Amazon introduces S3 Files to bridge AWS compute and S3 with native file system access

Amazon has unveiled Amazon S3 Files, a new file system that enables AWS compute resources to access Amazon Simple Storage Service (S3) buckets directly as file systems. This launch removes the previous tradeoff between S3’s durability and a file system’s interactive capabilities, transforming S3 into a centralized, interactive data hub for a range of AWS workloads.

S3 Files allows seamless, synchronized file system access to S3 data, automatically reflecting changes between the file system and underlying S3 buckets. The feature is available as a native mount on Amazon Elastic Compute Cloud (EC2), containers managed by Elastic Container Service (ECS) and Elastic Kubernetes Service (EKS), as well as AWS Lambda functions. File system access uses standard Network File System (NFS) version 4.1 or higher operations, including creating, reading, modifying, and deleting files.

Building on this integration, S3 Files makes use of high-performance storage to deliver approximately one millisecond latency for actively accessed data, with intelligent prefetching and fine-grained controls for data placement. For large sequential or byte-range reads, data is fetched directly from S3 to maximize throughput and contain costs.

These backend capabilities support concurrent access and NFS close-to-open consistency, which are essential for collaborative, mutable processes such as agentic artificial intelligence workflows and machine learning training pipelines.