Asus Released ExpertCenter Pro ET900N G3 Workstation
The new workstation integrates the Nvidia Grace Blackwell Ultra Superchip for high-performance computing workloads.
Updated on Sept. 27, 2026 in Semiconductors
Live Poll
Would you pay over $100,000 for a desktop computer capable of running AI models at home?
Asus has released the ExpertCenter Pro ET900N G3, a powerful workstation featuring the Nvidia GB300 Grace Blackwell Ultra Superchip. The system is designed for demanding environments and utilizes advanced liquid cooling.
Why it matters
The release of this workstation highlights the increasing integration of specialized artificial intelligence chips into high-end hardware for professional computing. It provides a platform for developers to leverage massive coherent memory for complex data processing tasks.
The system features 72 ARM cores, 748GB of coherent memory, and two 400-gigabit networking ports. It is powered by a 1,600-watt supply with 80 PLUS Titanium efficiency, with thermal management handled by a sealed liquid cooling loop and copper plates.
The players
Asus
A multinational computer and phone hardware and electronics company that develops high-performance computing equipment.
Nvidia
A global technology company that designs graphics processing units and high-performance supercomputing chips for artificial intelligence applications.
The details
Measuring 584 x 232 x 565 mm and weighing 27 kilograms, the workstation includes two 2TB NVMe drives in a mirrored array. An internal fan directs airflow specifically onto the network card's optical transceivers to maintain performance.
Timeline
September 27, 2026: Official publication of the workstation release announcement.
The Tech Race
The transition to Blackwell-based workstations signals a shift from purely cloud-based AI training to edge-based development for heavy workloads. This development positions Asus against other hardware vendors currently scaling the Nvidia architecture for enterprise deployment.
Professional developers and data scientists can expect enhanced localized performance, potentially bypassing the latency issues associated with remote server clusters. However, potential buyers should anticipate a significant investment, with industry estimates placing the cost between $100,000 and $120,000.
The takeaway
This workstation represents a milestone in bringing large-scale processing power directly to an individual or small-team workflow environment. Professionals looking to adopt this technology should factor in the high power requirements and significant physical space needed for the unit.
Further reading
For more information on the current state of specialized computing hardware, visit the Semiconductors section.
Source note: This article includes information reported by TechRadar.
Live Poll
Would you pay over $100,000 for a desktop computer capable of running AI models at home?
