NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 96GB GDDR7

NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition: Scalable Power for the Next Era of AI The NVIDIA RTX PRO™ 6000 Blackwell Max-Q Workstation Edition represents a major leap forward in high-density professional computing. Designed for professionals tackling the most demanding AI, data science, virtual production, and simulation workloads, this GPU delivers desktop workstation performance within a power-efficient 300W total board power envelope. Its dual-slot, full-height physical profile is optimized for multi-GPU density, allowing enterprises to scale up to four RTX PRO 6000 Max-Q GPUs inside a single workstation chassis.

Blackwell Architecture & Massive 96GB GDDR7 Memory Built on the NVIDIA Blackwell architecture, the card incorporates 24,064 CUDA cores and new Streaming Multiprocessors with neural shaders that integrate neural networks directly into programmable shaders. Equipped with 96 GB of GDDR7 memory featuring Error-Correcting Code (ECC) and an unprecedented 1.8 TB/s (1792 GB/s) of memory bandwidth across a 512-bit bus, professionals can train and fine-tune domain-specific LLMs locally, process colossal datasets, and render complex multi-app workflows without hitting memory walls.

Groundbreaking AI & Neural Graphics Fifth-generation Tensor Cores deliver up to 3x the performance of the previous generation, adding native support for FP4 precision and DLSS 4 Multi Frame Generation to achieve up to 3511 AI TOPS. Fourth-generation Ray Tracing Cores double the ray-triangle intersection rate, unleashing RTX Mega Geometry to render up to 100x more ray-traced triangles for physically accurate scenes and engineering simulations.

Enterprise Scalability & Multi-Display Precision With Universal Multi-Instance GPU (MIG) support, a single card can be partitioned into up to four isolated hardware instances (up to 4x 24GB, 2x 48GB, or 1x 96GB) with guaranteed quality of service for shared compute environments. Video workflows are accelerated by quad 9th Gen NVENC and quad 6th Gen NVDEC engines with native 4:2:2 encoding/decoding. Four DisplayPort 2.1b connectors provide next-generation display connectivity, driving multi-monitor setups at up to 8K at 240Hz or 16K at 60Hz with rich HDR color depth.

Key Specifications:

• Architecture & Cores: NVIDIA Blackwell architecture with 24,064 CUDA Cores and dedicated AI Management Processor.
• Compute Performance: 3511 AI TOPS (FP4 with sparsity), 110 TFLOPS single-precision, and 333 TFLOPS RT Core performance.
• Memory Configuration: 96 GB GDDR7 with ECC, 512-bit interface, 1792 GB/s (1.8 TB/s) bandwidth.
• Power & Thermals: 300 W total board power, 1x PCIe CEM5 16-pin power connector, active dual-slot cooling (4.4” H x 10.5” L).
• System Interface: PCIe 5.0 x16 for maximum host-to-device throughput.PDF+ 1
• Display Connectivity: 4x DisplayPort 2.1b outputs supporting high-refresh 4K, 8K, and up to 16K visual setups.
• Media Engines: 4x NVENC (9th Gen) and 4x NVDEC (6th Gen) with 4:2:2 support
• MIG Partitioning: Configurable into up to 4x 24GB, 2x 48GB, or 1x 96GB isolated instances.
• APIs Supported: CUDA 13.0, DirectX 12, Shader Model 6.6, Vulkan 1.3, OpenGL 4.6, OpenCL 3.0, and DirectCompute.