ReadingResearchRadarInvestment frameworkAI Company Primary Sources
Sign in / Sign up
← Back to reading
NewsMaterial3 min readDisconfirmAI Editorial

Pei-Jen Chang, Director of the Industrial Technology Research Institute: When Resources Are Limited, Technologies Are Categorized into Three Types—Those Where Taiwan Can Win Get Heavy Investment, Those with Supply-Disruption Risks Are Retained, and Others Are Simply Purchased

In an exclusive interview, Pei-Jen Chang, Director of the Industrial Technology Research Institute (ITRI) in Taiwan, outlined his core methodology for forward-looking planning: categorizing technologies based on their potential to succeed and the risk of supply disruption to determine resource allocation. Acknowledging the uncertainty of quantum computing routes, ITRI does not bet on a single winner; instead, it focuses entirely on developing cryogenic electronics—a component required by all major approaches—ensuring that "every leader needs me as their second-in-command."

2026-11-03 15:18 GMT+8

Pei-Jen Chang assumed the role of Director of the Industrial Technology Research Institute (ITRI) in Taiwan in October 2025, having previously served as Deputy Director. He is also a professor at the Institute of Applied Mechanics at National Taiwan University. In an exclusive interview with Vision Week Podcast, he systematically explained how a national-level research institution with limited resources decides "which technologies are worth investing in."

His core judgment is that forward-looking planning must consider four factors: future human and industrial needs, government policy, the development of emerging technologies, and Taiwan's industrial strengths and weaknesses. He believes the fourth factor is the most important because the first three "everyone does," and public literature can provide them; what is truly lacking is the ability to judge "whether Taiwan's industry can actually win."

Based on this judgment, he categorizes technologies into three types. The first type consists of technologies where Taiwan can win; most resources should be invested here, "otherwise you are wasting them." The second type includes technologies where winning is difficult but failure to possess them could lead to supply disruptions; "I must keep these in hand." This segment may not be profitable, but to enable profitability in the first category, it must be sustained. The third type comprises technologies with weak competitiveness and no supply-disruption risk: if three suppliers can be found globally at any time, do not develop them yourself, "just buy them."

Sign in / Sign up
ReadingResearchRadarInvestment frameworkAI Company Primary Sources

Taking quantum computing as an example, Chang believes quantum computers will not completely replace traditional computers but will serve a complementary role. There are multiple front-end routes, including ion traps, photonics, and superconductivity. ITRI collaborates with university professors and Academia Sinica, allowing the professors to study which route is more feasible. ITRI itself focuses on cryogenic electronics, which he believes are ultimately required by all routes—circuits that can operate at absolute zero temperatures of 4 K (approximately minus 269 degrees Celsius). He stated that whichever route achieves a breakthrough will become the leader of that path. ITRI may not be the leader in any single route, but "who wins is actually irrelevant to us." As long as "all global leaders cannot do without me," this is also a viable strategy.

Timing of entry is the second dimension of the framework. Investing too early means research funds may be exhausted before the industry takes off, causing the plan to stall; investing too late means essential patents may already be fully laid out by others. ITRI aims to see manufacturers stepping in "at the latest within four years" after investment. For technologies that may occur in ten or fifteen years but whose exact timing is uncertain, the approach is to "closely follow": rather than investing heavy resources, they continuously observe what the top three global players are doing and quickly catch up when the timing seems right.

This framework also extends to the R&D process itself. Chang requires teams working on robotics and drone projects to "interview at least 10 customers' customers" when defining plans, clarifying service workflows before setting specifications. He also established a quantitative rule: the price of key components for technology-transfer vendors must not exceed 30% of the expected selling price. Previously, expensive parts were used to demonstrate performance, but after technology transfer, vendors struggled to sell the products and had to spend a long time adjusting. Now, components intended for mass production are adopted during development.

It should be noted that this methodology is articulated by the ITRI Director himself and represents the institution's self-explanatory strategic logic. Whether actual resource allocation strictly follows this tripartite classification, and how geopolitical factors such as the "non-red supply chain" influence categorization judgments, require observation of specific project budget allocations for verification. Chang also mentioned that when categorizing technologies, requirements for a "non-red supply chain" must be considered—excluding Chinese-made components under specific procurement rules—which means whether certain technologies need to be mastered domestically cannot be judged solely by competitiveness or supplier count.

Sources

  • techorange.comTechOrange 科技報橘↗
Continue researching →