AWS AI Updates: April 28, 2026

1. SageMaker HyperPod Adds G7e (RTX PRO 6000 Blackwell) and r5d.16xlarge Instances

AWS. Amazon SageMaker HyperPod now supports G7e instances powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs and r5d.16xlarge for orchestration and preprocessing. AWS claims G7e delivers up to 2.3x better inference performance than G6e and exposes up to 768 GB of total GPU memory, making it a fit for LLM serving, agentic AI and multimodal generative AI workloads. Initial regions are US East (N. Virginia, Ohio), US West (Oregon) and Asia Pacific (Tokyo). Source

2. Amazon Redshift Serverless Makes AI-Driven Scaling the Default

AWS. Redshift Serverless now turns on AI-driven scaling by default for new workgroups, using machine learning to forecast compute demand and adjust resources rather than relying on static auto-scaling rules. AWS also widened the supported Base RPU range to 8–512, lowering the entry price for smaller analytics and AI/ML feature-engineering workloads on Redshift. Source