Surprisingly Useful AI Article Enhancements
Nvidia has announced that its RTX Pro 6000 Blackwell Server Edition GPUs will be available in new 2U mainstream servers from major system providers, targeting accelerated computing across enterprise workloads.
The company said Cisco, Dell Technologies, HPE, Lenovo and Supermicro will offer multiple configurations of RTX Pro Servers designed to deliver performance and efficiency gains for applications ranging from agentic Ai and content creation to data analytics, graphics and simulation.
Shift Toward Accelerated Computing
Nvidia CEO, Jensen Huang, said the move reflects a broader shift from CPU-based systems to accelerated platforms.
“Ai is reinventing computing for the first time in 60 years. With the world’s leading server providers, we’re making Nvidia Blackwell RTX Pro Servers the standard platform for enterprise and industrial Ai,” Huang said.
The RTX Pro Servers aim to bring GPU acceleration to workloads traditionally run on CPUs, including simulation, analytics and rendering. Nvidia said they can deliver up to 45 times better performance and 18 times higher energy efficiency compared with CPU-only 2U systems.
Targeting Constrained Data Centres
The systems are intended for customers building Ai-focused infrastructure within space, power and cooling limitations. They also form the backbone for the Nvidia Ai Data Platform, a reference design for enterprise Ai storage systems.
At SIGGRAPH, Dell announced updates to its Ai Data Platform, integrated with Nvidia’s reference design, alongside the PowerEdge R7725 server featuring two RTX Pro 6000 GPUs, Nvidia Ai Enterprise software and networking.
The new 2U servers join models announced at Computex in May, expanding a portfolio that supports configurations with two, four or eight GPUs.
Blackwell Architecture Features
The RTX Pro 6000 Blackwell GPUs incorporate fifth-generation Tensor Cores and a second-generation Transformer Engine with FP4 precision, delivering up to six times faster inference than the previous L40S GPU.
Other features include:
- Fourth-generation RTX technology for photorealistic rendering, with up to four times higher performance than the L40S
- Virtualisation and Multi-Instance GPU capability, allowing four isolated instances per GPU for multi-user Ai deployments
- Improved performance per watt for more efficient data centre operation
Physical Ai And Robotics
Nvidia said the new servers are suited to workloads using Omniverse libraries and Cosmos world foundation models, including digital twins, robotics simulation and large-scale synthetic data generation.
The company said simulation and synthetic data workflows can run up to four times faster than on L40S-based systems.
The servers also support new Nvidia blueprints for video search and summarisation through the Metropolis platform, as well as vision language models for productivity and safety use cases.
Ai Agents And Enterprise Deployment
RTX Pro Servers are certified for Nvidia Ai Enterprise software and positioned for running Ai agents using reasoning models such as Llama Nemotron Super. Nvidia said the model delivers up to three times better price performance using NVFP4 on a single RTX Pro 6000 compared with FP8 on an H100 GPU.
The Blackwell platform is backed by Nvidia’s CUDA-X libraries, developer ecosystem and a catalogue of nearly 6,000 applications.
Related: Best Business Laptops for work & school
Related: Best Gaming Laptops
Related: Best Portable Laptop
Availability
Cisco, Dell, HPE, Lenovo and Supermicro will offer Nvidia-certified RTX Pro Servers, with additional vendors including Advantech, Aetina, Airves, ASRock Rack, Asus, Compal, Foxconn, Gigabyte, Inventec, MiTAC Computing, MSI, Pegatron, Quanta Cloud Technology, Wistron and Wiwynn.
Configurations with eight GPUs in 4U form factors are available now, with the new 2U mainstream models expected later this year. More information is available on the Nvidia RTX Pro Servers page.
Last Updated on August 12, 2025 by Nick Ross
