G7e instances offer up to 2.3 times inference performance over previous generation, with up to eight Blackwell GPUs, 96 GB GPU memory each, and 5th Gen Intel Xeon processors. They support up to 192 vCPUs and high network bandwidth, optimized for large language models and multimodal AI applications.
Instances include NVIDIA GPUDirect features for improved multi-GPU and multi-node communication. Available now in Asia Pacific (Seoul) and Europe (Spain), these instances support on-demand, spot, and savings plans. Performance claims are vendor reported; independent benchmarks would provide comprehensive validation.
Source & further reading
Amazon EC2 G7e instances now available in Asia Pacific (Seoul) and Europe (Spain) regions ↗Source published 2026-03-09. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.