TSMC, Short-term Thinking and Strategic Decisions

While reading Chris Miller's great book Chip Wars, I became fascinated by the story of Morris Chang and the Taiwan Semiconductor Manufacturing Company, best known as TSMC.

Morris Chang was born in China in 1931. Escaping the advance of the Communist armies, he fled Shangai in the 1940s first to Hong Kong and then to Boston. After a short stay at Harvard, he transferred to MIT to study Mechanical Engineering. Chang joined Texas Instrument in 1958, where in a short time he rose to manager of the engineering section of the company. Later in his career at TI, he would become the group Vice President of the worldwide semiconductor business.

While working for Texas Instrument, as early as 1976, Chang had the vision that the low cost of semiconductors would expand their use in every direction, dramatically increasing the demand for chips. While most companies could design their circuits, most lacked the knowledge to fabricate the chips themselves and would prefer to outsource the fabrication to specialists. Also, the cost of a chip fabrication facility and the R&D involved was so high—and rising—that a company could only be cost-competitive in this market by producing a very large number of chips.

However, other executives at Texas Instrument were not interested in trying new business models. Texas Instrument designed and fabricated its own chips. In 1976 the company was making plenty of money and was in no mood to invest in markets that didn't yet exist. In the early 1980s, after 25+ years in Texas Instrument and being passed for CEO, Morris Chang left Texas Instrument.

In 1985, Taiwan's government officials hired Chang to lead Taiwan's chip industry. In 1987, with strong backing from Taiwan's government, Chang founded the Taiwan Semiconductor Manufacturing Company, or TSMC, the first dedicated semiconductor foundry. They would not design chips but laser-focus on forging semiconductors with cutting-edge technology for other companies. In Chang's mind, TSMC's business model would allow them to produce the huge number of chips required by the huge up-front investments in R&D and fabrication facilities.

Chang knocked on the door of both Intel and Texas Instrument for additional funding, but he was turned down because they weren't interested in the project. The rest of the funding came from Phillips, the Dutch semiconductor company, and from wealthy Taiwanese groups.

In 2005, after successfully leading TSMC to be the world's leading semiconductor foundry and making Taiwan a central player in the global semiconductor ecosystem, Chang stepped down from the role of CEO at 74. He remained as Chairman of the board.

Three years later, the 2008-2009 financial crisis lead to consumers buying less electronics, and tech firms stopped ordering chips. In the middle of the crisis, Chang's successor was following the CEO manual: laying off employees and cutting costs, reducing investment in R&D. That was what the rest of the industry was doing, and what the stock market investors expected.

But Chang had lived through every industry cycle since the invention of the integrated circuit. He was certain that this crisis would pass, sooner or later. The crisis would pass, and when that happened Chang wanted TSMC to be sitting at the center of the semiconductor industry. This required the capacity to produce chips for the biggest customers. He also wanted to win in the smartphone business, especially from Apple which at the time was buying its chips from Samsung.

Chang took the bold step to fire his successor and became again CEO of TSMC. The company's stock price fell. Investors thought Chang was launching a risky spending program, compromising profits. Chang thought instead that the real risk was accepting the status quo. He not only rehired the workers that had been laid off, but he announced multi-billion dollar increases in capital spending to double down on production capacity and R&D.

A couple of years later the crisis was over. Today TSMC manufactures 54% of all semiconductors in the world. Samsung stands in second (17%), followed by UMC (6.9%) and GlobalFoundries (5.9%). Intel Foundry Services division, established after years of hesitation in 2021, ranks n.7 only after Intel acquired Tower Semiconductor early in 2023.

In hindsight, it's easy to say that Chang made the right decisions. The reality, however, is that he faced the same complex environment and incomplete information about the future as other leaders of his time.

For leaders, procrastination many times takes the form of delaying important long-term decisions in favor of short-term rewards. Short-term thinking—like trying to please from quarter to quarter investors at all costs, or running industrial companies like they were only financial entities—hinders a leader's ability to take long-term strategic decisions. Someday what was the company's long-term future becomes the present reality.