ASUS and NVIDIA unveil desktop AI supercomputers with 784GB memory

ASUS and NVIDIA have unveiled a new class of AI desktop supercomputers featuring the NVIDIA Grace Blackwell superchips, including the ASUS ExpertCenter Pro ET900N G3 desktop and the ultra-compact ASUS Ascent GX10 mini-PC. These systems address the “last mile” challenge in AI computing by delivering data center-level performance in desktop form factors, with the ExpertCenter Pro offering up to 784GB of GPU memory and the palm-sized Ascent GX10 providing up to 1,000 AI TOPS of processing power.

The big picture: Traditional consumer graphics cards like the NVIDIA GeForce RTX 5090 with 32GB of VRAM can handle smaller AI models, but researchers and developers need significantly more memory to run large-scale AI workloads locally without relying on cloud-based data centers.

Key specifications: The ExpertCenter Pro ET900N G3 represents a breakthrough in desktop AI computing power.

• Powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, featuring an NVIDIA Blackwell Ultra GPU and NVIDIA Grace CPU connected via NVIDIA NVLink-C2C interconnect
• Offers up to 784GB of coherent system memory—more than twice the GPU memory of a workstation equipped with four NVIDIA RTX PRO 6000 GPUs
• Runs on NVIDIA DGX OS, a customized Ubuntu Linux installation optimized for AI, machine learning, and analytics applications

Ultra-portable powerhouse: The ASUS Ascent GX10 miniaturizes supercomputer capabilities into a palm-sized device.

• Built around the NVIDIA GB10 Grace Blackwell Superchip with a 20-core Arm Grace CPU and Blackwell GPU
• Delivers up to 1,000 AI TOPS (trillion operations per second) of processing power with 128GB of unified system memory
• Features integrated NVIDIA ConnectX-7 Network Technology for connecting two units to handle larger models like Llama 3.1 with 405 billion parameters

Why this matters: These systems bridge the gap between consumer-grade AI hardware and enterprise data center solutions, enabling researchers and businesses to run sophisticated AI workloads locally while maintaining privacy and control over their data.

Market positioning: ASUS positions these products as part of a comprehensive AI ecosystem spanning from consumer AI PCs to enterprise data center solutions, with the company emphasizing its unique breadth in the AI hardware market across all performance tiers.