Unleash the power of next‑generation workstation graphics with the NVIDIA RTX Pro 6000 Blackwell 96GB GDDR7 PCI Express 5.0 x16 Graphics Card. Engineered for creators, engineers, and AI researchers, this card delivers unmatched speed, precision, and reliability, turning even the most demanding projects into smooth, real‑time experiences.
NVIDIA RTX Pro 6000 Blackwell 96GB GDDR7 PCI Express 5.0 x16 Graphics Card
Featuring the cutting‑edge NVIDIA Blackwell architecture, the card packs 24,064 CUDA cores, 4th‑generation RT cores, and 5th‑generation Tensor cores. With 125 TFLOPS of FP32 performance and up to 4,000 AI TOPS, you can accelerate AI training, large‑language‑model fine‑tuning, and complex simulations without bottlenecks.
The massive 96 GB GDDR7 memory with ECC and a 512‑bit interface provides a staggering 1.8 TB/s bandwidth, ensuring seamless handling of massive datasets and ultra‑high‑resolution textures. Dual‑slot, extended‑height flow‑through cooling and a 600 W TDP keep the GPU stable under continuous heavy loads.
Benefits
- Maximum AI compute power with up to 4,000 AI TOPS for faster model training.
- Ultra‑high memory capacity (96 GB) and bandwidth (1.8 TB/s) for large datasets.
- 4K‑to‑16K display support with four DisplayPort 2.1b outputs for immersive visualisation.
- Robust reliability with ECC memory and advanced cooling for 24/7 workstation use.
Uses of the product
- AI training, inference, and large‑language‑model fine‑tuning.
- Professional 3D rendering, VFX, and animation pipelines.
- Engineering simulations, CAE/CAD, and scientific computing.
- Multi‑display workstations supporting >4 × 4K@120 Hz or up to 16K@60 Hz.
Features
- 24,064 CUDA cores, 4th‑gen RT cores, and 5th‑gen Tensor cores with FP4 sparsity.
- PCIe 5.0 ×16 interface for maximum data throughput.
- Dual‑slot, extended‑height (XHFL) flow‑through cooling with a single 16‑pin CEM5 connector.
- Four 9th‑gen NVENC and four 6th‑gen NVDEC video engines for real‑time encoding/decoding.
Specifications
| Product |
NVIDIA RTX PRO 6000 Blackwell (Workstation) |
| GPU Architecture |
NVIDIA Blackwell |
| CUDA Cores |
24,064 |
| Tensor Cores |
5th Generation Tensor Cores (FP4 support) |
| Ray Tracing Cores |
4th Generation RT Cores |
| AI Compute |
Up to 4,000 AI TOPS (FP4 with sparsity) |
| FP32 Performance |
125 TFLOPS |
| RT Core Performance |
380 TFLOPS |
| GPU Memory |
96 GB GDDR7 with ECC |
| Memory Interface |
512-bit |
| Memory Bandwidth |
1,792 GB/s (≈1.8 TB/s) |
| PCIe Interface |
PCI Express 5.0 x16 |
| Board Power (TDP) |
600 W |
| Power Connector |
1 × CEM5 16-pin |
| Cooling Solution |
Dual-slot, extended-height (XHFL) flow-through cooling |
| Card Dimensions |
5.4 in. height × 12 in. length, dual-slot |
| Display Outputs |
4 × DisplayPort 2.1b |
| Supported Max Resolution |
Up to 16K @ 60 Hz (DP 2.1 UHBR) |
| Multi-Display Support |
>4 × 4K @ 120 Hz
>4 × 5K @ 60 Hz
>2 × 8K @ 60 Hz
|
| Video Encoders |
4 × NVENC (9th generation) |
| Video Decoders |
4 × NVDEC (6th generation) |
| MIG (Multi-Instance GPU) |
Up to: 4× 24 GB
2× 48 GB
1× 96 GB partitions
|
| API / Software Support |
DirectX 12, Shader Model 6.6, Vulkan 1.3, OpenGL 4.6, OpenCL 3.0, CUDA 12.8 |
| Use Cases |
AI training & inference, LLM fine-tuning, 3D rendering, simulation, engineering (CAE/CAD), VFX, scientific compute |
| SKU |
900-5G144-2500-000 |
