AWS Launches Nvidia Blackwell-Based Servers For Next-Gen AI Workloads

Deepak

To accelerate innovation across emerging generative AI developments such as reasoning models and agentic AI systems, Amazon Web Services (AWS) has announced general availability of P6e-GB200 UltraServers, accelerated by NVIDIA Grace Blackwell Superchips. P6e-GB200 UltraServers are designed for training and deploying the largest, most sophisticated AI models. Earlier this year, AWS launched P6-B200 instances, accelerated by NVIDIA Blackwell GPUs, for diverse AI and high-performance computing workloads.

The P6e-GB200 UltraServers are AWS’s most powerful GPU offering to date, featuring up to 72 interconnected NVIDIA Blackwell Superchips. These UltraServers deliver massive compute power and memory capacity, ideal for training and deploying the largest, most complex AI models at the trillion-parameter scale, according to AWS VP David Brown.

Additionally, the P6-B200 instances have been launched as a versatile option for a wide range of AI and high-performance computing workloads. These instances offer 8 NVIDIA Blackwell GPUs with significantly improved processing power and memory compared to previous generations.

Both solutions build on AWS’s years of experience delivering secure and reliable GPU infrastructure at massive scale, supported by the AWS Nitro System for maximum security and stability. Customers can leverage these new capabilities through Amazon SageMaker HyperPod, Amazon Elastic Kubernetes Service (EKS), and NVIDIA DGX Cloud on AWS.

Related Items

Salesforce And AWS Expand Partnership

Salesforce And AWS Expand Partnership

Cloudera Collaborates With AWS

Cloudera Collaborates With AWS

Lost Password