ASUS has officially revealed the ExpertCenter Pro ET900N G3, which brings the power of NVIDIA DGX GB300 into offices, thanks to the desk-side friendly form factor.
It is made to support modern AI workloads, including LLM fine-tuning, generative AI pipelines, physical AI simulations, deep learning research, and autonomous agent-based systems. It leverages NVLink-C2C high-bandwidth interconnect technology to enable the massive 748GB coherent memory access across the platform, improving throughput and reducing bottlenecks.
As a result, the system can deliver up to 20 PFLOPS of AI performance while working seamlessly with NVIDIA AI software stack featuring machine learning experimentation, analytics, agentic AI development and deployment, such as NVIDIA Nemo and Claw workflows – all sitting securely within a locally managed environment. ASUS also provided some numbers here – vLLM through Qwen models enabled 864 TPS output and up to 1600 TPS for the whole process.
The ExpertCenter Pro ET900N G3 is now available for global order, but do check with your local offices to see if they are bringing it to your specific market.
