When U.S. Treasury Secretary Scott Bessent and Taiwan’s envoy to APEC, Lin Hsin-i (林信義), met on Friday, they discussed chips. Bessent wanted to know how Taiwan built its world-leading chip industry and listened with “great interest and patience” to the answer, Lin recounted.
Depending on how patient Bessant really was, the story Lin told him might have started in a Taiwanese breakfast shop, where in 1974, officials kicked off a series of discussions that eventually led to Morris Chang’s (張忠謀) recruitment and the founding of TSMC. Around this time, Taiwan was rapidly industrializing, with the government proactively investing in promising industries. This strategy was called “picking the winner.”
The government founded an R&D agency called the Industrial Technology Research Institute, or ITRI, in 1973. The next year, ITRI licensed integrated circuit technology from the Radio Corporation of America. This tech transfer, devised at the aforementioned Taiwanese breakfast shop, marked “the official introduction of semiconductor technology to Taiwan.” Engineers who were sent for training at the Radio Corporation of America returned to Taiwan to establish the country’s first integrated circuit production line.
However, ITRI didn’t have the commercial base to accelerate semiconductor development and keep pace with companies like Intel and Texas Instruments. So K.T. Li (李國鼎), minister without portfolio and the “father of Taiwan’s economic miracle,” recruited Morris Chang to build out Taiwan’s semiconductor industry. Originally from Ningbo, China, Chang moved to the U.S. in 1949 to study mechanical engineering. He worked for 25 years at Texas Instruments in Dallas, where he eventually became general manager of integrated circuits and then group vice president of its semiconductor business. Chang left the company in 1983, feeling like he’d been “put out to pasture” for failing to grow Texas Instruments’ consumer business.
Shortly after arriving in Taiwan to lead ITRI, Chang developed the concept of a pure-play foundry, or a company that manufactures chips designed by other companies. In contrast, American chip giants like Intel and Texas Instruments were vertically integrated, meaning that they both designed and manufactured chips.
In 1987, ITRI spun off the world’s first pure-play foundry, TSMC. The Taiwanese government provided $100 million in seed money and took a 49% stake in the company, just shy of the Taiwanese threshold for a state-owned enterprise. The Dutch company Philips provided $58 million to TSMC at the outset, becoming what Chang once referred to as the only significant “willing” investor.
Today, Taiwan’s National Development Fund remains TSMC’s largest shareholder, but this share has shrunk to about 6%. While the government believed in incubating industries in their infancy, its long-term plan encouraged privatization and competition.
As one expert told Domino Theory last year, Taiwan’s gamble on TSMC was a lesson in “throwing spaghetti on the wall to see what sticks.” The bet certainly paid off. When the fabless industry took off in the 1990s, that’s when TSMC began to grow quickly. These days, TSMC is the biggest and most important semiconductor manufacturer in the world, working closely with chip designers like Nvidia — the most valuable company by market capitalization — to produce at least 90% of the most advanced chips.
When Bessent asked Lin to tell him this story, his curiosity was probably more than politeness, as the Trump administration is keenly interested in building up the U.S.’s semiconductor manufacturing capacity. In March, C.C. Wei (魏哲家), the chief executive of TSMC, traveled to the White House to pledge an additional $100 billion toward expanding its semiconductor production in the U.S., on top of the $65 billion TSMC is already spending to build its first three fabs in Arizona. These investments were encouraged and facilitated by the U.S. government. In September, U.S. Commerce Secretary Howard Lutnick took it a step further, proposing a 50-50 split in chip manufacturing whereby half of Taiwanese chips made for the American market would be made by TSMC in the U.S. Taiwanese negotiators said that Taiwan would never make such a commitment.
Ideally, Trump wants the U.S. to achieve chip dominance not through TSMC, but through an American company: Intel. Intel has faltered since its days as a semiconductor powerhouse, unable to keep up with TSMC’s highly efficient foundry model and the comparatively low cost of producing chips at scale in Taiwan. But Intel is still Trump’s best bet. In August, the U.S. took a 10% equity stake in the company, amounting to an $8.9 billion investment.
By a stretch of the imagination, the U.S.’s proactive role in Intel’s revival echoes Taiwan’s early investment into TSMC. But in a world where TSMC already exists, does Intel have a chance at regaining its competitive edge?
Leave a Reply