NVIDIA RTX PRO Servers With Blackwell Coming to World’s Most Popular Enterprise Systems

VANCOUVER, British Columbia — Aug. 11, 2025 — SIGGRAPH—NVIDIA today announced that the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU is coming to the world’s most popular enterprise servers, speeding the shift from traditional CPU systems to accelerated computing platforms.

With these new 2U mainstream servers, enterprises worldwide can harness the NVIDIA Blackwell architecture in the most widely adopted rack-mounted systems for breakthrough performance and efficiency in their data centers.

Global system partners including Cisco, Dell Technologies, HPE, Lenovo and Supermicro will offer the 2U NVIDIA RTX PRO Servers — available in multiple configurations — to bring universal acceleration for enterprise workloads spanning agentic AI, content creation, data analytics, graphics, scientific simulation, as well as industrial and physical AI.

“AI is reinventing computing for the first time in 60 years — what started in the cloud is now transforming the architecture of on-premises data centers,” said Jensen Huang, founder and CEO of NVIDIA. “With the world’s leading server providers, we’re making NVIDIA Blackwell RTX PRO Servers the standard platform for enterprise and industrial AI.”

NVIDIA RTX PRO Server Family Brings Accelerated Systems to Data Centers
Every year, enterprises buy millions of servers for business workloads. They can now refresh these systems with accelerated servers, as AI becomes increasingly critical to operations.

RTX PRO Servers bring GPU acceleration to traditional CPU-based workloads like data analytics, simulation, video processing and graphics rendering — delivering up to 45x better performance, which results in 18x higher energy efficiency with lower cost of ownership compared with CPU-only 2U systems.

NVIDIA RTX PRO Servers represent a new class of on-premises infrastructure that brings groundbreaking Blackwell performance to enterprise customers building AI factories with space-, power- and cooling-constrained data centers.

These systems also provide the infrastructure backbone for the NVIDIA AI Data Platform, a customizable reference design for building modern storage systems for enterprise agentic AI. At SIGGRAPH, Dell is announcing updates to the Dell AI Data Platform — integrated with the NVIDIA AI Data Platform reference design — along with Dell PowerEdge R7725 2U servers featuring two RTX PRO 6000 GPUs, NVIDIA AI Enterprise software and NVIDIA networking.

The new 2U mainstream systems join a family of RTX PRO Servers announced in May at COMPUTEX, providing a full spectrum of rack-mounted designs capable of supporting two, four or eight NVIDIA RTX PRO 6000 Blackwell GPUs. The servers are ideal for enterprises looking to optimize performance, efficiency and costs.

Breakthrough AI Performance for Enterprise Data Centers
The new RTX PRO Servers provide a versatile, high-performance platform for a broad range of applications including AI and machine learning, data analytics, 3D graphics and scientific simulation. These servers incorporate the latest Blackwell architecture innovations, including:

• Fifth-generation Tensor Cores and second-generation Transformer Engine with support for FP4 precision, delivering up to 6x faster inference performance compared with the previous-generation NVIDIA L40S GPU
• Fourth-generation NVIDIA RTX technology for photorealistic rendering and visualization to deliver up to 4x higher performance than the L40S GPU
• Enterprise-grade scale for multi-user AI deployments, using virtualization and NVIDIA Multi-Instance GPU technology for four fully isolated instances per GPU
• Improved performance per watt for sustainable data center operations

Accelerating Physical AI and Robotics Workloads
RTX PRO Servers running NVIDIA Omniverse libraries and NVIDIA Cosmos world foundation models enable physical AI developers to build and deploy applications including digital twins for factory and robot simulation or large-scale synthetic data generation.

RTX PRO Servers can run simulation and synthetic data generation workflows up to 4x faster than systems with L40S GPUs.

In addition, to make spaces smarter and more secure, RTX PRO Servers can now support advanced blueprints — including the latest NVIDIA Blueprint for video search and summarization, part of the NVIDIA Metropolis platform — as well as vision language models and synthetic data generation extensions to boost productivity and enhance safety across physical AI environments.

Speed and Scale for Enterprise AI and Agents
All RTX PRO Servers are certified for NVIDIA AI Enterprise — the software layer that accelerates and secures AI development and deployment.

RTX PRO Servers are ideal for running AI agents that use AI reasoning models to act and automate complex tasks. Such models include Llama Nemotron Super, also announced today, which delivers up to 3x price performance when running with NVFP4 on a single NVIDIA RTX PRO 6000 GPU compared with FP8 on NVIDIA H100 GPUs. This enables more accurate reasoning at a lower cost.

The Blackwell platform builds on NVIDIA’s ecosystem of powerful development tools, NVIDIA CUDA-X libraries, over 6 million developers and nearly 6,000 applications to scale performance across thousands of GPUs.

Availability
Global system makers Cisco, Dell, HPE, Lenovo and Supermicro will offer a wide range of NVIDIA-Certified RTX PRO Servers. Additional data center system partners bringing RTX PRO Servers to market include Advantech, Aetina, Airves, ASRock Rack, ASUS, Compal, Foxconn, GIGABYTE, Inventec, MiTAC Computing, MSI, PEGATRON, Quanta Cloud Technology (QCT), Wistron and Wiwynn.

Customers can order RTX PRO Servers today from these system makers and channel partners worldwide. Configurations with eight RTX PRO 6000 GPUs in 4U form factors are available now. The 2U mainstream RTX PRO Servers are expected to be available later this year. Learn more about RTX PRO Servers.

Watch the NVIDIA Research special address at SIGGRAPH.

Source: Nvidia