TAIPEI, Taiwan - Intel has outsourced the production of about 15-20% of its non-CPU chips, with most of the wafer starts for these products assigned to TSMC and UMC...
As expected, most IC suppliers experienced weaker demand and generally poor sales results in the first half of this year due to the Covid-19 pandemic and strained U...
With Morris Chang's stepping down, TSMC will enter the performance of their second act, and the key protagonists will switch from having one strong leader to a partnership. Whether or not the story of this team can continue the win the confidence of clients and continue to win the leading market position will be the element of TSMC that will be most worthy of everyone's attention.
The Efficiency of Dual Leadership Still Needs to be Observed
TSMC will be under dual leadership following the retirement of Morris Chang, and it will truly be a new model of leadership. Under this model, the two leaders, CEO Mark Liu and President C.C. Wei will jointly lead TSMC with parallel decision making.
In this business structure, the CEO will lead the Board of Directors as the highest external representatives for TSMC and have the final say in policy decisions. Meanwhile the president will be responsible for the company's business strategies and leading the company, and the Chief Financial Officer will also report to the president.
The advantage is this dual leadership model is that they can avoid the risk of having decisions made by a single person so that operations will not be overly reliant on the acting leader's abilities and will. From a long-term perspective, this is a relatively stable and trustworthy model.
On the other hand, the disadvantage of dual leadership is that the effectiveness of implementation may face challenges.
If the two leaders are a good match, then operations will naturally be stable and smooth, and they may even be highly effective through through the leaders' complementarity and mutual assistance. However, if they are poorly matched, it will be difficult for internal management to proceed to external operations, and in the long term it will be difficult for TSMC to maintain their well-know penchant for discipline and efficiency.
However, observing the two successors' collaborative situation during the past three years, it currently does not look like there are any clashes or communication problems between the two leaders, and they are both willing to mutually respect one another's viewpoints. Furthermore, Morris Chang has commended the addition of the two successors and believes that their personalities are mutually complementary.
“It is difficult to get around the fact that sometimes opinions will differ; however, we will look at the strong points of each other's opinions,” replied Mark Liu to reporters' questions regarding potential issues in their collaboration.
The Continuation and End of Moore's Law
Morris Chang has no doubts that Moore's Law is the best spokesperson, and the explanation for the perfection that he has brought to TSMC is the course of semiconductor development in accordance with this law. Moreover, TSMC are the main profiteers under this law, it could even be said that they are the main champions of this law.
Nevertheless, under the continuous development of microtechnology, the physical limits of silicon chips are within sight, and the continuation and elimination of Moore's Law is becoming another key to development for TSMC.
It was not until September of this year that Morris Chang rediscussed the limits of Moore's Law. He believes that on flat surfaces, Moore's Law still has possibly ten years worth of time left. The switching of processes will be able to facilitate development to as little as 3nm; however, it is unknown what number can be reached below 3nm. This question is bound to become the turning point for TSMC's development.
Looking at the situation in terms of the schedule currently announced by TSMC, the mass production time for 3nm will be 2022. In an additional two years, it will be 2024, and by then the direction of change for semiconductor design and the kind of production equipment and semiconductor materials support used will determine when the end time for Moore's Law will occur.
However, it is certain that by 2024 TSMC will extend the reach of Moore's Law and push it to the limits of silicon chips. In addition, the functionality of integrated circuits will also continue to double, and unless an accident occurs, a large number of contemporary technology products will continue to have bright outlooks.
Continuing to Dominate the IC Foundry Market
Looking at the present, pure-play foundry is definitely a far-sighted business model. Regardless if viewed from the standpoint of manufacturing technology or business management, it is an inevitable and necessary link in the semiconductor industry chain.
This is especially the case with chip design tending towards greater complexity, manufacturing processes shrinking to tinier and tinier sizes, and chip manufacturing costs exceeding the burdens that single companies can shoulder. Taking 20nm as an example, casting costs have already reached US$130 million (approximately NT$4 billion). If there are no specialized chip OEM manufacturers, for semiconductor technology, if there are aspirations to make breakthroughs in semiconductor technology, they will have to be extremely small.
Semiconductor marker research institute IC Insights predicts that between 2016 and 2021 the Compound Annual Growth Rate (CAGR) will reach approximately 7.6% for global pure foundry OEMs with the market scope growing to US$72.1 billion in 2021. Hence, it is almost to the point where one could state that following the creation of Morris Chang's chip OEM model, TSMC claimed a place for itself which is difficult to be replaced. This is especially the case in the domain of advanced manufacturing processes where the experience and technology accumulated by TSMC has already turned the heads of competitors and reached a lead that is difficult for them to catch up to.
In the second half of 2018, TSMC's 20nm factory will commence mass production in Nanjing, China where they will be nearby to serve the largest future global semiconductor market, and this will bring about a large amount of momentum for TSMC's growth. Furthermore, it is certain that comprehensive IoT and next-generation computing will drive up demand for semiconductors, and TSMC has already established the following four major platforms: Mobile Computing, High Performance Computing, Internet of Things and Automotive Electronics, wherein they serve clients' needs in each distinct market.
The Search for Unknown Markets and Technologies
TSMC is of course sailing smoothly, yet their past investments in solar energy and LED enterprises have failed to yield results. However, this proves that TSMC is still searching for market opportunities outside their own industry, and even with the early sale of Toshiba Semiconductor, TSMC will evaluate and clearly seek out new markets as one of TSMC's internal goals.
As a result, as TSMC enters into a new dual-leadership system, it will be worth observing whether or not their evaluation foundation for investing in new industries changes.
In terms of technology, it will be necessary for TSMC to work out a plan for technological development after 3nm before there will be something for the business plans for their clients and themselves. It is currently possible that they will pivot towards 3-D IC development, and continuing to explore the world after 2nm is another possible choice.
If they pivot towards 3D IC, there is still the problem of through-silicon-via (TSV) perforated technology, and they are still waiting for improvements in thermal and circuit performance after stacking. In addition, whether or not the overall manufacturing costs of 3D IC will drop down within the ideal scope is also a key to the development of this technology.
If they continue to shrink below 2nm, however, all of the corresponding chip materials and equipment for related manufacturing will be within reach. As far as the materials go, the current popular substitute is graphene, and because it is much tougher than silicon and has better thermal conductivity, their chip frequency theory can reach 300GHz.
However, stone abrasive materials are highly expensive, and they are currently in the experimental phase; therefore, it is still unknown whether of not they can be used commercially. Nevertheless, companies, such as IBM, GlobalFoundries, Intel, and Samsung are all investing in this line of research.
The Challenge of Maintaining the Predecessor's Achievements
This year is the thirty-year anniversary of TSMC's founding, and reflecting on this time period, we can see an emerging company that has never ceased to grow, continually pushed forward, and moved into the unknown. Even so, “growth, life, decline, and emptiness“ are the basic principles of the world cycle, and they will inevitably turn to the next stage of development.
If during the era of Morris Chang TSMC went through the “growth” stage, then the next stage is “life”, and whether or not TSMC will be able to have a long life the under the next leaders will be a matter that all of Taiwan and the rest of the world will observe closely.
(TR/ Phil Sweeney)