with NVIDIA RTX PRO™ 6000 Blackwell Server Edition
The NVIDIA RTX PRO™ 6000 Blackwell Server Edition delivers scalable acceleration for enterprise infrastructure. Engineered to address intensive computational requirements, this passively cooled, dual-slot GPU provides the compute power required to scale generative AI models, process batch rendering, and sustain data center simulations.
Built on the NVIDIA Blackwell architecture, the NVIDIA RTX PRO™ 6000 Blackwell Server Edition accelerates data throughput. Utilizing a PCI Express Gen5 x16 interface alongside 96 GB of GDDR7 memory, the architecture establishes a 1,597 GB/s peak memory bandwidth. This structural capacity ensures high-parameter workloads and large-scale 3D visual datasets remain fully saturated in memory.
Designed for cloud-native deployment, the NVIDIA RTX PRO™ 6000 Blackwell incorporates virtualization features at the hardware level. Multi-Instance GPU (MIG) capabilities allow administrators to partition the physical GPU into up to four isolated instances, each with dedicated memory, cache, and compute cores. Single Root I/O Virtualization (SR-IOV) supports up to 48 virtual functions, enabling virtual machine consolidation on a single host node.
Engineered for server integration, the card operates within a 600 W power envelope and utilizes a bidirectional heat sink that accepts airflow in either left-to-right or right-to-left directions. Deployment can be precisely tailored using three operating modes. The default Display Off mode prioritizes pure compute scaling, while two distinct Display On modes allow systems to drive up to four VESA® DisplayPort™ 2.1b monitors for advanced enterprise visualization.
