AWS PCS Supports NVIDIA Blackwell UltraServers for Large-Scale GPU Workloads

AWS Parallel Computing Service now integrates with Amazon EC2 P6e-GB200 and P6e-GB300 UltraServer instances, enabling scalable HPC GPU clusters managed by Slurm.

By Noah Ellis · Editorial pseudonym · 2026-06-18 · 1 min read

AI-ASSISTED REPORTING · RETROSPECTIVE ARCHIVE

Retrospective edition: 2026-06-18. Source published: 2026-06-18. 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.

P6e UltraServers feature up to 72 NVIDIA Blackwell GPUs within an NVLink domain, delivering up to 360 petaflops FP8 compute and 13.4 TB high bandwidth memory. The P6e-GB300 variant offers increased GPU memory and compute over P6e-GB200.

Through PCS, customers can utilize reserved UltraServers on EC2 Capacity Blocks and deploy them into Slurm-managed node groups with automatic topology configuration. This combination supports large-scale AI and scientific workloads across multiple AWS regions, integrating compute, storage, and networking seamlessly.

Source & further reading

AWS Parallel Computing Service supports P6e-GB200 and P6e-GB300 UltraServers ↗

Source published 2026-06-18. 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.