Announcing Amazon EC2 G7e instances accelerated by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs

This article announces the general availability of Amazon EC2 G7e instances powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs for AI inference and graphics workloads.

• Up to 2.3x inference performance improvement over G6e instances
• NVIDIA RTX PRO 6000 Blackwell GPUs with 2x GPU memory and 1.85x bandwidth
• Support for 70B parameter models with FP8 precision on single GPU
• NVIDIA GPUDirect P2P enables direct GPU-to-GPU communication over PCIe
• Up to 4x inter-GPU bandwidth compared to L40s GPUs in G6e instances
• 4x networking bandwidth improvement for multi-node workloads
• Up to 8 GPUs per instance with 768GB total GPU memory
• Available in US East (N. Virginia) and US East (Ohio) regions
• Purchasable as On-Demand, Savings Plan, Spot, Dedicated Instances, or Dedicated Hosts

G7e instances provide significant performance improvements for large-scale AI inference and graphics workloads with enhanced multi-GPU capabilities and networking.