Technology news and discussion on TalkFuse.
ASUS targets businesses with a full-stack AI infrastructure package
ASUS has announced a full-stack AI package aimed at businesses building and running AI systems in-house. The move reflects growing enterprise interest in managing AI workloads directly, driven by data-control needs, predictable costs and the spread of agent-based applications.
The package, announced at NVIDIA’s GTC event in Berlin on 21–22 October, combines AI servers and infrastructure, software for managing AI workflows, and services through ASUS AI HUB. ASUS argues that agentic AI relies more heavily on CPUs as well as GPUs, because agents repeatedly plan tasks, run code, use tools and review results. The article cites IDC figures showing the global server market reached $122.6 billion in the first quarter of 2025, up 30 per cent year on year, and forecasts that inference will account for more than 70 per cent of computing demand by next year.
In short
- ASUS is targeting enterprise demand for in-house AI infrastructure.
- Its offer combines servers, workflow software and support services.
- ASUS says agentic AI increases the importance of CPU performance.
Explained simply
Artificial intelligence systems need powerful computers to train models and respond to users. Many businesses are considering running these systems themselves, rather than relying entirely on outside cloud services, to keep closer control of their data and costs.
ASUS is a technology company that makes computers and servers, while NVIDIA supplies widely used chips for AI computing. AI systems can use graphics processing units (GPUs) for heavy calculations and central processing units (CPUs) for general tasks; newer AI agents can require both as they plan work, use software tools and check results.
The wider shift matters because businesses need enough computing capacity to run AI reliably as use grows. The equipment, software and support they choose can affect how much control they have over their systems and what it costs to operate them.
Read the full article at the source →
Originally published by The Register as “Building AI factories at enterprise scale”.