News

Under the chip “winter” cycle, TSMC, which is blessed by Buffett, also needs to be cautious

  Buffett and Zhang Zhongmou, two strongmen who both advocate the Focus principle and have reached the pinnacle in their respective fields, have become “partners” and have attracted the attention of the capital market.
  In the third quarter of 2022, Buffett bought more than US$4 billion in Taiwan Semiconductor Manufacturing (TSM), and the average price paid was between US$70 and US$80 per share. Just like he waited decades to buy Coca-Cola, his acquaintance with the semiconductor field may have a longer history. Ten years ago, Berkshire had a brief “fate” with Intel, and now Buffett has bought semiconductor chip manufacturer TSMC on a large scale. Has the logic of betting on the semiconductor industry changed? In addition, although TSMC’s profitability and growth are relatively prominent, whether it is worth 15 times the valuation is still debatable.
Buffett’s relationship with semiconductors

  In 1967, Buffett’s partnership investment company has outperformed the Dow Jones Index by a large margin for ten years, completed its active intervention in the Sanborn map, and realized the release of value from the liquidation of the Dempster Mill. With $30 million in capital gains and 5% of Disney, he’s making a name for himself in the industry and planning to host the first ever Graham superinvestor’s California gathering.
  Introduced by a friend, he got to know Rosenfeld, the investment fund of Grinnell College, the top liberal arts college in the United States. He took his wife Suzie to attend an impassioned speech by Martin Luther King Jr., and was greatly shocked. He became one of the trustees of Grinnell College, and also met the founder of Intel Corporation, Mr. Robert Noyce, through Rosenfeld. It was an era when the semiconductor industry was in its infancy, and Buffett also watched Intel Corporation (originally named NM Electronics) grow more than 10,000 times from its listing.
  In 1993, he met Bill Gates, who recommended two stocks to him: Microsoft and Intel, one soft and one hard. Since then, including dividends, Intel has doubled by 13 times, and Microsoft has doubled by 100 times.
  In the 18th year after he met Gates, Berkshire bought Intel stock for the first time: in the third quarter of 2011, Berkshire bought Intel at $20 per share, and in the fourth quarter at $24 per share Sell ​​it all; in the first quarter of 2012, Berkshire bought Intel at $24/share and sold it at $28/share in the second quarter. Obviously, at that time, Berkshire had no “belief” in Intel.
  It has been another ten years since Berkshire last invested in Intel in the semiconductor field, and now it has bought semiconductor chip maker TSMC on a large scale.
Zhang Zhongmou’s business empire ambition

  As of the close on November 23, TSMC’s market value was US$425.1 billion, and Intel’s was US$123.1 billion. One TSMC was equal to 3.45 times Intel. Intel is now not as good as 1/3 of TSMC’s rising star, and even its market value has been repeatedly surpassed by AMD (Advanced Semiconductor Corporation, the second in the millennium). Why? Where is TSMC’s excellence?
  When it comes to TSMC, one cannot fail to talk about its founder Zhang Zhongmou. Zhang Zhongmou traveled to the mainland and Hong Kong to escape the war since he was a child. He was able to go to Harvard for his undergraduate degree because of his uncle’s introduction, and then transferred to the Massachusetts Institute of Technology to study mechanical engineering. After getting his master’s degree, he wanted to do a Ph.D. Exam, had to go to work. Because of a dollar dispute (really for a dollar!), he went to a semiconductor company, gave up Ford, and found that the semiconductor company’s technology was not good enough, and then jumped to Texas Instruments (TI). He worked part-time during the day and taught himself a few hours of electrical engineering at night. At TI, he took 4 months to improve the original yield of about 0% to 20% through process optimization, which was twice as high as IBM’s 10% yield at the time. As a result, Zhang Zhongmou became a star manager, managed a team of more than 20 people, and was sent by Texas Instruments to Stanford to study for a doctorate, and he got his doctorate in 2.5 years.
  Someone once asked him how he managed to obtain a Ph.D. in such a short period of time. He said a word, which was exactly what Buffett and Gates wrote down on paper when they met for the first time. The secret of their respective success – “Focus” (focus), he basically did nothing else when he was a Ph.D. student.
  After graduating from Ph.D., he returned to TI. At that time, these manufacturers generally required a relatively high operating profit to help customers produce, and he believed that economies of scale were more important, so he directly lowered the profit and was even willing to bear the initial loss to ensure that the demand was growing, and then he used learning When the curve lowers the cost, it starts to make profits, increase production, gain market share, and form a virtuous circle. In this way, he turned Texas Instruments into one of the largest semiconductor players.
  Many people thought that he would become a CEO (chief executive officer), but unfortunately he did not do so for various reasons. In 1978, he was transferred to the consumer group. After five years, he was unable to make a profit, and was directly “refrigerated” by the management of Texas Instruments. He knows that he is not suitable for analyzing customer preferences, because the changes are too fast, and manufacturing jobs, as long as they are low-cost and high-quality, are in line with his mechanical engineering background and competence circle. Knowing there was no hope at TI, he went to General Instruments as COO (Chief Operating Officer), relinquishing that position within 18 months.
  Later, he returned to Taiwan, China, and led the establishment of TSMC. At that time, 50% of TSMC’s equity was owned by the government, and Zhang Zhongmou was required to find investors for 50% of the equity, and Philips finally invested in them. At that time, Zhang Zhongmou did not have any equity, but now he is worth 2.1 billion U.S. dollars, thanks to his subsequent purchase of TSMC stock. At that time, Zhang Zhongmou was already 56 years old. If he failed, he would never have a chance to make a comeback.
  At this time, the gross profit margin of wafer production is only 5%. Everyone uses the IDM (Integrated Integrated Manufacturing) model, and the design and manufacture of chips are integrated. At that time, the boss of AMD even said that it is not a man who has no manufacturing capabilities (RealMenHaveFabs). IDM will only give orders to companies that simply do wafer manufacturing (also known as Fab) when the production capacity is insufficient, and once the production capacity is restored, it will immediately take the order back to its own side for production. Relying on his experience at Texas Instruments, Zhang Zhongmou knows that many IC designers want to go out to start a start-up company, but they are helpless because there is no Fab to help with production, and there is no ability to build a factory for their own production.
  Therefore, instead of being a venture capitalist for these design companies, he became their enabler, taking all the orders from these small companies for production, and hoping for a virtuous circle-more money coming in, better development Manufacturing capacity, better product quality, attracting more small companies, more money to invest in better equipment and improved processes, and so on.
  Their core philosophy is to always empower consumers, never use their own brand, and use the customer’s for all brand names. They have grown up with a group of Fabless giants, including Qualcomm, Broadcom, Nvidia, etc., greatly reducing the capital expenditure needs of these companies. For example, Nvidia only asked the market for 20 million US dollars throughout its life cycle.
  The first article of Moore’s Law is that the size of the chip is doubled every two years; the second article of Moore’s Law is that the manufacturing cost doubles every four years. Therefore, Zhang Zhongmou’s sentence very classicly summarizes the format of this industry—— You’re like running on a treadmill, and when you stop, you fall behind. This principle broke Intel’s “tick law” (tick law refers to the design and production are interlocking, overlapping upgrade evolution, firmly grasp the commanding heights of technology, so that opponents have nowhere to go), let Intel from fell from a height.

  Before and after the financial crisis, Zhang Zhongmou saw the huge opportunity brought by the mobile Internet and wanted to grasp it. He did indeed do it – he won the competition and secured Apple, a major customer (it is said that Intel rejected Apple at the time), making Intel, the old boss for many years, gradually show signs of fatigue.
Is TSMC with a price-earnings ratio of 15 times worth “copying homework”?

  This quarter, Buffett and Zhang Zhongmou, two strongmen who both advocate Focus and have reached the pinnacle in their respective fields, became “partners”, which aroused the attention of the market.
  We might as well look at TSMC as an investment target from two perspectives: profitability and growth.
  From Figure 1, we can see that in the past 16 years, TSMC has been catching up with 1/3 of Intel’s scale, and now it has equaled its revenue scale, and it will definitely surpass Intel in the future—after all, TSMC as a The existence of a platform is to bring together the power of the entire semiconductor industry to confront the last proud Intel of the IDM model. In fact, since Pat Kissinger became the head of Intel, he is also exploring the separation of design and manufacturing, and releasing the manufacturing to third parties. After all, delays in the process mean that Handing over the market to others – AMD has seized this miracle to achieve a counterattack on Intel.
Figure 1 Comparison of revenue between Intel and TSMC

Data source: Company annual report
Figure 2 Comparison of operating profit margins between Intel and TSMC

Data source: company financial report

  At the same time, we can see from Figure 2 that the gap between the operating profit margins of Intel and TSMC has also widened in the past few years. From the perspective of profitability, TSMC’s operating profit margin is 40%+, and the free cash flow rate of return (free cash flow/revenue) is about 25%, and this free cash flow is achieved when the company has been vigorously expanding its production capacity. , if once the capital expenditure is reduced, it will be even more astonishing. Correspondingly, the highest operating profit margin in AMD’s history (realized in 2021) is only 19%. Streaming income is no more than 20%. Therefore, in a sense, the best choice to bet on the semiconductor industry is TSMC.
  The remaining question is, is such a company worth 15 times earnings? Considering that the current semiconductor cycle is still in a downward phase, as a former practitioner of this industry, I would like to remind all investors who are bullish on semiconductors that when this industry is in a major downward cycle, the stock price can fall to a point of complete despair .

error: Content is protected !!