TSMC: Ending Semiconductor Cyclicality Through HPC

In this week’s article, we will dissect the trajectory, capex and the end market shift of the world’s most dominant leading-edge chipmaker.

Founding History

Morris Chang’s path to founding the world’s largest pure-play foundry business was meandering and atypical.

When he was a young man, he dreamt of becoming a novelist. His father, a frequently relocating banker, however, didn’t think it was a good idea.

With the help of an uncle in Boston, he facilitated Chang’s admission to Harvard. The young undergraduate became instantly enamored with American culture, and the country’s economic and political might of the time. He decided to abandon his childhood dream and elected to become a mechanical engineer instead.

To further his employment prospects in the field, he transferred to MIT a year later. Eventually, he obtained his Master’s degree but failed to defend his PhD thesis on multiple occasions.

Following graduation, after a short stint at an electronics company, he moved to Texas Instruments to embark on a 25-year corporate climb. While at the company, an early experiment led him to a breakthrough in improving manufacturing yield (i.e. increasing the ratio of non-defective semiconductors relative to the entire output) by 20% and propelled him to a string of management jobs. TI also paid for Chang to complete his PhD in electrical engineering at Stanford.

While at TI, Chang came up with his first novel idea. He observed that every new semiconductor manufacturing run had the same problem. It was capital intensive and, during the initial phase, had to go through a learning period, when it was difficult to improve the manufacturing yield.

As a result, companies would simply charge high prices for the initial semiconductor product from the beginning to cover for the high costs of production and the low output.

Chang thought such an approach was flawed. So he engaged the Boston Consulting Group to devise a unique approach of using the learning curve to drive the pricing strategy, early in the life cycle of a new product.

TI and BCG were able to predict the unit costs in volume production on the basis of the actual cost of the first 1,000 units. They discovered that the cost per unit improved, as cumulative unit production increased over time.

So TI decided to sacrifice early profits and introduce a lower price for the initial product. This way the company was able to squeeze out competitors, gain market dominance and secure profitability down the line.