Amazon EC2 G7 instances launched with NVIDIA RTX PRO 4500 Blackwell GPUs

Today, Amazon Web Services announced the general availability of Amazon EC2 G7 instances, delivering high performance GPU acceleration for AI inference, graphics, and data analytics workloads. AWS is the first major cloud provider to support NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs. G7 instances are accelerated by these GPUs with custom sixth-generation Intel Xeon Scalable processors, delivering up to 4.6x AI inference performance and up to 2.1x graphics performance compared to G6 instances. They also provide faster performance for GPU-accelerated analytics on Amazon EMR and Amazon Elastic Kubernetes Service (Amazon EKS). G7 instances are suited for a broad range of GPU-enabled workloads including AI inference, graphics rendering, video transcoding, spatial computing, virtual desktop infrastructure (VDI), and data analytics. Improvements over previous generation include faster GPU memory with 1.33 times the capacity and 2.45 times the bandwidth, 32 GB GPU memory per GPU, 5th Gen Tensor Cores, and 4th Gen RT Cores. Networking throughput is 700 Gbps with EFA-enabled networking, 7x higher than G6, supporting low-latency, high-bandwidth connectivity. Storage supports up to 7.6 TB local NVMe SSD. Advanced video encoding and decoding engines (9th-gen NVENC and 6th-gen NVDEC) support 4:2:2 encoding and decoding for high-resolution video workflows, delivering 1.5x concurrent video streams compared to G6. EC2 G7 instance specifications include up to 8 NVIDIA RTX PRO 4500 Blackwell GPUs with up to 256 GB total GPU memory, custom Intel Xeon Scalable processors, 7 sizes, up to 192 vCPUs, 700 Gbps network bandwidth, 768 GiB system memory, and 7.6 TB local NVMe SSD storage. The instance sizes range from g7.2xlarge with 1 GPU and 8 vCPUs to g7.48xlarge and g7.metal with 8 GPUs and 192 vCPUs. G7 instances support NVIDIA GPUDirect P2P for multi-GPU sizes and GPUDirect RDMA with EFA for Amazon FSx for Lustre, enabling low-latency GPU-to-GPU communication for multi-GPU and multi-node workloads. Users can get started with AWS Deep Learning AMIs or NVIDIA Workstation AMIs with prepackaged GPU drivers. For Amazon EKS, EKS AMIs with NVIDIA driver version R595 are supported. Operating systems supported include Amazon Linux, Ubuntu, RHEL, and Windows Server with NVIDIA driver integration supporting DirectX, Vulkan, and OpenGL. G7 instances are available today in US East (Ohio) and US West (Oregon) regions, with plans for expansion. Purchasing options include On-Demand, Savings Plans, Spot Instances, and Dedicated Instances for 12xlarge, 24xlarge, and 48xlarge sizes. More details and launch options are available on the Amazon EC2 console and EC2 G7 instances page. Feedback can be shared on AWS re:Post or through AWS Support contacts. – Daniel Abib