TAIPEI (Taiwan News) — National Science and Technology Council Minister Wu Cheng-wen (吳誠文) said the government will boost AI adoption among small and medium-sized enterprises and strengthen infrastructure as part of efforts to build a comprehensive AI ecosystem.

At a council meeting on Wednesday, Wu urged large companies to share their experience in deploying AI-powered systems to help the transformation of small and medium-sized firms. He added the council will continue working with the Ministry of Economic Affairs to provide subsidies and support for AI adoption in the manufacturing and service sectors, per CNA.

Wu highlighted the need for Taiwan to develop “Sovereign AI,” including building domestic data centers and computing infrastructure, keeping data stored locally for security, and establishing AI rules and guidelines.

The council said it has also partnered with the Financial Supervisory Commission and the Ministry of Justice to develop domestic AI models, ensuring they are trained to handle local languages and culture.

In December, the government began operations at a cloud computing and service center in the Southern Taiwan Science Park to bolster domestic AI infrastructure. The center supports industrial research and applications. It is built with earthquake-resistant construction, backup power, and advanced cooling systems to ensure uninterrupted service during emergencies.

President Lai Ching-te (賴清德) called on local tech companies to build their computing infrastructure, data storage systems, and communication networks. He emphasized the use of silicon photonics — a light-based transmission method that replaces electrical signals — to enhance data transfer in Taiwan.

The government also plans to replace all copper wiring in AI data centers with fiber-optic cabling to speed up data transfer and reduce heat and power loss. Wu said the council will unveil the first pilot site for the upgrade by the end of April, with phased expansion planned.

The council added that it will launch programs encouraging tech companies to share part of their computing resources with startups and research institutions. High GPU and electricity costs make it difficult for startups to train their own AI models, and the initiative aims to help lower these expenses.