Nvidia and Amazon Web Services (AWS) are announcing a new phase in their strategic partnership with several significant developments at the AWS re:Invent conference.
First, Nvidia is introducing DGX Cloud, bringing the Grace Hopper GH200 superchip to AWS for the first time. This project, known as Ceiba, aims to establish the world’s largest public cloud supercomputing platform on AWS, delivering 64 exaflops of AI power. Additionally, AWS will be adding four new kinds of GPU-powered cloud instances to its EC2 service.
Nvidia is also launching NeMo Retriever to help organizations construct superior large language models. The goal is to connect enterprise data to generative AI more effectively. Ian Buck, VP of Hyperscale and HPC at Nvidia, emphasizes that the two companies have been collaborating for over 13 years, primarily focusing on improving innovation and operation at AWS and for mutual customers.
The DGX Cloud is an optimized deployment of Nvidia hardware and software designed to provide supercomputing capabilities for AI. The DGX Cloud will functionally enable supercomputing type capabilities for AI, a first for AWS.
Additionally, Nvidia and AWS will work together to build Project Ceiba, aiming to construct the world’s largest cloud AI supercomputer. This supercomputer will be equipment with 16,000 Grace Hopper superchips and have up to 9.5 Petabytes of total memory.
Moreover, Nvidia is launching NeMo Retriever, an offering that aims to help construct enterprise grade chatbots. The technology allows businesses to connect large language models with enterprise data, making the chatbots more productive and useful. This technology also includes accelerated vector search for optimizing the performance of vector databases.
These announcements represent a significant advancement in the partnership between Nvidia and AWS, showcasing the potential for improved future collaboration in fields such as AI and cloud computing. The continued collaboration between these two industry leaders suggests that more innovative solutions can be expected in the future.
