Amazon Web Services plans to launch the Project Ceiba supercomputer in partnership with GPU-maker Nvidia.
The system will feature 16,384 GH200 NVL32 Grace Hopper Superchips and is expected to be the world’s largest cloud AI supercomputer. Update: The supercomputer has been upgraded with Nvidia's latest chips. Find out more here.
The system will also feature the AWS Elastic Fabric Adapter (EFA), the AWS Nitro system, and Amazon EC2 UltraCluster scalability technologies.
Amazon said that the supercomputer is expected to provide 65 exaflops of AI performance and have up to 9.5 petabytes of total memory.
“This new supercomputer will be set up inside of AWS infrastructure hosted by AWS and used by Nvidia’s own research and engineering teams," Ian Buck, Nvidia's VP of hyperscale and HPC, said.
It will be used "to develop new AI for graphics, large language model research, image, video, 3D, generative AI, digital biology, robotics research, self-driving cars, and more.”
The most powerful cloud supercomputer that has been publicly disclosed so far is Microsoft's Eagle system, ranked as the third most powerful supercomputer in the world. It features Intel Xeon Platinum 8480C 48C CPUs and 14,400 Nvidia H100 GPUs.