
TITAN GS685R-G6 B200
AI & GPU Server
- 8U
- Dual AMD EPYC™ 9005/9004 and 97x4 Series Processors
- NVIDIA HGX™ B200 8-GPU with NVIDIA NVSwitch™
- Up to 6TB DDR5-6400 MT/s memory
- 12 x 2.5" NVME hot-swap drive bays and 2 x M.2
- ASPEED AST2600 management
- 3000W (6+6) 80 PLUS Titanium Power Supply
- Perfect for AI Training, AI Inference, Deep Learning and more

Product Overview
The CIARA TITAN GS685R-G6 B200 combines dual AMD EPYC™ 9005/9004 and 97×4 Series processors (up to 500 W TDP each) with the NVIDIA HGX™ B200 8-GPU platform to deliver high throughput for AI and HPC workloads.
Featuring NVSwitch™ interconnect technology, up to 6 TB of DDR5 6400 MT/s memory, and 12 hot-swap NVMe bays, it minimizes latency and ensures consistent performance across demanding workloads, making it a reliable choice for large-model training and enterprise-scale AI infrastructure.


Key Features and Benefits
High-performance Compute Power
Dual-socket AMD EPYC™ 9005/9004 and 97×4 Series processors (up to 500 W TDP each) provide multi-core performance for AI, HPC, and scientific workloads, enabling faster model training and simulation at scale.
NVIDIA HGX™ B200 Platform with NVSwitch™
The NVIDIA HGX™ B200 platform with 8 GPUs connected via NVSwitch™ provides full-bandwidth GPU-to-GPU communication, minimizing latency and accelerating large-model training, inference, and parallel computation.
Optimized Cooling and Power Efficiency
Advanced airflow and efficient power delivery ensure consistent GPU performance under sustained load while reducing energy use and maintaining reliability across demanding workloads.
Flexible Storage and Expansion
Twelve hot-swap NVMe bays and two M.2 slots enable fast data ingestion, checkpointing, and model loading, while 9 PCIe Gen5 x16 slots support future GPUs, accelerators, or network adapters.
Perfect For
AI Training
AI Inference
Deep Learning
Machine Learning
Simulation
Research
Tech Specs
Resources
Product Downloads
Service & Support
Related Products
















