A workstation for running AI models on your own premises
The InfoTelica AI Tower is a GPU workstation built on NVIDIA RTX PRO 6000 Blackwell (Max-Q) graphics cards and AMD Ryzen Threadripper PRO processors. It runs large language models, vision models and AI agents locally, so your data never leaves your office.
We deliver it ready to use, with the operating system, drivers and inference software installed, and we can build your first AI application on it.
1x NVIDIA RTX PRO 6000 Max‑Q · Threadripper PRO 9975WX · 256 GB DDR5 ECC · 2 TB NVMe
Three starting points
Example configurations. Every component can be changed to fit your workload and budget.
Starter
For a team starting with AI
- GPU: 1x NVIDIA RTX PRO 6000 Max‑Q
- CPU: AMD Ryzen Threadripper PRO 9975WX
- Memory: 256 GB DDR5 ECC
- Storage: 2 TB NVMe SSD
- Good for: Document Q&A, summarisation and coding assistants with mid-size models
Professional
For a department in production
- GPU: 2x NVIDIA RTX PRO 6000 Max‑Q
- CPU: AMD Ryzen Threadripper PRO 9985WX
- Memory: 512 GB DDR5 ECC
- Storage: 2x 4 TB NVMe SSD
- Good for: Several applications at once, larger models and more concurrent users
Maximum
For large models and fine-tuning
- GPU: 4x NVIDIA RTX PRO 6000 Max‑Q
- CPU: AMD Ryzen Threadripper PRO 9995WX
- Memory: 1024 GB DDR5 ECC
- Storage: 4x 4 TB NVMe SSD + HDD
- Good for: Large and multimodal models, fine-tuning and many users
Component options
| Component | Options |
|---|---|
| Graphics | 1x, 2x, 3x or 4x NVIDIA RTX PRO 6000 Max-Q |
| Processor | AMD Ryzen Threadripper PRO 9975WX, 9985WX or 9995WX |
| Memory | 256 GB (4x64), 384 GB (4x96), 512 GB (8x64), 768 GB (8x96) or 1024 GB (8x128) DDR5 ECC registered |
| SSD storage | 2 TB, 2x 2 TB, 4x 2 TB, 4 TB, 2x 4 TB or 4x 4 TB NVMe |
| HDD storage | 4, 8, 12, 16 or 20 TB |
| Platform | ASUS Pro WS WRX90E-SAGE SE motherboard, FSP 2500W Cannon Pro ATX 3.1 power supply, Arctic Freezer 4U-M cooling |
What comes with it
Ready to run
Linux, NVIDIA drivers, container runtime and an inference server such as vLLM are installed and tested before delivery.
Your first application
We can deliver a working document assistant, translation tool or analytics agent on the tower, built on your own data.
Support
Remote support for the system and software, and help to scale up to a shared GPU platform when demand grows.
Not sure which configuration you need?
Tell us which models and how many users you plan to serve. We will recommend a configuration.