NVIDIA RTX PRO™ 6000 Blackwell Server Edition

NVIDIA RTX PRO™ 6000 Blackwell Server Edition 96 GB GDDR7 with ECC (900-2G153-0000-000) - 3 Years Local Warranty

Enterprise AI and Data Center GPU Performance

NVIDIA RTX PRO™ 6000 Blackwell Server Edition 96GB GDDR7 with ECC (900-2G153-0000-000) is an enterprise-grade GPU accelerator engineered for AI computing, high-performance computing (HPC), generative AI, simulation, and large-scale data center workloads. Built on the latest Blackwell architecture, this server GPU delivers exceptional compute performance for modern AI infrastructure and enterprise deployments.

96GB GDDR7 Memory with ECC for Large AI Models

Equipped with 96GB GDDR7 memory and ECC protection, the RTX PRO 6000 Blackwell Server Edition supports demanding AI workloads including large language models (LLMs), generative AI inference, model training, digital twins, simulation, and scientific computing. The large memory capacity enables handling of complex datasets and larger AI models while improving reliability and data integrity.

Optimized for AI, HPC and Enterprise Applications

The RTX PRO 6000 Blackwell Server Edition is designed for data centers, AI labs, research institutions, and enterprises requiring accelerated computing resources. It supports AI development workflows, rendering, visualization, engineering simulations, digital content creation, and accelerated analytics while enabling scalable deployment within server environments.

Built for Modern Data Center and AI Infrastructure

Designed for enterprise deployment, the server edition integrates into modern AI clusters and accelerated computing environments. It enables organizations to build scalable AI platforms supporting model development, inference, virtualization, and next-generation compute workloads while maintaining high reliability and enterprise readiness.

Technical Specifications

General
Model: 900-2G153-0000-000
Product Name: NVIDIA RTX PRO™ 6000 Blackwell Server Edition 96 GB GDDR7 with ECC
Product Type: Professional Server GPU / AI Accelerator
GPU Architecture: NVIDIA Blackwell
Deployment: AI Factories, Enterprise AI, HPC, LLM Inference, Data Centers, Visualization Workloads
Form Factor: Full Height Full Length (FHFL), Dual Slot
Cooling Design: Passive Cooling (Server Optimized)
Interface: PCI Express 5.0 x16

GPU Architecture & Processing
GPU Architecture: NVIDIA Blackwell
CUDA Cores: 24,064
Tensor Cores: 752 (5th Generation)
RT Cores: 188 (4th Generation)
AI Performance Features: FP4, FP8, FP16, BF16, Tensor Operations Support
Neural Rendering Support: Supported
DLSS 4 Multi Frame Generation: Supported
RTX Mega Geometry: Supported

Memory Specifications
GPU Memory: 96 GB GDDR7 with ECC
Memory Interface: 512-bit
Memory Bandwidth: Up to 1,597 GB/s (~1.6 TB/s)
ECC Memory Protection: Supported
High Capacity AI Memory Pool: Supported

AI & Compute Features
AI Inference Acceleration: Supported
Generative AI Workloads: Supported
LLM Fine Tuning: Supported
RAG Workloads: Supported
Multimodal AI: Supported
Physical AI: Supported
Scientific Computing: Supported
Multi-Instance GPU (MIG): Up to 4 MIG partitions @ 24 GB each

Rendering & Visualization
Real-Time Ray Tracing: Supported
Photorealistic Rendering: Supported
3D Modeling & CAD: Supported
Media & Entertainment Workloads: Supported
Architecture / Engineering Visualization: Supported
VR Environment Processing: Supported

Video Engine Features
NVENC: 9th Generation
NVDEC: 6th Generation
AV1 Encoding: Supported
HEVC Encoding: Supported
4:2:2 H.264 Encoding: Supported
4:2:2 HEVC Decode: Supported

Performance Specifications
Single Precision Performance (FP32): Up to 120 TFLOPS
RT Performance: Up to 355 TFLOPS
AI Processing Optimization: FP4 Precision Support
Transformer Engine Support: Supported

Power & Thermal
Maximum Power Consumption: Up to 600W (Configurable)
Power Range: 400W – 600W
Thermal Solution: Passive Cooling
Server Rack Deployment: Optimized
Air-Cooled Variant: Available
Liquid-Cooled Variant: Available

Security Features
Confidential Compute: Supported
Secure Boot: Supported
Root of Trust: Supported
ECC Protection: Supported
Enterprise Security Integration: Supported

Connectivity & Interface
Bus Interface: PCIe Gen 5 x16
Bidirectional Bandwidth: Up to 128 GB/s
Server Integration: Supported
Multi-GPU Deployments: Supported

Physical Specifications
Form Factor: FHFL Dual Slot
Card Height: 4.4 in
Card Length: 10.5 in
Mounting Type: Server Rack Optimized
Cooling Style: Passive Server Cooling

Supported Workloads
LLM Fine Tuning (<70B Parameters)
LLM Inference
RAG Deployments
Enterprise AI Factories
Generative AI
Scientific Computing
Data Analytics
3D Rendering
Simulation Workloads
Video Processing

Package Contents
NVIDIA RTX PRO 6000 Blackwell Server Edition GPU
Server Integration Documentation
Warranty Documentation

Key Features
NVIDIA Blackwell Architecture
24,064 CUDA Cores
96 GB GDDR7 ECC Memory
1.6 TB/s Memory Bandwidth
PCIe 5.0 x16
752 Tensor Cores (5th Gen)
188 RT Cores (4th Gen)
Up to 600W Configurable Power
Supports up to 4 MIG partitions
Enterprise AI & HPC Ready

Typical Use Cases
Enterprise AI Infrastructure
DGX / AI Factory Environments
LLM Training & Inference
RAG Systems
Scientific Simulation
Media Rendering Farms
Data Analytics Platforms
High Performance Computing Clusters