AWS Deploys P6-B300 EC2 Instances in Jakarta, Enhancing AI Training for Large Foundation Models

The new P6-B300 EC2 instances with NVIDIA Blackwell Ultra GPUs deliver increased memory and networking for AI model training in Jakarta and other regions.

By Oliver Grant · Editorial pseudonym · 2026-09-03 · 1 min read

AI-ASSISTED REPORTING · RETROSPECTIVE ARCHIVE

Retrospective edition: 2026-09-03. Source published: 2026-09-03. Added to this archive: 2026-11-03. This briefing was prepared later from the linked announcement; it was not published on this website on the historical edition date.

AWS expanded its P6 series with P6-B300 instances available in Asia Pacific (Jakarta) and additional regions. These instances feature eight NVIDIA Blackwell Ultra GPUs, 2.1 TB of GPU memory, and 6.4 Tbps Elastic Fabric Adapter networking, improving speed and scale for training trillion-parameter foundation models.

Compared to P6-B200, P6-B300 offers 2x networking bandwidth and 1.5x GPU memory capacity and TFLOPS (FP4), enhancing throughput for large language models and AI workflows. This positions P6-B300 as optimized hardware for sophisticated AI training and deployment at scale.

Source & further reading

Amazon EC2 P6-B300 instances are now available in the AWS Asia Pacific (Jakarta) Region ↗

Source published 2026-09-03. 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.