Taiwan approves 10-year innovation program to address semiconductor industry weakness

Bryan Chuang, Taipei; Judy Lin, DIGITIMES Asia 0


The Taiwan Chip-Driven Industrial Innovation Program (TCIIP) was approved by Taiwan's National Science and Technology Council (NSTC), with implementation to start in 2024.

The NSTC pointed out three challenges the semiconductor industry in Taiwan is facing: 1) there is a lack of system specification to determine power and direct access to markets; 2) there is insufficient capability in advanced IC design and lack of self-developed EDA tools; 3) advanced processing technology advantages have to be maintained while the mature-node IC manufacturing sector is susceptible to cut-throat competition in the market.

The TCIIP, which is to be implemented within a 10-year period, plans to address those challenges with four approaches, including attracting international startups and venture capitalists to come to Taiwan, accelerating innovations in heterogeneous integration and advanced technologies, synthesizing AI and chips to stimulate innovation within the industry, and enhancing the talent incubating environment to attract global R&D talent.

The NSTC has started its first step in connecting international startup ecosystems with local innovation hubs. In the future, more collaboration opportunities will be created for startups to work with Taiwan's IC manufacturers and IC designers.

There are also various projects related to heterogeneous integration and advanced technologies in government-sponsored research entities such as the Industrial Technology Research Institute (ITRI). the program will help upgrade the equipment the researchers need for those projects.

As for the synthesis between AI and IC, the National Center for High-Performance Computing has obtained a budget of NT$230 million (US$7.14 million) to engage in the development of generative AI models based on Taiwan's language characteristics, and integrate with various technology services to promote the development of "trusted" generative AI model.