ASUS ExpertCenter Pro ET900N G3

ASUS ExpertCenter Pro ET900N G3

The ASUS ExpertCenter Pro ET900N G3 is a next-generation deskside AI supercomputer, bringing data center–class performance into a workstation form factor. Designed for AI engineers, researchers, and enterprise teams, it enables local model training, fine-tuning, and inference without relying on cloud infrastructure.

Key Highlights

• Powered by NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip (CPU + GPU unified architecture)
• Up to ~775–784GB unified coherent memory for large-scale AI workloads
• Up to 20 PFLOPS AI performance for training and inference
• High-speed NVIDIA ConnectX-8 networking (up to ~800Gb/s)
• Runs NVIDIA DGX OS with full AI software stack support

Designed for Modern AI Workflows

• Train, fine-tune, and run large language models locally
• Eliminates cloud latency and data privacy concerns
• Supports multi-user environments with GPU partitioning (MIG)
• Seamless transition from local development to data center or cloud deployments

Unified AI Architecture

• Grace CPU and Blackwell GPU connected via NVLink-C2C for high-bandwidth, low-latency communication
• Large unified memory pool removes traditional CPU–GPU bottlenecks
• Fifth-generation Tensor Cores with FP4 precision for next-gen AI models

Enterprise-Ready Design

• Deskside/tower form factor for office or lab environments
• Scalable with PCIe expansion and NVMe storage options
• Optimized cooling and power delivery for continuous AI workloads
• Ideal for organizations requiring on-prem AI with full data control

Why It Stands Out

The ET900N G3 represents a new category of AI infrastructure — bridging the gap between traditional workstations and full-scale data center clusters. With up to 20 PFLOPS of AI performance and a massive unified memory pool, it allows developers to run significantly larger models locally while maintaining enterprise-grade performance and scalability.