



Leadtek Nvidia Rtx Pro 6000 Blackwell Server Edition 96Gb Gddr7 Ecc Gpu
The Leadtek NVIDIA RTX PRO 6000 Blackwell Server Edition is a professional-grade GPU built on the NVIDIA Blackwell architecture, featuring 24,064 CUDA cores, 752 Tensor Cores, and 188 RT Cores. It delivers 120 TFLOPS of single-precision performance and 3.8 PFLOPS of peak FP4 AI performance, making it suitable for demanding AI, rendering, and compute workloads.
Equipped with 96 GB of GDDR7 memory with ECC support on a 512-bit interface, the card achieves a memory bandwidth of 1.6 TB/s. It supports up to 4 Multi-Instance GPU (MIG) partitions at 24 GB each, confidential compute, and secure boot with root of trust for enhanced security in virtualised environments.
The GPU connects via PCI Express 5.0 x16 and features four DisplayPort 2.1 outputs. It has a passive thermal solution, a dual-slot form factor measuring 4.4 inches by 10.5 inches, and a configurable power consumption of 400-600W via a single PCIe CEM5 16-pin connector. The product weighs 1.194 kg and includes 4x NVENC, 4x NVDEC, and 4x JPEG engines.
| Architecture | NVIDIA Blackwell |
| CUDA Cores | 24,064 |
| Tensor Cores | 752 |
| RT Cores | 188 |
| GPU Memory | 96 GB GDDR7 with ECC |
| Memory Interface | 512-bit |
| Memory Bandwidth | 1.6 TB/s |
| Single-Precision Performance (FP32) | 120 TFLOPS |
| Peak FP4 AI Performance | 3.8 PFLOPS |
| Graphics Bus | PCI Express 5.0 x16 |
| Display Connectors | 4x DisplayPort 2.1 |
| Form Factor | 4.4" (H) x 10.5" (L), dual slot |
| Thermal Solution | Passive |
| Max Power Consumption | 400-600W (Configurable) |
| Power Connector | 1x PCIe CEM5 16-pin |
| Multi-Instance GPU | Up to 4 MIGs @ 24GB |
| Confidential Compute | Supported |
| Secure Boot with Root of Trust | Yes |
| NVENC | NVDEC | JPEG | 4x | 4x | 4x |
| Product Weight | 1.194 kg |
Original: $23,474.00
-65%$23,474.00
$8,215.90Product Information
Product Information
Shipping & Returns
Shipping & Returns
Description
The Leadtek NVIDIA RTX PRO 6000 Blackwell Server Edition is a professional-grade GPU built on the NVIDIA Blackwell architecture, featuring 24,064 CUDA cores, 752 Tensor Cores, and 188 RT Cores. It delivers 120 TFLOPS of single-precision performance and 3.8 PFLOPS of peak FP4 AI performance, making it suitable for demanding AI, rendering, and compute workloads.
Equipped with 96 GB of GDDR7 memory with ECC support on a 512-bit interface, the card achieves a memory bandwidth of 1.6 TB/s. It supports up to 4 Multi-Instance GPU (MIG) partitions at 24 GB each, confidential compute, and secure boot with root of trust for enhanced security in virtualised environments.
The GPU connects via PCI Express 5.0 x16 and features four DisplayPort 2.1 outputs. It has a passive thermal solution, a dual-slot form factor measuring 4.4 inches by 10.5 inches, and a configurable power consumption of 400-600W via a single PCIe CEM5 16-pin connector. The product weighs 1.194 kg and includes 4x NVENC, 4x NVDEC, and 4x JPEG engines.
| Architecture | NVIDIA Blackwell |
| CUDA Cores | 24,064 |
| Tensor Cores | 752 |
| RT Cores | 188 |
| GPU Memory | 96 GB GDDR7 with ECC |
| Memory Interface | 512-bit |
| Memory Bandwidth | 1.6 TB/s |
| Single-Precision Performance (FP32) | 120 TFLOPS |
| Peak FP4 AI Performance | 3.8 PFLOPS |
| Graphics Bus | PCI Express 5.0 x16 |
| Display Connectors | 4x DisplayPort 2.1 |
| Form Factor | 4.4" (H) x 10.5" (L), dual slot |
| Thermal Solution | Passive |
| Max Power Consumption | 400-600W (Configurable) |
| Power Connector | 1x PCIe CEM5 16-pin |
| Multi-Instance GPU | Up to 4 MIGs @ 24GB |
| Confidential Compute | Supported |
| Secure Boot with Root of Trust | Yes |
| NVENC | NVDEC | JPEG | 4x | 4x | 4x |
| Product Weight | 1.194 kg |





















