Amazon S3 Files and AI spend

Amazon just changed a quiet rule of cloud computing: on April 7, 2026, Amazon Web Services said teams can open an Amazon Simple Storage Service bucket like a normal folder tree instead of treating it like a pile of web objects. That product is called Amazon S3 Files. (aws.amazon.com) Amazon Simple Storage Service, usually shortened to Amazon S3, has long been the cheap giant warehouse for data, but many programs want a file cabinet with folders, locks, and rename commands. That mismatch is why companies kept copying the same data into separate storage systems just so older software could read it. (aws.amazon.com) Amazon S3 Files is Amazon’s attempt to remove that copy step. Amazon says the data stays in Amazon S3 while compute services see it with full file-system behavior and about 1 millisecond latency. (aws.amazon.com) Amazon’s documentation says the service is built using Amazon Elastic File System, which is Amazon’s managed network drive, but pointed at data already stored in Amazon S3. That means a machine learning training job, a container, or a script can use familiar file commands without exporting the bucket somewhere else first. (docs.aws.amazon.com) Amazon is aiming this at teams that run containers on Kubernetes, build data pipelines, or train artificial intelligence models on giant datasets. Those jobs often break when one tool speaks “files” and the storage layer speaks “objects,” so engineers spend time building glue code instead of shipping products. (aws.amazon.com) The timing is not random. In a note published April 9, 2026, Amazon chief executive Andy Jassy said Amazon Web Services’ artificial intelligence revenue run rate is now above $15 billion in the first quarter of 2026 and still rising. (aboutamazon.com) That number helps explain why Amazon is smoothing out storage, because artificial intelligence workloads are hungry for one thing above all else: fast access to lots of data without extra handling. If every model team has to duplicate petabytes into a second system before training starts, cloud bills and delays both climb. (aboutamazon.com) (aws.amazon.com) Amazon is also spending at a scale that matches that push. In Amazon’s February 5, 2026 fourth-quarter results, the company said 2025 capital expenditures were $83 billion, and Jassy said he expects 2026 spending to be “reasonably representative” of that annualized fourth-quarter pace, with most of it going to artificial intelligence for Amazon Web Services. (ir.aboutamazon.com) (aboutamazon.com) Around that spending wave, vendors are selling the picks and shovels. StackGen said on April 10, 2026 that it is using Amazon Web Services Summit events in Bengaluru, Mumbai, and New York City to demo an artificial intelligence agent that generates Terraform infrastructure, detects drift, triages incidents, and enforces compliance policies. (stackgen.com) Put those pieces together and the pattern is simple: Amazon is trying to make storage look simpler to the application, while partners try to make infrastructure look simpler to the operator. If both layers get easier at the same time, more artificial intelligence projects reach production before platform teams become the bottleneck. (aws.amazon.com) (stackgen.com)