News & Events

Structural advantage

TSMC's had 24.9% growth in sales in 2021 and it is estimated that it may be close to 30% in 2022. TSMC chairman Mark Liu has said the foundry has a structural advantage. I don't know how people understand the characteristics of the pure-play foundry, and what "structural advantage" means.

TSMC is almost the only company in the world that has all three advantages essential to competition at the same time: manufacturing capability, client structure, and ecosystem. As we all know, TSMC's manufacturing capability is not only capable of advancing to 5nm or 3nm, but also the industry leader in production capacity. And more than half of the world market share means that TSMC is far ahead in terms of customer structure.

In terms of ecosystem, for example, TSMC usually maintains R&D funding of about 8% of its sales. Granted that Samsung Electronics' foundry market share is about 30% oof TSMC's, Samsung would have to invest at least 24% of its revenue in the System LSI division to achieve TSMC's R&D scale. The upstream and downstream equipment, materials, and EDA/IP industries are all following closely to form an ecosystem.

TSMC can even neatly divide its R&D team into two teams and deploy them in a "leap frog" mode, so that even if its rivals have a chance to grab customers in a single project, they will still need much more time to bring about structural changes. That's why TSMC funder Morris Chang said that Intel CEO Pat Gelsinger needs time. As a challenger in the market, it would be a dream come true to overtake your rivals; but very often it is more likely you crash out first during the overtaking attempt.

If Samsung's recent leadership adjustments at its foundry business resulted from the unit's below-par performance, it is easy to understand how high the possibility of making a mistake is. We all understand the basic reasoning, but trying to master the three competitive factors at the same time is very likely to fall into the dilemma of three monks picking water. Furthermore, is there any chance that TSMC's model can be implemented in other industries?

When TSMC started its foundry business in 1987, it seemed to be a relatively small market segment compared to IDMs. However, the key is that those at the helm saw the opportunity for IC design companies to grow and the reality that they could not afford to operate their own foundries. TSMC has evolved from a "trusted partner" to an "indispensable partner" over a period of at least 20 years.

From my own experience, DIGITIMES Asia chose to be a relatively small company, and to focus only on the electronics industry. Our business logic is to concentrate our resources to achieve an overwhelming advantage in the local market, as there are no obvious competitors due to differentiation.

Therefore, DIGITIMES Asia can focus on creating daily real-time news and publish more than 300 reports a year across semiconductor design, manufacturing, and various major ICT industries. In addition to speed and depth, our constantly optimized database gives us an advantage in the changing Internet of Everything and the decentralized production mechanism brought by G2.

Secondly, the market demand exists and is getting more and more diversified, and more and more large companies are coming forward to discuss "strategic plans": the more and better the customer structure, the better one is armed. The difficult industry analysis is similar to the advantage in manufacturing capability. Customer's expectations of diversified business opportunities are our opportunities, and that's what we have learned from TSMC.

We can't become as big as TSMC, but proper understanding and learning can be very helpful to the company's operation.