Fall GTC 2020 announcements

At NVIDIA GPU Technology Conference 2020 NVIDIA announced several new Enterprise solutions. Below is a summary of the most interesting announcements.


NVIDIA RTX A6000 is the latest GPU built on the NVIDIA Ampere architecture to power the most graphics and compute intensive workflows from the desktop. The RTX A6000 delivers everything professionals need to meet today’s design, creative, and scientific challenges.

  • The RTX A6000 features the latest generation RT Cores Tensor Cores, and CUDA Cores for unprecedented rendering, AI, graphics, and compute performance.
  • The RTX A6000 includes 48 GB of graphics memory, scalable to 96 GB with NVIDIA NVLink, so professionals can tackle the most memory intensive workloads from their desktop.
  • The RTX A6000 also supports PCIe Gen 4 for double the bandwidth of PCIe Gen 3, which improves data-transfer speeds from CPU memory for data-intensive tasks like AI and data science.
Nvidia RTX A6000

nvidia A40


NVIDIA A40 is the world’s most powerful data center GPU for visual computing, delivering unparalleled performance and capabilities to professionals anywhere for graphics-based workloads such as ray traced rendering, high-performance virtual workstations, simulation, 3D design, VR, and virtual production.

  • Built on NVIDIA’s Ampere architecture, the NVIDIA A40 includes the latest generation NVIDIA RT Cores, Tensor Cores, and CUDA cores for state of the art graphics, compute, and AI acceleration.
  • 48GB of GDDR6 memory for massive models and datasets, up to 96GB with two-way NVIDIA NVLink.
  • NVIDIA A40 is configured for virtualization by default, but also features 3 display ports and support for Quadro Sync for customers requiring display capabilities such as  powering large-scale display wall experiences or virtual production setups.
  • The NVIDIA A40 is at the heart of powering visualization workloads in the EGX Server for Professional Visualization, formerly branded as RTX Server.

NVIDIA DGX SuperPOD Solution for Enterprise

For enterprises that need the fastest path to AI innovation at scale, NVIDIA DGX SuperPOD Solution for Enterprise is a turnkey hardware, software, and services offering that removes the guesswork from building and deploying AI infrastructure.

  • DGX SuperPOD Solution for Enterprise is built on the DGX SuperPOD reference architecture—the world’s fastest, commercially available AI infrastructure, as proven in the MLPerf benchmark suite. It enables businesses to acquire world-class AI infrastructure, hand-guided by NVIDIA and deployed in a fraction of the time so data science teams can be up and running on day one.
  • Data science teams need the right tools, platform, and infrastructure to streamline AI workflows and speed time to insights. And IT teams need the right partner to integrate AI infrastructure within the existing environment. With DGX SuperPOD Solution for Enterprise, our professional services team and NVIDIA partners will help customers adapt our proven infrastructure solution to their environment, including flexible deployment options tailored to their business.
  • More than a reference architecture, teams need a faster path to making AI infrastructure operationally useful to the business. With DGX SuperPOD Solution for Enterprise, customers will benefit from full-service data center planning and infrastructure delivery expertise that speeds deployment, from sizing, to installation, to training, to ongoing optimization and beyond, all backed by NVIDIA.

nvidia ngc support services

NVIDIA-Certified Systems

The current NGC-Ready program is evolving and transitioning into the NVIDIA-Certified Systems program. This program is introduced with the NVIDIA A100 GPU, the latest Mellanox network adapters (ConnectX-6, ConnectX-6 Dx , BlueField-2 DPU), the Quantum family of InfiniBand switches and Spectrum family of Ethernet switches and cables.

This allows enterprises to confidently choose GPU optimized systems that have been tested for single node and multi-node GPU performance on a wide range of popular AI workloads.  These secure systems are available in smaller configurations and as scalable building blocks backed by enterprise-grade support.

  • A Unified AI Platform: Simplifies the development experience, reduces time to software deployment and ensures GPU-optimized software from the NGC catalog works out of the box.
  • Performance Optimized: Tested for high performance on important AI and data analytics workloads including, deep learning, inference, machine learning, and industry-specific AI workloads.
  • Scalable: Enables faster time to solution by allowing you to scale up and scale out, using multi node cluster configurations, with NVIDIA Ampere architecture GPUs and NVIDIA Mellanox SmartNICs and switches.
  • Built-in Security: NVIDIA Mellanox SmartNICs offer the ability to offload and accelerate multiple layers of network security with Transport Layer Security (TLS) and Internet Protocol Security (IPsec) in-line cryptography that help secure AI workflows.
  • Enterprise-Grade Support: Minimizes system downtime during infrastructure deployment and maximize user productivity during development

Jetson Nano 2GB Developer Kit

  • The developer kit Is ideal for teaching, learning and developing AI and Robotics.  It benefits from an active developer community with ready-to-build open-source projects.
  • For educators, students and enthusiasts new to AI and robotics, it delivers the ability to create hands-on projects with the same NVIDIA JetPack used for breakthrough AI-based products across all industries.
  • Jetson Nano 2GB Developer Kit delivers incredible AI performance at a low price.

JetPack 4.4.1

  • Jetson Linux Driver Package (L4T) with familiar desktop Linux environment
  • CUDA-X accelerated libraries and APIs for Deep Learning, Computer Vision, Accelerated Computing and Multimedia
  • Includes samples, documentation, and developer tools