AWS First Steps
Region expansion

Amazon EC2 P6-B300 instances now available in additional AWS Regions

Amazon EC2 introduces P6-B300 instances in new regions, offering high-performance computing ideal for training and deploying large AI models with enhanced networking and memory

Amazon EC2 P6-B300 instances are now available in the Asia Pacific (Hyderabad) and South America (Sao Paulo) Regions. These instances feature 8xNVIDIA Blackwell Ultra GPUs, 2.1 TB of high bandwidth GPU memory, 6.4 Tbps EFA networking, 300 Gbps dedicated ENA throughput, and 4 TB of system memory. Compared to P6-B200 instances, P6-B300 offers twice the networking bandwidth, 1.5 times the GPU memory size, and 1.5 times the GPU TFLOPS (at FP4, without sparsity), making them well-suited for training and deploying large trillion-parameter foundation models and large language models efficiently.

Why it matters

This update impacts users who require AI model training and deployment, offering high-performance computing suitable for applications needing large computational resources and high-bandwidth networking.

Read the original AWS announcement

Related articles and terms

GPU-Based Machine Learning Training with AWS Batch - Cost-Efficient Large-Scale TrainingRun GPU training with your existing Docker containers, and cut costs by up to 90% using Spot Instances and checkpointing. Includes guidance on when to choose Batch over SageMaker.Amazon EC2A cloud-based virtual server service from AWS that lets you choose from hundreds of instance types and launch servers in minutesAmazon Elastic Inference[Discontinued] Service that attached fractional GPU accelerators to EC2 and SageMaker for inference. Inferentia (Inf1 / Inf2), GPU instances, and SageMaker AI inference are the replacementsAmazon SageMakerA fully managed platform for building, training, and deploying machine learning models end-to-end, significantly boosting productivity for data scientists and ML engineersThe Layered Architecture of AWS AI/ML Services - Flexibility Through the Three Tiers of SageMaker, Bedrock, and API ServicesThis article organizes AWS AI/ML services into three layers - SageMaker (full control), Bedrock (managed generative AI), and Rekognition/Comprehend/etc. (API-based) - and explains AWS's flexibility through comparisons with GCP Vertex AI and Azure OpenAI Service, including custom silicon integration.