Amazon EC2 P6-B300 instances are now available in the US East (N. Virginia) Region

Amazon Web Services has expanded availability of its EC2 P6-B300 instances to the US East (N. Virginia) Region, bringing high-performance AI computing infrastructure closer to East Coast enterprises and government agencies. The P6-B300 instances feature eight NVIDIA Blackwell Ultra GPUs with 2.1 TB of high bandwidth GPU memory, 6.4 Tbps Elastic Fabric Adapter (EFA) networking, 300 Gbps dedicated Elastic Network Adapter (ENA) throughput, and 4 TB of system memory. Compared to the previous P6-B200 instances, the P6-B300 delivers significant performance improvements including double the networking bandwidth, 1.5x larger GPU memory capacity, and 1.5x more GPU TFLOPS at FP4 precision. These enhancements are specifically designed to support training and deployment of trillion-parameter foundation models and large language models, with the increased networking capacity and memory enabling faster training times and higher token throughput for demanding AI workloads. The P6-B300 instances are available in p6-b300.48xlarge configuration across three AWS regions: US West (Oregon), AWS GovCloud (US-East), and the newly added US East (N. Virginia). This expanded regional availability provides organizations with more geographic options for deploying large-scale AI training and inference workloads while maintaining low-latency access to AWS services.