AWS has unveiled S3 Files, a novel file system that lets you access Amazon S3 buckets as if they were a conventional file storage space, without moving data out of S3. The promise is to combine the durability and cost-effectiveness of object storage with the interactivity of a file system—a trade-off that AI and data engineering teams have been trying to narrow for years.
A bridge between object storage and files
Until now, architectures often had to choose between object storage and file storage depending on the expected access patterns. With S3 Files, AWS says it eliminates this split by exposing S3 data through a shared file layer, reachable from any AWS compute resource, whether it be instances, containers, or functions.
The service maintains a view of the objects present in the bucket and translates file operations into optimized S3 requests. AWS also claims that hot data benefits from low-latency storage, with latencies of around 1 ms for frequently accessed data, while heavier sequential access continues to be served directly from S3 to maximize throughput.
AI and data-focused
The positioning of S3 Files is initially aimed at interactive workloads: AI agents collaborating on file-based tools, preparing datasets for training, or existing applications that rely on libraries and scripts manipulating files. AWS highlights the absence of upfront migration and the ability to access the same data both through the file interface and via the S3 APIs.
In the current context, this launch fits into AWS’s strategy around AI and the streamlining of data pipelines. The cloud provider stresses that teams would no longer need to duplicate datasets between object and file storage, reducing synchronization complexity and storage silos.
A service ready at scale
AWS states that S3 Files is now generally available in 34 regions. The service supports simultaneous access by thousands of compute resources, with consistency mechanisms tailored to shared usage, and integrates with IAM controls, TLS 1.3 encryption, and AWS’s server-side encryption options.
Technically, AWS relies on Amazon EFS under the hood to provide file access, while preserving S3 as the single source of truth. This approach enables AWS to present S3 Files as a hybrid building block, aimed at enterprises that want to keep their data in S3 without giving up the file-hierarchy tools they rely on.
S3 Files does not appear as a standalone paid service, but rather as a capability that makes S3 available in filesystem mode, with pricing following the standard Amazon S3 pricing model.
