Product features
CPU & GPU
- Processor: 1x NVIDIA Grace ™ 72-core Neoverse V2 CPU (ARM-based)
- GPU: 1x NVIDIA Blackwell Ultra GPU
- MIG: Multi-instance GPU for up to 7 MIGs
- optional: 1x dedicated NVIDIA Blackwell GPU for graphics rendering, e.g. RTX Pro 2000 Blackwell
RAM
- total coherent memory: 748GB, linked via NVLink ®-C2C, at 900 GB/s interconnect bandwidth
- CPU memory: 496GB ECC LPDDR5X, 396 GB/s memory bandwidth
- GPU memory: 252GB ECC HBM3e, 7.1 TB/s memory bandwidth
Networking
- Dual-port QSFP112 400GbE LAN via NVIDIA ConnectX-8 SuperNIC (= 800 Gbit/s bandwidth total)
- 1x RJ45 1GbE Dedicated BMC LAN port
- 1x RJ45 10GbE LAN port
- WIFI optional
Storage
- 4x M.2 PCIe 5.0 x4 NVMe (M-key 2280)
- populated with 4x 2TB SSDs - configurable
PCIe Slots
- 1x PCIe Gen 5 x16 FHFL double-width, for GPUs
- 2x PCIe Gen 5 x8 (in x16) FHHL
- 1x M.2 PCIe 2.0 x1 slot (E-Key 2230))
Front IO Ports and Connectors
- 2x USB 3.2 Gen2 port (Type-C)
- 2x USB 3.2 Gen2 ports (Type-A)
- 1x USB 2.0 port
- 1x headphone port
- 1x microphone port
- 1x power button
- 2x message LEDs
Rear IO Ports and Connectors
- 4x USB 3.2 Gen2 ports (Type-A, 10 Gbit/s)
- 1x microUSB port (for BMC)
- 2x QSFP ports 400 Gbit/s (CX-8 NIC)
- 1x RJ45 port 10 Gbit/s Ethernet (Marvel)
- 1x RJ45 port 1 Gbit/s Ethernet (Realtek)
- 1x mini-DP (video port, for BMC)
- 3x Audio jacks (Audio in/Audio out/Mic)
OS & security
- OS: Ubuntu 24.04 LTS with NVIDIA AI Developer Tools
- pre-installed with essential CUDA-X™ libraries, including tools for data science
- pre-installed with complete NVIDIA AI Software Stack
- out-of-the-box support for NVIDIA NemoClaw open source stack allowing organizations to build and deploy always-on autonomous AI assistants locally
- supports NVIDIA Nemotron
IT manageability and enterprise features
- out-of-band telemetry via BMC
- Redfish API support
- NVIDIA Data Center GPU Manager
- visibility, monitoring, remote management capabilities
- hardware root of trust
- enterprise secure boot
- onsite hardware support (next business day) included
Dimensions, weight, form factor
- Form factor: desktop tower with carrying handle
- height: 565 mm / 22.3 inch
- width: 232 mm / 9.1 inch
- depth: 584 mm / 23 inch
- net weight: 27 kg
- gross weight: 32 kg
System cooling
- cooling system: closed-loop AIO liquid cooling (suitable for office use)
- data-center grade thermal architecture
- long-duration, full-load local AI computing with zero thermal throttling, consistent peak performance, 24/7 reliability for mission-critical workloads
Power supply
- power supply: 1x 1600W Titanium ATX power supply, 100 - 240V
- compatible with C19 power cord
Misc
- Audio: Realtek® ALC4080 Audio Codec
- operating temperature: TBD
- storage temperature: TBD
- operating relative humidity: TBD (non-condensing)
- storage relative humidity: TBD (non-condensing)
- certification: CE, FCC (Class A), BSMI, UL, KCC
packing content:
- 1x ASUS ExpertCenter Pro ET900N G3 AI Supercomputer (NVIDIA DGX Station GB300)
- 1x power cord: IEC C19 to EU standard plug (US NEMA 5-20P available on request)
AI performance
- FP4 Tensor core: 20 PFLOPS / 15 PFLOPS without sparsity
- FP8 / FP6 Tensor core: 10 PFLOPS
- Int8 Tensor Core: 330 TOPS
- FP16/BF16 Tensor Core: 5 PFLOPS
- TF32 Tensor Core: 2.5 PFLOPS
- FP32: 80 TFLOPS
- FP64 / FP64 Tensor Core: 1.3 TFLOPS
Product description
The ASUS ExpertCenter Pro ET900N G3 AI Supercomputer is an NVIDIA DGX Station based on the GB300 Grace™ Blackwell superchip. It can be used for running local AI applications for development and production purposes.
Example AI applications include LLM inference for frontier models of up to 1 Trillion parameters, scientific workloads (e.g. BioNeMo), finetuning models, hosting AI agents, and running RAGs for AI-enhanced document workflows. In particular it is well-suited to run MoE (mixture of experts) models.
Thanks to its MIG feature and its massive memory throughput it can be run headless to support several colleagues in a local workgroup.
Compared to server systems, or heavy multi-GPU systems, it runs cooler and quieter, and is uniquely suited for office use. Its server-grade management features, such as the integrated BMC, the Redfish API support, and enterprise secure boot allow to integrate it into a corporate IT management landscape.
Two ExpertCenter Pro ET900N G3 DGX Stations can be joined using the Dual-port QSFP112 400GbE LAN to serve even bigger models, increase precision, or general model performance.
As a special feature the ASUS ExpertCenter Pro ET900N G3 always includes onsite hardware support in its purchase price.
ASUS also claims to plan future support for Windows-based AI development.
A separately obtainable NVIDIA AI Enterprise software license provides a suite of software tools, libraries, and frameworks to fast-track your AI deployment on the ASUS DGX Station GB300.
Four strategic advantages over cloud data centers:
- Enhanced Data Privacy & Governance - sensitive corporate data, proprietary code, and intellectual property never leave the local corporate network
- Zero cloud latency - real-time AI inferencing and complex simulations operate at local bus speeds without regional network lag
- Predictable Operational Costs: eliminates ongoing variable cloud consumption, egress fees and API token billing in favor of a fixed hardware asset and electricity consumption
- Uncapped Experimentation: AI research labs and developers can run continuous, heavy fine-tuning and token-heavy autonomous agent testing 24/7 without risking budget overruns
Optional GPUs
One of the following optional GPUs can be added into the system to add graphics output options, and for scientific simulation and rendering workloads:
- NVIDIA RTX PRO 6000 Blackwell Workstation Edition
- NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
- NVIDIA RTX PRO 4000 Blackwell SFF Edition
- NVIDIA RTX PRO 2000 Blackwell
Datasheets, manuals and additional downloads
Lead time & purchase
This product is available on pre-order with a current lead time of about 8-9 weeks. Contact us to purchase it.
Additional SKUs and codes
- model number: ExpertCenter Pro ET900N G3