Percepta / Workstation GPU
SKU:900-5G153-2500-000
96GB ECC GDDR7 and a 300W active-blower design for professional workstations. New card in retail packaging. Paid-in-full pre-order; see the delivery estimate and agreed deadline below.
Explore the detailsPre-order
Max-Q Workstation Edition / Retail
The RTX PRO 6000 Max-Q Workstation Edition brings 96 GB of ECC GDDR7 memory to local LLM inference and development, GPU data science and 3D production. Its 300 W active blower design prioritises power and space efficiency for workstations built around one or several GPUs.
Each GPU provides memory for model weights, analysis datasets and detailed scenes. Fifth-generation Tensor cores support FP4 inference in compatible LLM software; CUDA-X data science libraries enable GPU-accelerated data exploration and model evaluation. Capacity remains 96 GB per card, with workload size depending on precision and working memory.
The full-height, dual-slot card is 10.5 inches long, compared with the extended-height, 12-inch, 600 W Workstation Edition. Its 300 W blower-cooled design targets denser multi-GPU workstations, where separate cards can run independent AI experiments or render jobs. Check slot spacing, airflow and total system power with your system manufacturer.
Fourth-generation RT cores support ray-traced lighting for architectural visualisation and product design reviews. Ninth-generation NVENC and sixth-generation NVDEC handle video encoding and decoding, including 4:2:2 H.264/HEVC for editing and AV1 for streaming in compatible software. Four DisplayPort outputs connect monitors for 3D and video work.
Specifications per card
| Edition | Max-Q Workstation Edition |
|---|---|
| GPU architecture | NVIDIA Blackwell / GB202 |
| CUDA compute capability | 12.0 |
| CUDA cores | 24,064 |
| Tensor cores | 752 / fifth generation |
| RT cores | 188 / fourth generation |
| Graphics memory | 96 GB GDDR7 with ECC |
| Memory interface | 512-bit |
| Memory bandwidth | 1,792 GB/s |
| Host interface | PCIe 5.0 x16 |
| Total board power | 300 W |
|---|---|
| Cooling | Active blower fan |
| Form factor | Full-height, dual-slot |
| Nominal dimensions (height x length) | 4.4 x 10.5 inches; additional installation clearance required |
| Auxiliary power | 1 x PCIe CEM5 16-pin |
| Display connectors | 4 x DisplayPort 2.1b |
| Display operation | Direct display output; enabling MIG compute mode disables physical display outputs |
| NVLink bridge | Not supported |
| Frame synchronisation | Supports NVIDIA RTX PRO Sync; separate hardware required and not included |
| FP32 compute | 110 TFLOPS |
|---|---|
| RT core throughput | 333 TFLOPS |
| AI compute (FP4, with sparsity) | 3,511 AI TOPS |
These peak figures use GPU Boost clocks. Application performance depends on software, workload, power settings and operating conditions. AI TOPS uses effective FP4 with sparsity; compute, ray-tracing and AI figures measure different operations and should be compared like for like.
| Video encoding | 4 x ninth-generation NVENC engines |
|---|---|
| Video decoding | 4 x sixth-generation NVDEC engines |
| JPEG engines | 4 |
| Video processing | AV1 support; 4:2:2 encode/decode support with compatible software |
| Multi-Instance GPU (MIG) | Up to 4 instances; supported profiles include 4 x 24 GB, 2 x 48 GB or 1 x 96 GB |
| NVIDIA vGPU technology | Not supported on this edition |
| Manufacturer part number | 900-5G153-2500-000 |
|---|---|
| UPC | 810152850220 |
| Supplied item | One new GPU card |
| Packaging | Retail packaging |
Allow additional installation space for cabling and airflow beyond the card dimensions. MIG requires a supported driver, profile and system configuration. It partitions an individual GPU; it does not combine memory across separate cards.
The card is full height and dual slot, measuring 4.4 x 10.5 inches. Allow additional room for the power cable and blower airflow. It requires one PCIe CEM5 16-pin connection.
The 300 W rating is for the GPU. Check PSU capacity for the complete system and, for a multi-GPU build, check card spacing and cooling with the system manufacturer.
Max-Q retains 96 GB of memory and the same core counts, but uses a 300 W blower-cooled design with lower peak throughput. The 600 W Workstation Edition has a larger double-flow-through cooler. Check each card's dimensions, power and airflow requirements for your system.
Each card has its own 96 GB of graphics memory. Applications must support multi-GPU processing or model distribution to use several cards for one job. The memory is not automatically pooled, and NVLink bridging is not supported.
MIG supports up to four instances with a supported software stack. Enabling compute mode for MIG disables the physical display outputs. NVIDIA vGPU technology is not supported on Max-Q Workstation Edition; MIG does not automatically make instances independently available to separate virtual machines.
This listing is for one new NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 96 GB GPU card only, not a complete workstation. Supplied in retail packaging.
A compatible workstation, power connection, GPU drivers and application software are required. Any cables, adapters, accessories or software licences must be confirmed separately before ordering. Ask us about your installation requirements.
Send us your system details and the applications you use to discuss whether this edition suits your build.
Discuss your requirement ↗Discuss the details
Tell us about your application or the system you are putting together.
Workstation GPU
Condition: New
Workstation GPU
Condition: New - OEM/bulk packaging
Workstation GPU
Condition: New - OEM/bulk packaging
