| Product Code: | |
| Options: | |
| Qty: | |
| Unit Price: | £ |
| Part No: | NVTCSRTXPRO6000SE-PB |
| Manufacturer No: | TCSRTXPRO6000SE-PB |
| Delivery: | Call for Leadtime |
![]() |
The NVIDIA TCSRTXPRO6000SE-PB server graphics card delivers high-performance GPU computing for data centre environments, artificial intelligence, scientific computing and advanced visual processing. Powered by NVIDIA Blackwell, it combines 96GB of GDDR7 memory with 24,064 CUDA cores, fifth-generation Tensor Cores and fourth-generation Ray Tracing Cores. With PCI Express 5.0 connectivity, passive cooling and up to 4 PFLOPS of FP4 AI performance, it provides substantial processing capacity for demanding server applications and GPU-accelerated workloads.
Powered by NVIDIA Blackwell, the RTX PRO 6000 combines dedicated processing resources for artificial intelligence, graphics rendering and ray tracing. These capabilities support demanding server applications, including generative AI, scientific computing, advanced visualisation and GPU-accelerated data processing.
The 96GB of GDDR7 memory provides substantial capacity for large AI models, extensive datasets, complex simulations and memory-intensive computing tasks. Error-Correcting Code (ECC) helps detect and correct memory errors, supporting data integrity during sustained server workloads.
Featuring 24,064 CUDA cores, the graphics card delivers extensive parallel processing capacity for compatible applications. CUDA acceleration supports scientific calculations, data analysis, rendering, simulation and other GPU-intensive computing workloads.
The 752 fifth-generation Tensor Cores provide dedicated acceleration for artificial intelligence and deep learning applications. Support for advanced precision formats, including FP4, helps increase processing throughput in compatible AI development, training and inference workloads.
The graphics card delivers up to 4 PFLOPS of peak FP4 AI performance. This processing capacity supports demanding generative AI, machine learning, deep learning and model inference applications. Performance depends on supported precision formats and workload configurations.
With 188 fourth-generation RT Cores, the card accelerates ray tracing calculations used in professional rendering and visualisation. These dedicated processing units support realistic lighting, reflections and shadows in compatible 3D applications and graphical workloads.
The 512-bit memory interface provides bandwidth of up to 1,597GB/s. This allows the GPU to access large volumes of graphics data at high speed, supporting complex models, extensive datasets, high-resolution content and memory-intensive AI applications.
The PCI Express 5.0 x16 interface provides high-speed communication between the graphics card and a compatible server motherboard. With bidirectional bandwidth of up to 128GB/s, it supports demanding data transfers between GPU resources and other system components.
Four DisplayPort 2.1 connectors provide display connectivity for supported visualisation and monitoring applications. These outputs support multiple high-resolution displays when used with compatible monitors and system configurations.
Multi-Instance GPU (MIG) technology allows the available GPU resources to be divided into separate instances for compatible workloads. The card supports up to four 24GB instances, allowing computing resources to be allocated across multiple applications or users.
Four ninth-generation NVIDIA NVENC encoders and four sixth-generation NVDEC decoders provide dedicated video processing capabilities. Four JPEG engines also support accelerated image processing tasks. These resources are useful for compatible media production, video processing and streaming applications.
The Server Edition uses passive cooling without an onboard cooling fan. It relies on the server chassis airflow system to remove heat generated during intensive GPU operations. Compatible server platforms must provide sufficient directed airflow to maintain appropriate operating temperatures.
The graphics card supports configurable power consumption between 400W and 600W, depending on the server configuration. Its PCIe CEM5 16-pin power connector requires a compatible power delivery system capable of supporting the selected operating power level.
The full-height, full-length graphics card measures approximately 4.4 inches high by 10.5 inches long and occupies two expansion slots. Its dimensions and passive cooling requirements make server chassis compatibility, expansion clearance and airflow important considerations before installation.
Confidential computing support provides hardware-based security capabilities for compatible workloads. This helps protect sensitive information during processing and is particularly relevant to server environments handling private datasets, proprietary information and AI applications.
Secure Boot with Root of Trust provides security verification during the system startup process. This helps establish confidence in the server's boot components and supports security-conscious data centre environments.
The combination of CUDA processing, high-bandwidth GDDR7 memory and dedicated Tensor Cores supports GPU-accelerated scientific and engineering applications. Compatible workloads include computational research, complex simulations, data analysis and advanced modelling.
With 96GB of graphics memory and up to 4 PFLOPS of FP4 AI performance, the card supports demanding generative AI and inference workloads. Its processing capabilities are suitable for compatible large language models, machine learning applications and other data-intensive AI tasks.
Dedicated ray tracing hardware and substantial parallel processing resources support professional graphics and visual computing applications. The card can be used for compatible rendering, 3D visualisation, engineering workloads and complex graphical environments.
Installation requires a compatible PCI Express 5.0 x16 expansion slot, sufficient dual-slot clearance and a suitable PCIe CEM5 16-pin power connection. The server must also provide appropriate power delivery and directed chassis airflow to accommodate the passive cooling system. NVIDIA recommends system memory equal to or greater than the GPU memory, with twice the GPU memory recommended.
The NVIDIA TCSRTXPRO6000SE-PB server graphics card delivers high-performance GPU computing for data centre environments, artificial intelligence, scientific computing and advanced visual processing. Powered by NVIDIA Blackwell, it combines 96GB of GDDR7 memory with 24,064 CUDA cores, fifth-generation Tensor Cores and fourth-generation Ray Tracing Cores. With PCI Express 5.0 connectivity, passive cooling and up to 4 PFLOPS of FP4 AI performance, it provides substantial processing capacity for demanding server applications and GPU-accelerated workloads.
Powered by NVIDIA Blackwell, the RTX PRO 6000 combines dedicated processing resources for artificial intelligence, graphics rendering and ray tracing. These capabilities support demanding server applications, including generative AI, scientific computing, advanced visualisation and GPU-accelerated data processing.
The 96GB of GDDR7 memory provides substantial capacity for large AI models, extensive datasets, complex simulations and memory-intensive computing tasks. Error-Correcting Code (ECC) helps detect and correct memory errors, supporting data integrity during sustained server workloads.
Featuring 24,064 CUDA cores, the graphics card delivers extensive parallel processing capacity for compatible applications. CUDA acceleration supports scientific calculations, data analysis, rendering, simulation and other GPU-intensive computing workloads.
The 752 fifth-generation Tensor Cores provide dedicated acceleration for artificial intelligence and deep learning applications. Support for advanced precision formats, including FP4, helps increase processing throughput in compatible AI development, training and inference workloads.
The graphics card delivers up to 4 PFLOPS of peak FP4 AI performance. This processing capacity supports demanding generative AI, machine learning, deep learning and model inference applications. Performance depends on supported precision formats and workload configurations.
With 188 fourth-generation RT Cores, the card accelerates ray tracing calculations used in professional rendering and visualisation. These dedicated processing units support realistic lighting, reflections and shadows in compatible 3D applications and graphical workloads.
The 512-bit memory interface provides bandwidth of up to 1,597GB/s. This allows the GPU to access large volumes of graphics data at high speed, supporting complex models, extensive datasets, high-resolution content and memory-intensive AI applications.
The PCI Express 5.0 x16 interface provides high-speed communication between the graphics card and a compatible server motherboard. With bidirectional bandwidth of up to 128GB/s, it supports demanding data transfers between GPU resources and other system components.
Four DisplayPort 2.1 connectors provide display connectivity for supported visualisation and monitoring applications. These outputs support multiple high-resolution displays when used with compatible monitors and system configurations.
Multi-Instance GPU (MIG) technology allows the available GPU resources to be divided into separate instances for compatible workloads. The card supports up to four 24GB instances, allowing computing resources to be allocated across multiple applications or users.
Four ninth-generation NVIDIA NVENC encoders and four sixth-generation NVDEC decoders provide dedicated video processing capabilities. Four JPEG engines also support accelerated image processing tasks. These resources are useful for compatible media production, video processing and streaming applications.
The Server Edition uses passive cooling without an onboard cooling fan. It relies on the server chassis airflow system to remove heat generated during intensive GPU operations. Compatible server platforms must provide sufficient directed airflow to maintain appropriate operating temperatures.
The graphics card supports configurable power consumption between 400W and 600W, depending on the server configuration. Its PCIe CEM5 16-pin power connector requires a compatible power delivery system capable of supporting the selected operating power level.
The full-height, full-length graphics card measures approximately 4.4 inches high by 10.5 inches long and occupies two expansion slots. Its dimensions and passive cooling requirements make server chassis compatibility, expansion clearance and airflow important considerations before installation.
Confidential computing support provides hardware-based security capabilities for compatible workloads. This helps protect sensitive information during processing and is particularly relevant to server environments handling private datasets, proprietary information and AI applications.
Secure Boot with Root of Trust provides security verification during the system startup process. This helps establish confidence in the server's boot components and supports security-conscious data centre environments.
The combination of CUDA processing, high-bandwidth GDDR7 memory and dedicated Tensor Cores supports GPU-accelerated scientific and engineering applications. Compatible workloads include computational research, complex simulations, data analysis and advanced modelling.
With 96GB of graphics memory and up to 4 PFLOPS of FP4 AI performance, the card supports demanding generative AI and inference workloads. Its processing capabilities are suitable for compatible large language models, machine learning applications and other data-intensive AI tasks.
Dedicated ray tracing hardware and substantial parallel processing resources support professional graphics and visual computing applications. The card can be used for compatible rendering, 3D visualisation, engineering workloads and complex graphical environments.
Installation requires a compatible PCI Express 5.0 x16 expansion slot, sufficient dual-slot clearance and a suitable PCIe CEM5 16-pin power connection. The server must also provide appropriate power delivery and directed chassis airflow to accommodate the passive cooling system. NVIDIA recommends system memory equal to or greater than the GPU memory, with twice the GPU memory recommended.