TAIPEI (Taiwan News) — Taiwan’s electronics manufacturers are expanding beyond their traditional roles as AI reshapes demand across the computing stack, pushing the industry toward integrated system solutions.
Adata Technology will showcase AI applications spanning data centers to edge devices at Computex in June, as it broadens its focus from memory components to AI system optimization. The company highlighted a shift in demand from model training toward inference and real-time deployment, according to CTEE.
At the exhibition, Adata’s enterprise storage brand Trusta will introduce its AI computing integration solution to allocate computing, memory, and storage resources through software. The company said testing indicates the approach could reduce reliance on high-end graphics processing units and lower AI deployment costs by more than 50%.
Originally developed for graphics rendering, GPUs have become a key part of AI computing. They can handle large volumes of calculations simultaneously, helping accelerate AI model training and processing, according to Google Cloud.
Meanwhile, Compal Electronics is also repositioning itself within the AI infrastructure buildout, moving beyond its role as a contract server manufacturer. The company has announced a collaboration with AI infrastructure firm Exascale Labs to develop next-generation integrated AI infrastructure solutions, which will also be showcased at Computex, according to CNA.
Compal said rising AI workloads are increasing power consumption and thermal density in data centers, driving demand for integrated systems that combine computing, cooling, and power management. The company plans to expand in this direction, aiming to offer deployment-ready platforms for cloud service providers and large-scale AI facilities.
Both companies said AI infrastructure is increasingly defined by system-level design rather than standalone hardware performance. They added that customers are seeking scalable, unified solutions.
