Published on March 22, 2024, 6:25 am

AWS and NVIDIA have come together to offer a game-changing solution in the form of the NVIDIA Grace Blackwell GPU-based Amazon EC2 instances to turbocharge performance when constructing and executing inferences on multi-trillion parameter LLMs. This groundbreaking collaboration will see AWS introducing the cutting-edge NVIDIA GB200 Grace Blackwell Superchip and B100 Tensor Core GPUs, expanding their well-established partnership to deliver top-notch infrastructure, software, and services tailored to unlock new generative artificial intelligence (AI) capabilities.

The alliance between NVIDIA and AWS showcases a fusion of their top technologies, incorporating NVIDIA’s latest multi-node systems featuring the state-of-the-art NVIDIA Blackwell platform and AI software alongside AWS’s Nitro System, AWS Key Management Service (AWS KMS) advanced security features, Elastic Fabric Adapter (EFA) for petabit scale networking, and Amazon EC2 UltraCluster hyper-scale clustering. These combined offerings aim to empower customers with the tools required to efficiently build and operate real-time inference on multi-trillion parameter large language models (LLMs), all at a faster pace, expansive scale, and reduced cost compared to previous-generation NVIDIA GPUs on Amazon EC2.

Through this synergistic partnership that dates back more than 13 years with the inception of the world’s first GPU cloud instance on AWS, both companies are continuously innovating. Adam Selipsky, CEO at AWS expressed his enthusiasm regarding the potential that NVIDIA’s next-generation Grace Blackwell processor brings to generative AI and GPU computing. Pairing this advancement with AWS’s robust Elastic Fabric Adapter Networking, Amazon EC2 UltraClusters’ hyper-scale clustering capability, and the unique Nitro system’s virtualization features forms an unparalleled offering enabling customers to propel multi-trillion parameter large language model projects faster, at massive scale, more securely than ever before.

Jensen Huang, founder and CEO of NVIDIA pointed out that AI is propelling breakthroughs across various industries at an unprecedented rate. The collaboration between AWS and NVIDIA is geared towards accelerating advanced generative AI capabilities by providing customers with unmatched computing power to explore new frontiers.

Furthermore, initiatives such as Project Ceiba underscore the commitment between AWS and NVIDIA in driving future generative AI innovation by leveraging cutting-edge technologies like the Blackwell-powered DGX Cloud instances offered on AWS. These instances will play a pivotal role in expediting the development of next-gen generative AI models enabling breakthroughs that transcend previous limitations.

The strategic alignment between AWS and NVIDIA paves the way for advancements in diverse sectors including healthcare and life sciences through offerings like Amazon SageMaker integration with NIM inference microservices powered by NVIDIA AI Enterprise. This collaboration aims to revolutionize computer-aided drug discovery with innovative solutions like BioNeMo FMs for generative chemistry on platforms like HealthOmics integrated within AWS ecosystem.

In conclusion, this partnership marks a significant milestone in harnessing leading-edge technology from both entities to drive innovation across industries while providing customers with unmatched capabilities in Generative Artificial Intelligence applications.


Comments are closed.