Morris Chang
The Texas Instruments executive who, passed over for the top job, moved to Taiwan and invented the foundry business that now makes the world's advanced chips.
The most consequential business decision of the last fifty years was made by a man who had already been passed over. Morris Chang, born in Ningbo in 1931 and moved as a child through a China at war, studied at Harvard, MIT, and Stanford, spent twenty-five years at Texas Instruments, and rose to run its worldwide semiconductor business. When the chief executive's job went to someone else, he left. In 1985 the Taiwanese government asked him to come and run its Industrial Technology Research Institute, and to do something about the fact that Taiwan had no chip industry to speak of.
The idea nobody wanted
Chang's diagnosis was that Taiwan could not win at what everyone else was doing. The industry ran on integrated device manufacturers: Intel, Motorola, and Texas Instruments each designed chips and built them in their own fabs. Taiwan had no world-class designers and no chance of catching up. What it did have was disciplined manufacturing, an engineering workforce, and government money.
So he proposed splitting the business in half. A company that only manufactured, that designed nothing of its own and therefore competed with none of its customers, could take orders from everybody. Designers without factories could then exist at all, since the capital barrier that kept them out of the industry would be someone else's problem. Chang pitched a version of this at Texas Instruments years earlier and got nowhere; when he went looking for partners in 1986, both Intel and Texas Instruments turned him down. Philips came in for just under 28 percent in exchange for technology licenses, Taiwan's National Development Fund put up roughly 48 percent, and Taiwan Semiconductor Manufacturing Company opened in 1987 with Chang, then in his mid-fifties, as chairman.
What it created
For its first decade TSMC was a modest business making other people's older designs. Then the thing Chang had bet on happened: the fabless companies he had made possible grew up. Nvidia, Qualcomm, Broadcom, Apple's silicon group, and later AMD had no fabs and never needed any, because TSMC was theirs. Each generation of chipmaking got more expensive, which favored whoever had the most customers to spread the cost across, which brought more customers, which paid for the next generation. Intel, still building only for itself, eventually could not keep up. By the 2020s TSMC was manufacturing the overwhelming majority of the world's most advanced logic chips, on an island roughly a hundred miles off the Chinese coast.
Chang retired in 2005, came back in 2009 when the company stumbled during the financial crisis, ordered an enormous counter-cyclical capital investment, and retired again in 2018. In 2021 he described Taiwan's semiconductor position in public as a matter of strategic significance rather than mere commerce, and he has been notably skeptical that American fabs can match Taiwanese cost and discipline, calling the effort expensive and wasteful even as TSMC built in Arizona.
The result is that a single company, on a single contested island, sits at the center of the world's computing supply, and that this arrangement is the accidental product of one executive's career disappointment in Dallas. Every argument about chip export controls, Taiwan's silicon shield, and where the next fab gets built runs back to the split Chang made in 1987 between the people who design chips and the people who make them.