Understanding the Query
The ASUS ExpertCenter Pro ET900N G3 is a deskside AI workstation built on the NVIDIA GB300 Grace Blackwell Superchip. According to ASUS and its published specifications, it pairs a 72-core Grace CPU with 496 GB of LPDDR5X system memory and a Blackwell GPU with 252 GB of HBM3E, delivering roughly 20 PFLOPS of local AI compute, with up to 748 GB of coherent unified memory across the platform. It is a high-end, high-power machine aimed at AI researchers and developers running very large models locally.
Where Thinvent Fits
Thinvent operates in a very different part of the computing landscape. Rather than deskside AI supercomputers, Thinvent builds compact, energy-efficient, fanless and industrial-grade computers — mini PCs, thin clients, all-in-one PCs and industrial PCs — designed to run reliably for years in offices, factories, retail counters, clinics and classrooms. These systems are typically powered by efficient Intel processors, draw a fraction of the power of a workstation-class machine, and are built to survive dust, heat and continuous operation.
| Aspect | Deskside AI workstation (e.g. ET900N G3 class) | Thinvent compact / industrial computers |
|---|---|---|
| Typical role | Local AI model training and inference | Endpoint computing, VDI, kiosks, industrial control |
| Power profile | Very high (dedicated PSU, active cooling) | Low power, often fanless |
| Footprint | Large deskside chassis | Small, mountable, space-saving |
| Compute focus | GPU-accelerated AI | Efficient general-purpose CPU + integrated graphics |
| Deployment | Research lab, AI development | Office, shop floor, branch, classroom |
Choosing the Right Class of Machine
If your workload is large-language-model training, fine-tuning or heavy GPU inference, you need a workstation-class platform with discrete acceleration. If your workload is running applications, browsing, billing, ERP, POS, digital signage, thin-client sessions, data-entry terminals or machine control, a compact fanless computer is the more sensible, quieter, cooler and lower-maintenance choice. Many organisations run both: a small number of powerful AI or server machines in the back office, and a large fleet of efficient Thinvent-class endpoints at the edge.
Key Considerations for Compact Computers
When specifying a compact or industrial computer, the parameters that matter most are processor efficiency and core count, memory capacity, storage type and size, fanless versus actively cooled design, operating temperature range, mounting options, number and type of display outputs, wired networking, and long-term OS and supply stability. Thinvent systems are commonly specified with Intel Celeron and N-series processors, 4–16 GB memory, SSD storage, dual display outputs, gigabit Ethernet, and a choice of Windows, Linux or no OS.
Thinvent's Products in This Space
Thinvent's range covers fanless mini PCs, thin clients, all-in-one PCs and rugged industrial computers built for round-the-clock duty. Typical configurations use efficient Intel processors with 4 to 10+ cores, 4–32 GB of memory, SSD storage from 128 GB upward, gigabit Ethernet, HDMI and USB connectivity, and a choice of Microsoft Windows 11 Pro, Ubuntu Linux, an embedded Linux build, or no operating system. For organisations that need dependable, low-power endpoints rather than GPU-heavy AI workstations, Thinvent offers a practical, serviceable alternative with worldwide availability and long product lifecycles.
