Your company account is blocked and you cannot place orders. If you have questions, please contact your company administrator.

NVIDIA’s Powerful AI Supercomputers

Right at Your Fingertips

NVIDIA’s DGX Systems are purpose-built for the complex challenges of AI. As the world’s most powerful GPU computing platforms, they set a new bar for enterprise AI infrastructure and next-gen data centers. NVIDIA’s DGX Systems are designed to give data scientists the most powerful tools for AI exploration, delivering groundbreaking performance and results. Get faster training, larger models, and more accurate results.

Test Before You Invest

Trace3 is a select solutions provider for DGX loaner access, which includes the DGX-1 and DGX A100. Benefits of our exclusive program include:

  • Remote POC sessions available with VPN access.
  • Competitive rates.
  • Up to 8 GPUs per user (GPUs can also be divided among users).
  • Access to multiple deep learning platform containers.
  • Test before you invest—get the results you expect before spending the money.
  • GPU-based hardware/software POC in our lab.


From data analytics to inference to training, the world’s most advanced system for AI workloads has arrived. The unprecedented compute power of the DGX A100 delivers the fastest time-to-solution, allowing users to tackle challenges that were previously not possible. Highlights of the DGX A100 include:

  • NVIDIA’s Ampere architecture optimized for sparse AI Tensor Ops, offering 2X faster execution overall versus the previous generation Volta GPU architecture for machine learning applications.
  • Five petaFLOPS of AI performance into a 6U form factor, replacing legacy compute infrastructure with a single, unified system.
  • A100 GPUs bring a new precision TF32 providing 20X higher FLOPS for AI vs. the previous generation.
  • Eight NVIDIA A100 Tensor Core GPUs for unmatched acceleration.
  • A100 GPU with a class-leading 1.6 terabytes per second (TB/s) of memory bandwidth, more than a 70% increase over the last generation.
  • 40MB Level 2 cache on-chip memory 7x larger than the previous generation.
  • The world’s most robust security posture for AI enterprise.
  • Unmatched data center scalability.
  • Fastest I/O architecture of any DGX system.

Learn More


Gain insights in hours instead of weeks or months. DGX-1 is the integrated software and hardware system that supports AI research with an optimized combination of computing power, software, and deep learning performance. Features/highlights of the DGX-1 include:

  • Most common deep learning frameworks specifically optimized for NVIDIA DGX-1 are available as containers.
  • Simplifies portability with NVIDIA Docker containers.
  • The power to run multiple frameworks at once.
  • DGX-1 with 8 x V100 NVLink delivers 96X faster training.
  • NVIDIA NVLink high-speed interconnect delivers five times more GPU-to-GPU bandwidth.
  • Cuts infrastructure footprint by 250X and reduces cost by 20X.
  • Reduces deep learning training time.
  • Vast amounts of data can be analyzed and visualized.
  • Accelerates deep learning frameworks.
  • Designs more sophisticated neural networks.


AI in Dairy Operations and Animal Health - Read Article >>

Contact Us for More Information on NVIDIA DGX