firm’s “networking innovations” like Elastic Fiber Adapter for inter-node communications and Nitro, a lightweight hypervisor.
Beyond its “deep partnership with NVIDIA” on GPUs for AI infrastructure computing, Jassy referenced its custom chips for training, called Tranium, and for inference, called Inferentia. AWS customers “want better price/performance on their AI workloads,” he said. “As customers approach higher scale in their implementations, they realize quickly that AI can get costly.”
He also highlighted Amazon SageMaker to support AI workflows from data prep through training and optimization then production deployment. Specifically, Jassy mentioned AWS’s “unique HyperPod capability, which automatically splits training workloads across more than 1,000 AI accelerators, prevents interruptions by periodically saving checkpoints, and automatically repairing faulty instances from their last saved checkpoint.” In sum, he said, this reduces training time up to 40%.
For customers looking to use existing models and tune them using proprietary datasets, Jassy touted Amazon Bedrock as having “the broadest selection of leading foundation models and most compelling modules for key capabilities like model evaluation, guardrails, RAG and agents.”
Detailing capex plans, CFO Brian Olsavsky noted that capex figures include cash outlay and equipment finance leases. “The majority of the spend is to support the growing need for technology infrastructure. This primarily relates to AWS as we invest to support demand for our AI services while also including technology infrastructure to support our North America and international segments.” Importantly, the total figure also supports non-AWS needs for Amazon’s fulfillment and transportation operation.
Asked about demand for AI services as compared to AWS’s capacity to deliver AI services, Jassy said he believes AWS has “more demand that we could fulfill if we had even more capacity today. I think pretty much everyone today has less capacity than they have demand for, and it’s really primarily chips that are the area where companies could use more supply.” See earlier detail on AWS’s custom silicon.
“I actually believe that the rate of growth there has a chance to improve over time as we have bigger and bigger capacity,” Jassy said.
