The federal government’s Sovereign AI Compute Strategy already accepts part of that premise, including up to $1 billion for public supercomputing infrastructure.⁷ The next step is connecting compute policy to semiconductor policy, energy policy, and industrial policy before another country makes those decisions for it.
The most important part of the supply chain, however, is not the machine. It is the people who know how to make the machine work.
The 90 percent figure is usually described as a supply-chain fact. It is also a human fact. Taiwan’s semiconductor power lives in engineers who understand yield, technicians who recognize process drift, managers who keep clean rooms running, suppliers who know what can go wrong, and students who become specialists through years of repetition.
A semiconductor facility without trained people is an expensive building.
If Taiwan is blockaded, invaded, or brought under Beijing’s political control, the democratic world could face a semiconductor talent crisis as well as a chip shortage. Canada should prepare before that question becomes desperate.
A Taiwan Semiconductor Talent Corridor would be a practical start: expedited work permits, permanent-residence pathways, university placements, family reunification, relocation support for Taiwanese firms, recognition of technical credentials, and serious security screening. It should be targeted, not rhetorical, and tied to a national industrial plan.
That welcome would also have to be governed. Canada cannot absorb unlimited immigration without consequences for housing, schools, health care, infrastructure, and public confidence. The lesson is planning, not retreat.
Canada needs one table where the pieces stop being pieces: Ottawa, Quebec, Ontario, Indigenous governments, universities, pension funds, IBM, AI companies, power authorities, and potential partners such as TSMC and Nvidia. Its job should be narrow: decide what Canada can build, where it can build it, who will power it, who will finance it, and which people must be trained or welcomed to make it real.
Call it a Semiconductor Resilience Authority if a name is needed. The name matters less than the discipline. Its first milestone should be a TSMC-Canada proposal within twelve months. Its horizon should be ten years.
A Taiwan Strait crisis will not wait for Canada to finish studying its options. It will move through prices, factories, hospitals, data centers, defense systems, and the hidden machinery of ordinary life.
Canada cannot control China’s ambitions, Taiwan’s choices, or America’s reliability. It can decide what capacity to build before the emergency arrives.
If Canada waits, it will buy chips from whoever controls the bottleneck.
If it acts now, it can help make sure the knowledge that built Taiwan’s semiconductor miracle has somewhere else to stand.