I just saw something amazing
New compact workstation supports up to dual Nvidia RTX 6000 Ada GPUs for running large language models locally.
ASUS has unveiled the ExpertCenter Pro ET900N G3, a compact workstation specifically engineered for AI development and local large language model deployment. The system's standout feature is its support for up to two Nvidia RTX 6000 Ada Generation GPUs, each featuring 48GB of GDDR6 memory—critical for running parameter-heavy models without constant cloud offloading.
Technically, the workstation measures just 10 liters but packs substantial power with 13th/14th Gen Intel Core i9 processors, support for 128GB DDR5 RAM, and multiple storage configurations including dual M.2 slots. The dual-GPU configuration provides 96GB of total VRAM, enabling local execution of models like Llama 3 70B or GPT-4 class models through quantization techniques. This addresses the growing demand for on-premises AI infrastructure amid rising cloud costs and data privacy concerns.
The ET900N G3 represents a strategic shift toward democratizing high-performance AI compute. While cloud platforms like AWS and Azure dominate enterprise AI, this hardware enables research labs, financial institutions, and healthcare organizations to maintain full data control while developing proprietary models. The system's compact form factor makes it suitable for office environments where traditional server racks aren't feasible, potentially accelerating AI adoption in regulated industries where data sovereignty is paramount.
- Supports dual Nvidia RTX 6000 Ada GPUs with 48GB VRAM each (96GB total)
- Compact 10-liter chassis with Intel Core i9 processors and 128GB DDR5 RAM
- Enables local execution of 70B+ parameter LLMs without cloud dependency
Why It Matters
Reduces AI inference costs by 60-80% versus cloud services while ensuring complete data privacy for sensitive applications.