The universal system for all your AI workloads and infrastructure is here. The NVIDIA DGX™ A100 sets a new standard for compute density, performance and flexibility. The single platform brings together training, inference, and analytics into a consolidated and easy-to-deploy system with optimized software that is the foundational building block for AI infrastructure.
The DGX A100 packs five petaFLOPS of AI performance into a 6U form factor, replacing legacy infrastructure with a single, unified system. Offering the fastest time to solution, the DGX A100 is the world’s first AI system built on the A100 Tensor Core GPU, the world’s most advanced accelerator. With the unprecedented acceleration of computing power, the DGX A100 enables IT administrators to assign resources that are right-sized for specific workloads, ensuring that the largest and most complex jobs are supported, along with the simplest and smallest.
Setting a new bar for data center scalability, the DGX A100 also features Mellanox ConnectX-6 VPI HDR InfiniBand/Ethernet network adapters with 450 gigabytes per second of peak bi-directional bandwidth.
Backed by integrated access to a global team of over 14,000 DGX experts that come with every DGX system, the DGX A100 accelerates the end-to-end lifecyle of enterprise AI across planning, deployment and on-going optimization.