中文網
Taipei
Wed, Feb 1, 2023
21:49
cloudy
18°C
CONNECT WITH US

Opinion: Can Gelsinger's wishful goal be achieved?

Rocky Uriankhai, DIGITIMES Asia, Taipei 0

Intel CEO Pat Gelsinger expressed at the DAVOS Forum on the 17th of this month (January 2023) that he hopes the global chip market will reduce its dependence on Taiwan.

The CEO also believes it will help the world economy.

The US is the global hegemony that dominates the international order. Intel is the leading chip manufacturer in the US. It is not difficult to understand Gelsinger's expectations. The question is, what is the ideal regional ratio of global chip production capacity distribution in Gelsinger's mind?

The answer to this question was answered by Gelsinger himself in an interview with CNBC in May last year (2022). He first emphasized that the world will "never" be completely independent of Taiwan, which showed that he saw the trend of "decentralization" of the supply chain caused by the competition between the U.S. and China.

In other words, chip manufacturing will no longer be excessively concentrated in one single region or country, but Taiwan's importance in the global semiconductor supply chain will not just vanish either.

Secondly, he expected that ten years from the present time, the chip production capacity of the US plus the EU and Asia will be evenly divided. However, the EU and the US together account for only 20% now, while China, Japan, South Korea, and Taiwan in Asia account for 80%.

In addition, Gelsinger emphasized that this 50% to 50% goal was approved by US Secretary of Commerce Raimondo and European Commission President Ursula von der Leyen. However, he did not mention if his vision of "half and half" has been recognized by leaders of other Asian countries.

Recognition means intention. Whether or not the goal of strengthening the US plus Europe can be achieved depends not only on the capabilities of the US and Europe but also on the capabilities of its competitors, Asia because international relations focus on relative strength.

The following is the analysis of the strength and willingness of the four Asian countries.

Taiwan first - which we can find out from the most significant chip maker TSMC. On January 12 (2023), TSMC held an earnings call. In addition to explaining the financial report for the fourth quarter of 2022, it also proposed, in a 5-year period, to adjust the overseas production capacity of 7nm and below to 20%, while keeping 80% of capacity in Taiwan. This is far from Gelsinger's goal of equal distribution. In other words, Gelsinger will hope that TSMC continues to move its production capacity abroad for another five years. TSMC seems to have launched a global layout, but talent and manpower are a big problem when building factories overseas.

What Gelsinger pays attention to is geopolitical security. Therefore, when TSMC and Samsung build factories in the US, their production capacity can be counted as 'Made in USA.' Therefore, if these two major chip manufacturers continue to expand their production capacity in the US, which will definitely help Gelsinger's goal of a 'balanced ratio' between Europe plus America, and Asia. This is also the reason why TSMC's 21st factory in Arizona, US has announced that it will increase production capacity before the launching ceremony.

There's also South Korea when it comes to Asia. Samsung and SK Hynix, whose main products are memory chips, have not only promised to expand production capacity in the US but have also built and expanded factories in Xi'an and Wuxi, China. In other words, if South Korea hopes to make profits from the two major chip markets at the same time, it will add weight to both ends of the 'Europe+US vs Asia' balance, which will not help Gelsinger's goal of reducing the proportion of chip production capacity in Asia. This is the reason why the US dissuades South Korea from expanding its production in China, and it is also the reason why South Korea resists joining Chip4.

The third one, Japan, which once dominated more than half of the world's chip production capacity in the 1980s. After close cooperation between the government and the private sector, Japan is currently adopting a two-pronged strategy in terms of research and development. It has established Leading-edge Semiconductor Technology Center (LSTC) in cooperation with IBM. Regarding manufacturing, in addition to inviting TSMC to set up a factory in Kumamoto, Japan, eight Japanese companies were also called upon to establish Rapidus, focusing on making chips.

Japan's goal is to start producing 2-nanometer chips from the current 40-nanometer by 2027, with the intention of restoring its former glory in one fell swoop. Although the estimated production capacity is still uncertain, it will definitely not be satisfied with the current global chip production capacity accounting for only about 6%. In other words, Japan, which is located in Asia, also hopes to increase its local production capacity.

The fourth one is China, which lags behind the West in terms of semiconductor industry. Judging from the past financial reports and some public reports of SMIC, the leading chip manufacturer in China, its 14nm chip production capacity - if the US keeps controlling the upper stream of semiconductor supply chain, especially the lithography system of EUV - may be very difficult to make a great breakthrough in the advanced process within 10 years.

But on the other hand, its 28nm and legacy process chip production has made great progress. Regardless of SMIC's level of technological advance - only in terms of production capacity, China's total chip production capacity in the next ten years can still add weight to the Asian side of the balance. But the 70% self-sufficiency target set in the 'Made in China 2025' plan released in 2015 should not be achieved.

Finally, Europe, which is the teammate of the US on the same side of the 'Europe+U.S. vs Asia' balance. In March of 2021, the European Commission announced the '2030 Digital Compass: the European way for the Digital Decade' in which the EU aims to increase its chip production capacity from the current 10% of the global production capacity to 20% in 2030. The EU is also targeting 2 nanometers, which is as ambitious as Japan.

However, this is not the first time that the EU has felt that its position in the global semiconductor industry was too backward. As early as the end of May 2013, the EU proposed a 'New Electronics Strategy', but the result was not very good. Europe has fallen far behind Taiwan and South Korea, and can only join with Japan and China to become the third leading group.

In other words, Europe has not been able to catch up in the past ten years. In the next ten years, if the objective conditions such as industrial structure, total population, and R&D capability, hold the same, it will still be difficult to increase production capacity based on strategy or plan alone. This is also the reason why Germany tried to win over TSMC to set up a factory.

Based on the details above, we can imagine that Gelsinger expects that ten years later, the ratio of chip production capacity between Europe plus America and Asia will reach parity, and the 'EU+US' side has to be doubled. At the same time, the Asian side, namely China, Japan, South Korea, and Taiwan, must also cooperate. The challenge is still huge.

Global wafer foundry industry analysis and forecast, 2022
Global server shipment forecast and industry analysis, 2022