Four major challenges facing TSMC

Latest update time：2021-09-01 10:04

Reads：

Source: The content comes from " Caixun ", thank you.

On November 4, TSMC's stock price exceeded 300 yuan, and its market value reached 8 trillion yuan, surpassing semiconductor giant Intel and becoming the number one in the semiconductor industry.

However, behind such a dazzling halo does not mean that the future is bright. There are still many challenges waiting for TSMC.

Challenge 1: Has Moore’s Law of production process reached its limit?

台积电一向以技术领先著称，因此外界最关切者，莫过于台积电向来依据摩尔定律延续晶圆代工业务，随着纳米制程持续微缩，线宽尺寸恐怕已接近极限，这会影响到台积电的竞争优势吗？

On November 2, TSMC founder Morris Chang did not mince words at the TSMC sports meeting: "No one knows when Moore's Law will end. " Because as early as 1998, someone asked Morris Chang and former Intel CEO Craig R. Barrett when Moore's Law would become invalid?

At that time, Craig and Morris Chang both answered about 20 years, but now it has been proven that both were wrong. In the future, 5nm and even 2nm will be developed. But what will happen after 2nm? Will the industry's anxiety about Moore's Law repeat itself?

In response to this question, TSMC Chairman Liu Deyin reiterated: "In the past, Moore's Law relied on density. Now, semiconductor progress is not only about process miniaturization, but also about logic operation density or computing power as an indicator of progress. "

In order to solve the problems of computing density and heterogeneous architecture integration, advanced packaging is regarded as one of the best solutions to continue Moore's Law. Intel and Samsung have made great breakthroughs in related technologies. TSMC has also entered the field of advanced packaging since 2009, such as CoWoS (wafer-on-substrate packaging), InFO (integrated fan-out packaging), SoIC (system integrated on chip), etc. In April this year, it also officially launched the world's first 3D IC (three-dimensional chip) packaging technology, which is expected to be mass-produced in 2021.

TSMC President Wei Zhejia also pointed out that simply reducing line width is not enough to meet the standards required by new technologies. TSMC has 6 to 7 years of production experience in packaging layout, and future 3D ICs will be able to achieve the performance and structure required by customers.

In other words, the line width density of the chip is no longer the only factor that determines whether Moore's Law will continue. Instead, more diversified chip design and manufacturing methods can make Moore's Law applicable to a wider range of areas.

Challenge 2: Talent Gap: Where will the 8,000 researchers come from?

9月，在「科技智库领袖高峰会」会末，刘德音以半导体协会理事长的身份，向经济部部长沈荣津提出建言，提到半导体人才短缺的问题。

Liu Deyin said that the way for Taiwan's semiconductor manufacturers to survive is to lead in technology. However, in recent years, both China and South Korea have increased their semiconductor research funding. Taiwan should be prepared for danger and expand its basic research layout. "The fundamental problem of Taiwan's semiconductor talent shortage is the lack of research funding each year, which leads to a shortage of professors willing to conduct in-depth research, which in turn affects the number of students and ultimately the number of semiconductor professionals. "

Liu Deyin has confirmed at the Semiconductor Annual Conference that a research and development center will be established in Baoshan Township, Hsinchu County, which can accommodate 8,000 engineers to invest in basic research in technology, material research and development, and the semiconductor industry in the next 20 to 30 years. "This will be a center similar to Bell Labs," Liu Deyin has high hopes for the laboratory. But where will these talents come from?

In April this year, TSMC put up a whole row of recruitment flags along the road leading to Micron from the Central Science Park plant. However, because it was "too direct", it alarmed Micron's senior management. Finally, after a direct call to TSMC's senior management, the flags were finally removed.

Caixun asked the Central Science Park Administration for confirmation, and spokesperson Shi Wenfang responded: "We did receive an application from TSMC to plant a Roman flag on the road in Central Science Park. " It is clear that the competition for talent in the market has reached the level of "hand-to-hand combat."

Challenge 3: The influence of China and South Korea, how to respond to market threats?

As ZTE and Huawei have successively suffered difficulties from the United States, China is stepping up its efforts to develop its own semiconductor technology industry chain.

Through policy promotion and capital injection, SMIC and Huali Microelectronics have successively announced that their 14nm process will enter mass production. SMIC's process layout has taken a big step forward. Although the yield rate is still not comparable to TSMC, it may still produce a certain substitution effect.

In addition to the competition from technology, another challenge comes from the market. As Chinese wafer foundries mature in mainstream and high-end processes, TSMC's revenue from Chinese customers, which accounts for nearly 20%, may be further affected. More urgently, in the current situation, TSMC may face the risk of choosing sides. No matter which side it stands on, it may face severe challenges. On the other hand, South Korea is also trying to take advantage of Samsung through policy layout. How to reduce the related negative impact is a test of the wisdom of TSMC's senior management.

Challenge 4: With the rise of emerging technologies, will quantum computers and satellite communications disrupt the market?

Since quantum computers are related to national competitiveness and whether they can gain an advantage in information warfare, they have become the core of strategic technology development for countries such as the United States and China. In the past, both academia and industry have made every effort to develop them, with investment reaching hundreds of billions of dollars each year. In contrast, Taiwan's Ministry of Science and Technology only approved a budget of NT$70 million for the first time in 2018, which is just a drop in the bucket compared to the United States and China.

Barry Lam, chairman of Quanta Computer, once said that even quantum computers will definitely not be separated from silicon in the future. TSMC ’s mastery of silicon materials can be said to be one of the best on the planet. If quantum computers really become popular, TSMC will also benefit from it. This is also the biggest reason why Liu Deyin said that TSMC will not be absent from quantum computers in the future.

Although the mainstream is superconducting quantum close to absolute zero, room-temperature quantum computing technology based on silicon materials is still indispensable and has greater development potential, such as the spin electronics technology proposed by Intel and the optical quantum architecture that China is fully developing.

The manufacturing process of both is similar to that of existing mainstream semiconductor chips and does not require the most advanced process nodes. Perhaps the Taiwanese industry can also start research and development from this technology, and combined with TSMC's process development, it will bring better overall results.

In addition, although 5G and AI are hot topics, satellite communication technologies such as Starlink led by SpaceX, which have the opportunity to subvert traditional network infrastructure thinking, are also worthy of attention. Taiwan already has the ability to design satellites on its own and has a place in the international satellite parts and technology supply chain. As the most upstream manufacturing technology developer, how TSMC will lead Taiwan to deepen its layout in related markets is being watched with bated breath by the world.

*Disclaimer: This article is originally written by the author. The content of the article is the author's personal opinion. Semiconductor Industry Observer reprints it only to convey a different point of view. It does not mean that Semiconductor Industry Observer agrees or supports this point of view. If you have any objections, please contact Semiconductor Industry Observer.

Today is the 2130th issue of content shared by "Semiconductor Industry Observer" for you, welcome to follow.

Latest articlesabout

■SiC giant, rebirth: how to predict the future?

■Apple chips may hit Qualcomm hard

■Chip cost per car: soaring to $1,000

■TSMC 2nm, important information

■Huang Renxun's latest views

■The risks of this type of chips that are promising have increased significantly!

■NPU, how to see it?

■Storage giants are abandoning DDR 4

■Intel, why?