AWS Snowball Edge ML Support enables ML inference on edge devices by running pre-trained models locally, ideal for disconnected environments or latency-sensitive applications.

https://aws.amazon.com/snowball-edge/