Nvidia’s latest RTX Pro 6000 Blackwell Server Edition GPU will soon be available on enterprise servers.
Cisco, Dell Technologies, HPE, Lenovo, and Supermicro systems ship a variety of configurations to 2U servers. Nvidia says it will increase performance and efficiency for AI, graphics, simulation, analysis, industrial applications, and support tasks such as AI model training, content creation and scientific research.
“AI is reinventing computing for the first time in 60 years. What started in the cloud was transforming the architecture of on-premises data centers,” said Jensen Huang, founder and CEO of Nvidia. “One of the world’s leading server providers makes Nvidia Blackwell RTX Pro Servers the standard platform for enterprise and industrial AI.”
Accelerate GPU for business workloads
Millions of servers are sold for business operations each year, with most still using “traditional” CPUs. The new RTX Pro server provides accelerated system GPUs and improves performance in analysis, simulation, video processing and rendering, the company says. Nvidia said Server Edition GPUs can deliver 18 times more energy efficient and 45 times more performance than CPU-only systems.
The RTX Pro line is aimed at companies building “AI factories” where space, power and cooling may be limited. The server also provides the infrastructure for NVIDIA’s AI data platform for storage systems. For example, Dell has updated its AI data platform to use NVIDIA designs. The PowerEdge R7725 server comes with two RTX Pro 6000 GPUs, NVIDIA AI Enterprise software, and NVIDIA networking.
The new 2U server, which can accommodate up to eight GPU units, was among those announced at Computex in May.
Blackwell architecture features
The new server is built around Nvidia’s Blackwell Architecture.
- A 5th generation tensor core with FP4 accuracy and a second generation transformer engine that can perform inference at up to six times the speed of an L40S GPU.
- Up to 4x the performance of the L40S GPU, 4th generation RTX technology for photo rendering.
- Virtualization and NVIDIA multi-instance GPU technology allow four separate workloads per GPU.
- Improved energy efficiency to reduce data center power usage.
For physical AI and robotics
Nvidia’s Omniverse library and Cosmos World Foundation models for RTX Pro servers can perform digital twin simulations, robot training routines, and large-scale synthetic data creation. It also supports Nvidia Metropolis Blueprints for video search and summary and vision language models, among tools for use in physical environments.
Nvidia’s updated its Omniverse and Cosmos products with new Omniverse SDKS and compatibility with Mujoco (MJCF) and Universal Scene Description (OpenUSD). The company says this will allow more than 250,000 MJCF developers to run robot simulations on the platform. The new Omniverse Nurec library brings 3D Gaussian Splatz of raytraced for model structures from sensor data, and adds new OpenUSD-based schemas for robots and sensors.
NUREC rendering is already integrated into the Carla Autonomous Vehicle Simulator and is being adopted by companies like Foretellix, which use it to generate synthetic AV test data. Used by car manufacturers such as Ford and Porsche, Voxel51’s Fiftone Data Engine supports Nurec. Boston Dynamics, Figure AI, Hexagons, and Amazon Devices & Services are one of those that already employ libraries and frameworks.
COSMOS WFMS has been downloaded over 2 million times. This software helps you generate synthetic training data for your robot using text, images, or video prompts. The new Cosmos Transfer-2 model speeds up the generation of image data from spatial inputs such as simulation scenes and depth maps. Companies such as Lightwheel, Moon Surgical and Skild AI are beginning to create large-scale training data using Cosmos Transfer-2.
Nvidia has also introduced Cosmos Reason, a 7 billion parameter vision language model to help robots and AI agents combine prior knowledge and understanding of physics. It can automate dataset curation, support multi-step robot task planning, and run video analysis systems. Nvidia’s proprietary robotics and drive teams use data filtering and annotation reasons for COSMOS reasons, while Uber and Magna deploy them in autonomous vehicles, traffic surveillance and industrial inspection systems.
AI agents and large-scale deployments
The RTX Pro server can run the newly announced Llama Nemotron supermodel. Running with NVFP4 accuracy on a single RTX Pro 6000 GPU, it offers 3x the price performance of FP8 on an NVIDIA H100 GPU.
(Photo by photo)
See: nvidia regains the most valuable company title with AI momentum
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London. The comprehensive event will be held in collaboration with other major events, including the Intelligent Automation Conference, Blockx, Digital Transformation Week, and Cyber Security & Cloud Expo.
Check out other upcoming Enterprise Technology events and webinars with TechForge here.
