AWS launches G7e instances with NVIDIA Blackwell GPUs on SageMaker AI, enabling cost-effective deployment of 120B+ parameter foundation models

G7e instances feature NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs with 96 GB GDDR7 memory per GPU

Available configurations include 1, 2, 4, and 8 GPU node options to accommodate different workload requirements

Single G7e.2xlarge instances can now host large open-source models including GPT-OSS-120B, Nemotron-3-Super-120B-A12B, and Qwen3.5-35B-A3B

New capability provides organizations with high-performance generative AI inference at reduced costs compared to multi-node deployments