AWS is set to offer NVIDIA Grace Blackwell GPU-based Amazon EC2 instances and NVIDIA DGX Cloud to enhance the performance of building and running inference on multi-trillion parameter LLMs. This collaboration between AWS and NVIDIA was announced at GTC, showcasing the new NVIDIA Blackwell GPU platform coming to AWS.
The partnership between AWS and NVIDIA aims to provide customers with secure and advanced infrastructure, software, and services for unlocking new generative AI capabilities. With NVIDIA’s Grace Blackwell Superchip and B100 Tensor Core GPUs, AWS will extend its offerings to include the latest NVIDIA technologies to accelerate AI innovation.
By leveraging NVIDIA’s Blackwell platform and AI software along with AWS’s Nitro System, AWS Key Management Service (AWS KMS), Elastic Fabric Adapter (EFA), and Amazon EC2 UltraCluster, customers can build and run real-time inference on multi-trillion parameter large language models faster and at a lower cost than before.
AWS CEO, Adam Selipsky, emphasized the long-standing collaboration between AWS and NVIDIA, highlighting the continuous innovation in providing customers with cutting-edge GPU solutions in the cloud. NVIDIA’s CEO, Jensen Huang, echoed the sentiment, recognizing the impact of AI in driving breakthroughs across industries.
The latest innovations from AWS and NVIDIA focus on accelerating the training of cutting-edge LLMs with over 1 trillion parameters. AWS will offer the NVIDIA Blackwell platform featuring GB200 NVL72, allowing customers to scale to thousands of GB200 Superchips for faster inference workloads.
To enhance AI security, AWS Nitro System, AWS KMS, encrypted EFA, and Blackwell encryption are combined to protect customer data and model weights. The collaboration also includes Project Ceiba, a supercomputer built exclusively on AWS for NVIDIA’s research and development, to advance generative AI innovation.
AWS and NVIDIA’s collaboration extends to healthcare and life sciences, offering high-performance inference for generative AI applications. By integrating Amazon SageMaker with NVIDIA NIM inference microservices and BioNeMo FMs for generative chemistry, the partnership aims to accelerate drug discovery and advance healthcare use cases.
Overall, the collaboration between AWS and NVIDIA is driving innovation in AI, providing customers with the tools and infrastructure needed to unlock new possibilities in generative AI across various industries.
Source link
