As an Amazon Associate, we earn from qualifying purchases. TweakTown may also earn commissions from other affiliate partners at no extra cost to you.
The ASUS ExpertCenter Pro ET900N G3 desktop PC might look like a sleek computer that you'd find in an office, but it's a desk-sized supercomputer powered by the latest NVIDIA Grace Blackwell superchips. With GPU memory being the current limiting factor in running complex local AI models, this is the sort of rig where even the GeForce RTX 5090 32GB GPU isn't enough.

The ASUS ExpertCenter Pro ET900N G3 desktop PC is an AI supercomputer.
The ASUS ExpertCenter Pro ET900N G3 desktop PC is based on the NVIDIA DGX station and has been designed from the ground up to accelerate AI workflows. It's powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, which includes an NVIDIA Blackwell GPU and an NVIDIA Grace CPU connected via NVIDIA NVLink-C2C interconnect.
You're looking at data center-level performance in the size of a desktop PC with access to up to 784GB of memory. The system is optimized for NVIDIA's AI Software Stack and comes with the custom Linux-based NVIDIA DGX OS to maximize its performance, which is a whopping 20 PFLOPS of AI performance.

The palm-sized ASUS Ascent GX10 delivers 1,000 AI TOPS of power.
If a desktop-sized cutting-edge AI data center is too big, there's also the compact ASUS Ascent GX10 - a palm-sized supercomputer powered by the NVIDIA GB10 Grace Blackwell Superchip. With up to 1,000 AI TOPS of power and 128GB of system memory, ASUS notes that this mini-PC or mini-supercomputer can handle 200 billion parameter AI models.
If you want to work with even larger models, two ASUS Ascent GX10 systems can be connected using cutting-edge NVIDIA ConnectX-7 Network Technology to double their performance to support 400 billion parameter AI models.
ASUS notes that aspects of these systems are still in development, and more information on pricing availability will be available later this year.