TensorFlow Workstations
Recommended Workstations
-
CPU
Intel Xeon W-2500/W-3500 (60 Cores)
-
GPU
Up to 2x NVIDIA RTX 4090, 4080 or 4x RTX 6000 Ada
-
Memory
Up to 2TB DDR5 Memory
-
Cooling
GPU: Air-cooling | CPU: Water-cooling
-
Features
NVIDIA AI Workstation for AI Research
-
CPU
AMD Threadripper PRO 5000WX/7000WX (96 Cores)
-
GPU
Up to 2x NVIDIA RTX 4090/4080 or 4x RTX 6000 Ada
-
Memory
Up to 1 TB DDR5 Memory
-
Cooling
GPU: Air-cooling | CPU: Water-cooling
-
Features
Optimized for stable diffusion, LLaMA, Alpaca. ChatGPT-like AI On Your Local Computer
-
CPU
AMD Threadripper Pro 5000WX/7000WX (96 Cores)
-
GPU
Up to 7x Water-cooled NVIDIA RTX 4090, 6000 Ada, A100, H100
-
Memory
Up to 1 TB DDR5 Memory
-
Cooling
Enterprise-class water-cooling
-
Features
Up to 3x times lower noise vs. air-cooling. Maximum GPU power for inference and training
- Liquid‑cooled GPUs + CPU
- Simple maintenance
- Quick disconnect fittings for easy upgrades
- Smart controller
- Up to 3x times lower noise vs. air-cooling
-
CPU
2x Intel Xeon 4th/5th Gen (128 Cores)
-
GPU
Up to 5x NVIDIA RTX 4080, 4090, A6000, A100, H100
-
Memory
Up to 4 TB
-
Cooling
Professional-grade water-cooling
- #1 world's fastest ranked server (luxmark benchmark)
- Designed for 24/7/365 operation at 100% load
- Liquid-cooled 8x GPUs Deep Learning, AI or Rendering Rackmount Server
- Up to 3x times lower noise vs. air-cooling
- Enterprise-class custom liquid-cooling system
- Simple maintenance
- Quick disconnect fittings for easy upgrades
- Modular design
- Smart controller
- You can place the server in a lab or office
-
CPU
2x Intel Xeon 4th/5th Gen (128 cores)
-
GPU
Up to 8x water-cooled NVIDIA RTX A6000, A100, H100 GPUs
-
Memory
Up to 6 TB Memory
-
Cooling
Server-grade water-cooling
Enterprise-class custom liquid-cooling system
- Compact liquid‑cooled AI workstation with 2x NVIDIA A100, H100 94GB
- Simple maintenance
- Quick disconnect fittings for easy upgrades
- Smart controller
- Up to 3x times lower noise vs. air-cooling
- Optimized for AI, deep learning, data science, Llama
- Multi-GPU System for AI Research
- Optimized for stable diffusion, LLaMA, Alpaca. “ChatGPT” On Your Local Computer.
-
CPU
AMD Threadripper Pro (96 Cores)
-
GPU
Up to 2x Water-cooled NVIDIA RTX A100, H100
-
Memory
Up to 1 TB DDR5 Memory
-
Cooling
Enterprise-class water-cooling
-
Features
Up to 3x times lower noise vs. air-cooling. Maximum GPU power for inference and training