NVIDIA OVX Systems

NVIDIA RTX™-Accelerated AI and Graphics Performance From [digital twins](https://www.nvidia.com/en-in/omniverse/solutions/digital-twins/) and [physical AI](https://www.nvidia.com/en-in/glossary/generative-physical-ai/) to [generative AI](https://www.nvidia.com/en-in/ai-data-science/generative-ai/) and LLM [inference](https://www.nvidia.com/en-in/deep-learning-ai/solutions/inference-platform/), NVIDIA OVXTM systems deliver industry-leading graphics and compute performance to accelerate 3D, industrial AI, and simulation workloads in the data center scale.

Unlock the Power of NVIDIA RTX for Your Workloads Generative AI Develop new services, insights, and original content. With next-generation AI, graphics, and media acceleration capabilities, OVX with L40S delivers higher inference performance than the previous generation. LLM Training and Inference Accelerate AI training and inference workloads. Get exceptional AI computing performance to accelerate training and inference of state-of-the-art LLM and generative AI models. Industrial Digitalization Create and operate physical AI applications. NVIDIA Omniverse™ makes it possible to connect, develop, and operate the next wave of industrial digitalization applications. With powerful RTX graphics and AI capabilities, OVX delivers exceptional performance for OpenUSD-based industrial digital twins and robotics simulation applications built with Omniverse libraries. Rendering and 3D Graphics Power high-fidelity creative workflows with NVIDIA RTX graphics. Power the creation of stunning visual content and high-fidelity creative workflows, from interactive rendering to real-time virtual production.

Pixar, Adobe, Apple, Autodesk, and NVIDIA Form Alliance for OpenUSD The Alliance for OpenUSD (AOUSD) aims to foster the standardization, development, evolution, and growth of Pixar’s Universal Scene Description (OpenUSD) technology. NVIDIA Releases Major Omniverse Upgrade With Generative AI and OpenUSD New Omniverse platform updates, connections to Adobe Firefly, and Ada Generation systems to accelerate interoperable 3D workflows and industrial digitalization. NVIDIA Omniverse Opens Portals to Vast Worlds of OpenUSD New Omniverse Cloud APIs—ChatUSD LLM for conversing in USD, RunUSD for translating USD to images, and DeepSearch LLM, enabling semantic 3D search—to help developers adopt Universal Scene Description (OpenUSD).

Unified Architecture for Accelerating Generative AI and Industrial Digitalization Universal Acceleration Accelerate time to market and improve TCO with high-performance infrastructure powered by L40S GPUs,​ delivering breakthrough performance for complex AI and graphics-intensive workloads. Validated Architecture Optimized to accelerate the full stack of NVIDIA software, from training, fine-tuning, and deploying NVIDIA NIM™ microservices with NVIDIA AI Enterprise to powering industrial digitalization with NVIDIA Omniverse libraries. Performance at Scale Ultra-fast, high-bandwidth communication for distributed workloads with ConnectX-7 network adapters and Bluefield-3 DPUs. Efficiently scale across 100s of OVX servers with low-latency Ethernet or Infiniband platforms.

Start Accelerating Your AI and Graphics Workloads NVIDIA AI Enterprise Supercharge production AI, improve productivity, and achieve insights faster with NVIDIA AI Enterprise. With over 100 frameworks, pretrained models, and development tools optimized for building and running AI, NVIDIA AI Enterprise accelerates every stage of the AI journey, from data prep and model training through inference and deployment at scale. NVIDIA Omniverse NVIDIA Omniverse is a collection of libraries and microservices for developing industrial digital twins and physical AI simulation applications. Software makers can integrate pre-built functionality for open data interoperability (OpenUSD), real-time RTX rendering, and advanced physics simulation to safely validate and test AI systems before real-world deployment.

Take a Deeper Dive Into NVIDIA OVX Featured Exploring Digital Twins: Key Insights From Five Industry Leaders Explore five real-world use cases and get insights and best practices for developing digital twins. News New Ethernet Networking Platform for AI Available Soon Dell Technologies, Hewlett Packard Enterprise, and Lenovo will be the first to offer systems featuring NVIDIA’s full stack for generative AI, which includes the latest GPUs, AI software, and, now, the NVIDIA Spectrum-X™ Ethernet networking platform for AI. NVIDIA BlueField-3 SuperNICs Accelerate OVX Networking for AI Peak AI performance demands optimized networking. The NVIDIA BlueField-3 SuperNIC is a new class of network accelerators tailored for AI. As part of the Spectrum-X networking platform, BlueField-3 SuperNICs deliver peak AI workload efficiency and supercharge hyperscale generative AI workloads running on OVX systems. Global Data Center System Manufacturers to Supercharge Generative AI and Industrial Digitalization New OVX servers with NVIDIA L40S GPUs deliver breakthrough multi-workload performance, combining powerful AI compute with best-in-class graphics and media acceleration. Training Robots and Streamlining Assembly These technologies streamline communication, optimize layouts, and accelerate robot training within digital twins of their factories. They also enhance operational efficiency, process optimization, and production capacity in factories from design to deployment.​ Real-Time Computer-Aided Engineering Digital Twins Build interactive workflows for manufacturing design and simulation with physics-ML, accelerated solvers, and real-time rendering. Building Autonomous Railway Networks To achieve climate neutrality by 2050, German national railway operator, Deutsche Bahn, is building autonomous railway networks and trains to maximize efficiency of existing railroad capacity for the transport of goods and passengers.

OVX L40S Server with four GPUs CPUs: 2x 32c Intel Xeon Gold 6448Y<br>2x 32c AMD EPYC 9354 GPUs: 4x NVIDIA L40S Networking — E/W: 2x NVIDIA ConnectX®-7 (2x200Gb) or 2x NVIDIA BlueField™-3, B3140H (1x400Gb) Networking — N/S: 1x BlueField-3, B3220 (2x200Gb) Host Memory: Min 384 GB DDR5 ECC (1 DIMM per slot) Host Boot Drive: 1x 1 TB NVMe Host Storage: 2x 4 TB NVMe OVX L40S Server with eight GPUs CPUs: 2x 56c Intel Xeon Gold 8480+<br>2x 64c AMD EPYC 9554 GPUs: 8x NVIDIA L40S Networking — E/W: 4x ConnectX-7 (2x200Gb) or 4x BlueField-3, B3140H (1x400Gb) Networking — N/S: 1x BlueField-3, B3220 (2x200Gb) Host Memory: Min 768 GB DDR5 ECC (1 DIMM per slot) Host Boot Drive: 1x 1 TB NVMe Host Storage: 2x 4 TB NVMe