NVIDIA RTX PRO 6000 Server Blackwell 96GB GDDR7

Detailed Product Description

The NVIDIA RTX PRO 6000 Blackwell Server Edition 96GB GDDR7 is a high-performance professional GPU designed for enterprise data centers, AI infrastructure, scientific computing, rendering, 3D graphics, video processing and virtual workstation workloads. Built on the advanced NVIDIA Blackwell architecture, it combines powerful AI computing with professional visual computing capabilities for demanding enterprise applications.

Featuring 96GB of GDDR7 memory with ECC, the NVIDIA RTX PRO 6000 Server Edition is designed to handle large AI models, complex datasets, demanding 3D projects and multi-application workloads. Its high memory capacity and bandwidth make it suitable for workloads that require substantial GPU resources in enterprise server environments.

NVIDIA Blackwell Architecture

The RTX PRO 6000 Server Edition features the NVIDIA Blackwell architecture, with 24,064 CUDA cores, fifth-generation Tensor Cores and fourth-generation RT Cores. NVIDIA designed the GPU to accelerate both AI and visual computing workloads, including agentic AI, generative AI, scientific computing, rendering and engineering simulation.

96GB GDDR7 ECC Memory

The GPU provides 96GB GDDR7 memory with ECC and a 512-bit memory interface. NVIDIA specifies up to 1,597 GB/s memory bandwidth, enabling high-speed access to large datasets and complex workloads. This large memory capacity is particularly useful for AI inference, data analytics, 3D graphics and scientific computing.

Enterprise AI & Data Center Workloads

The NVIDIA RTX PRO 6000 Server Edition is designed for enterprise AI workloads including LLM inference, agentic AI, generative AI, data analytics and AI-powered applications. NVIDIA also supports Multi-Instance GPU technology, allowing the GPU to be divided into isolated GPU instances for multiple workloads.

Professional Rendering & Visual Computing

The GPU combines AI and RTX graphics technologies to accelerate 3D rendering, professional visualization, animation, engineering design and simulation. Its fourth-generation RT Cores provide dedicated hardware acceleration for ray tracing and advanced visual workloads.

Scientific Computing & Data Analytics

With high FP32 and AI compute performance, the RTX PRO 6000 Server Edition is suitable for scientific computing, engineering simulation, data analytics, financial workloads, genomics and large-scale computational applications.

Video Processing

The GPU includes four 9th-generation NVIDIA NVENC encoders and four 6th-generation NVDEC decoders, enabling accelerated professional video encoding, decoding and video-processing workloads. NVIDIA also highlights support for 4:2:2 H.264 and HEVC workflows.

Server & Data Center Deployment

The RTX PRO 6000 Server Edition is designed specifically for server and data-center deployments, including high-density multi-GPU systems. NVIDIA specifies air-cooled and liquid-cooled configurations, allowing system builders to select the appropriate thermal solution for their infrastructure.

Technical Specifications

Ideal For

• Enterprise Artificial Intelligence
• LLM Inference
• Generative AI
• Agentic AI
• Machine Learning
• Data Analytics
• Scientific Computing
• Engineering Simulation
• 3D Rendering
• Professional Visualization
• Virtual Workstations
• Video Processing
• High-Performance Computing
• Enterprise Data Centers
• Multi-GPU Server Infrastructure

Key Features

• NVIDIA Blackwell Architecture
• 96GB GDDR7 ECC Memory
• 24,064 CUDA Cores
• 512-bit Memory Interface
• 1,597 GB/s Memory Bandwidth
• 188 Fourth-Generation RT Cores
• 5th Generation Tensor Cores
• 4× 9th Gen NVENC
• 4× 6th Gen NVDEC
• PCIe Gen 5 x16
• 4× DisplayPort 2.1
• MIG Support
• 400–600W Configurable Power
• Designed for Enterprise AI, Data Center, Rendering, Scientific Computing & Virtual Workstations