Text/JING Review/ZiYang Proofreader/Zhiqiu
Mass production of 2nm is expected in 2024! Intel's desire to catch up with TSMC may be a pipe dream
On March 24, 2021, Intel CEO Kissinger announced the "IDM2.0" strategy at a global live event.
It is reported that this strategy mainly consists of three parts. First, expand the network of manufacturing plants. Second, the use of third-party OEM methods to increase production capacity. Third, create a first-class foundry business.
Among them, the third target is the core of Intel's IDM 2.0 strategy, which is also the most difficult strategy to achieve.
According to Intel's official statement, the ultimate goal of the IDM2.0 strategy is to catch up with TSMC and get orders from Apple.
It is worth mentioning that when Intel announced the strategy, it had not yet achieved mass production of the 7nm process, while TSMC was already mass-producing 5nm chips, and the technology gap between the two companies was very large.
However, Intel soon unveiled a roadmap for technology upgrades. And according to Taiwan's Economic Daily, Kissinger confirmed that by 2024, Intel will be able to mass-produce 2nm chips, and the time point will be earlier than TSMC.
The industry generally believes that the acceptance of orders by Intel's advanced process chips will not affect the development of TSMC.
In fact, in the author's opinion, it is difficult for Intel to achieve the goal of advanced 2nm in 2024. After all, going from 7nm to 2nm requires at least three technology updates. According to the conventional upgrade idea, Intel's mass production of 2nm chips will need to wait at least four years.
Moreover, even if Intel catches up with TSMC in 2024, TSMC will not lose its status as a foundry brother.
Because TSMC has been sitting on the throne of chip brother for many years, it relies not only on technology, but also on production capacity and customers.
These two advantages are more difficult to surpass than technology. Taking "customers" as an example, TSMC is backed by giants such as Apple, NVIDIA, AMD, and MediaTek, and its annual orders are guaranteed.
Not only that, TSMC customers basically do not take the initiative to introduce the accumulated power production capacity and contact Intel. After all, the relationship between TSMC's production capacity and order demand is "more wolves and less meat", and once a company gives up its share, it will inevitably be supplemented by manufacturers. Choosing Intel is a risk for the customer.
Qualcomm is a typical case, Qualcomm in the early days to save costs, choose Samsung OEM, but two consecutive generations of processors have been Samsung process "pit miserable". So much so that Qualcomm began to communicate with TSMC at the end of 2021 and returned to the TSMC camp.
In addition, the production capacity factor cannot be ignored. In fact, many manufacturers can not do without TSMC, a very important reason is that in addition to TSMC, other manufacturers simply do not have the ability to receive large orders. TSMC monopolizes more than 90% of the advanced chip production capacity of the foundry industry, which is based on the large number of high-end chip talents and the number of high-end chip manufacturing equipment procurement.
And high-end talents need to spend time slowly cultivating, high-end chip manufacturing equipment, not only limited production capacity, even manufacturing manufacturers are few.
Although Intel has a chip foundry business foundation, in terms of high-end chip foundry, the basic conditions are far from TSMC. In the author's opinion, the best result of Intel's foundry business in the next decade is not to replace TSMC, but to replace Samsung and become the second in the industry.
Of course, in the long run, Intel is not completely without a chance of winning. After all, Intel is larger than TSMC, which means that other businesses can transfuse blood into the foundry business and increase the investment in the foundry business.