Dear customers, please be informed that PC Themes is closed on Mondays. We will be open on every Tuesday to Sunday, as per our usual business hours. For PH, kindly check with us prior to coming to our store. Thank you. 
Shop
Nvidia Graphics Card
PNY NVIDIA RTX PRO™ 6000 Blackwell Server Edition Graphics Card
- $26,999.00
Pre-order: Requires 1 to 2 week for fulfillment
| PNY NVIDIA RTX PRO 6000 Blackwell Server Edition 96GB Graphics Card Specifications | |
|---|---|
| Architecture | NVIDIA Blackwell |
| CUDA Cores | 24,064 |
| Tensor Cores | 752 (5th Generation) |
| RT Cores | 188 (4th Generation) |
| GPU Memory | 96GB GDDR7 with ECC |
| Memory Interface | 512-bit |
| Memory Bandwidth | Up to 1.6 TB/s |
| Power Consumption | 400 W – 600 W (Configurable) |
| Cooling Solution | Passive |
| Form Factor | 4.4" H x 10.5" L, FHFL Dual Slot |
| Weight | 1.219 kg (1219g) |
| Graphics Bus | PCIe 5.0 x16 |
| Interconnect Interface | PCIe Gen 5 x16, 128 GB/s Bidirectional |
| Display Connectors | 4x DisplayPort 2.1 |
| Power Connector | 1x PCIe CEM5 16-pin |
| NVENC Engines | 4x (9th Generation) |
| NVDEC Engines | 4x (6th Generation) |
| JPEG Support | 4x |
| Multi-Instance GPU | Up to 4 MIG @ 24GB |
| Confidential Compute | Supported |
| Secure Boot | Yes |
Share Now
Out of Stock!
This product is out of stock, submit your email and we will update once it restock.
PNY NVIDIA
RTX PRO™ 6000 Blackwell Server Edition Graphics Card
Unique Features
- » NVIDIA Blackwell Architecture: Purpose-built for the data center, leveraging the Blackwell architecture to deliver unmatched AI training, inference, and rendering capabilities at scale.
- » 96GB GDDR7 VRAM: Outfitted with an expansive 96GB of ultra-fast GDDR7 memory with Error-Correcting Code (ECC), delivering a monumental 1,792 GB/s of bandwidth for massive LLMs and data science workloads.
- » Passive Thermal Design: Expertly engineered with a high-performance passive heatsink, relying on system-level airflow to maximize density and reliability in 24/7 enterprise server environments.
- » Enterprise Virtualization: Fully optimized for NVIDIA vGPU software and hardware-level Multi-Instance GPU (MIG) partitioning, allowing seamless resource allocation across multiple virtual machines.
Core Specifications
GPU Engine
- • Architecture: NVIDIA Blackwell
- • CUDA Cores: 24,064
- • Hardware Acceleration: 752 Tensor Cores & 188 RT Cores
- • Media Engines: 4x NVENC & 4x NVDEC with advanced AV1 Encode/Decode
Memory Subsystem
- • Memory Size: 96 GB GDDR7
- • Memory Interface: 512-bit
- • ECC Support: Yes (Error-Correcting Code, Enterprise Grade)
- • Bandwidth: 1,792 GB/s
Virtualization & Software
Display Outputs
- • Connectors: None (Headless Server Design)
- • Target Use Case: Remote rendering, cloud computing, and AI data centers
Software Ecosystem
- • MIG Capability: Multi-Instance GPU partitioning (up to 4x 24GB or 2x 48GB instances)
- • Supported Software: NVIDIA vGPU, NVIDIA AI Enterprise, CUDA 12.8, Direct X 12 Ultimate
Connectivity & Power
System Interface
- • Bus Type: PCI Express 5.0 x16
- • Form Factor: 4.4” H x 10.5” L, FHFL Dual Slot
Electrical Specs
- • Power Consumption: 300W (Max)
- • Power Connectors: CPU 8-pin (Standard server power delivery)
- • Thermal Requirement: System airflow is strictly required for cooling
Build & Warranty
Physical Details
- • Thermal Solution: Passive Heatsink
- • Server Compatibility: Certified for major OEM rackmount and tower server chassis
- • Scalability: Designed for high-density multi-GPU installations
Support
- • Certifications: NVIDIA Certified Systems compatible, NEBS Ready
- • Warranty: 3-Year PNY Professional Enterprise Warranty