AWS published a technical guide for deploying Qwen3.8-2.4T-A95B, a 2.4-trillion-parameter open-weight model, on SageMaker HyperPod using vLLM, featuring quantization and reasoning capabilities.

AWS released Ray Serve Deep Learning Containers to replace TorchServe, providing a supported, pre-assembled container for GPU inference workloads on Kubernetes clusters.

AWS announced August 2026 product updates including million-token context windows, multi-region inference, extended agent runtimes, and expanded GovCloud availability across Bedrock and related services.

Heurist Finance deployed a conversational AI investment workbench on Amazon Bedrock AgentCore, leveraging agentic AI for market data analysis with sandboxed code execution and full auditability.