The NVIDIA RTX PRO 6000 Blackwell Server Edition is a professional-grade data center GPU designed to accelerate enterprise AI, visual computing, and scientific workloads .

Here is a brief overview of its key specifications and features:

Key Characteristics

Based on the groundbreaking NVIDIA Blackwell architecture, this GPU provides a significant generational leap in performance. Its fifth-generation Tensor Cores are designed to deliver up to 3x the AI performance of the previous generation, supporting FP4 precision, which is crucial for accelerating local LLMs and generative AI applications .

The card is built for massive data handling. The large 96 GB VRAM capacity allows it to tackle extensive 3D models, complex AI datasets, and multi-application workflows . It also supports Multi-Instance GPU (MIG) technology, enabling the GPU to be partitioned into up to four fully isolated instances to maximize resource utilization for multiple users .

As a Server Edition card, it is designed for data center environments. It features a passive cooling solution, meaning it relies on server chassis airflow for cooling, and is optimized for 24/7 operation in systems from leading vendors .

Target Workloads

The RTX PRO 6000 Blackwell Server Edition is optimized for a wide range of demanding professional workloads. These include agent-based AI, scientific computing, high-fidelity rendering, and video processing, leveraging its powerful AI and visual computing capabilities . It also excels in professional visualization and simulation workflows, supporting applications in fields like engineering, healthcare, and media and entertainment, and is capable of driving high-resolution displays via its DisplayPort 2.1 outputs . Furthermore, it is tailored for Universal Scene Description (OpenUSD)-based 3D workflows and synthetic data generation within the NVIDIA Omniverse platform, facilitating the development of industrial digitalization applications .