TSMC accepts the reality: suspend 7nm, turn to 28nm, and grab the job of the mainland fab?

According to the data in the third quarter of 2022, if 28nm is used as the dividing point in TSMC's revenue, then advanced processes below 28nm have contributed 66% of the share, equivalent to two-thirds.

and the process of 28nm and above only contributed 34%, of which 28nm contributed 10%, and above 28nm contributed only 24%.

Judging from the proportion of revenue contribution in recent years, TSMC is constantly advancing towards advanced technology, and can't wait to convert all of them into advanced technology, because the competition of advanced technology is small, and the profit of is high.

However, TSMC has also encountered a problem at present, that is, fewer and fewer customers use advanced technology, and everyone stays on mature technology and is unwilling to switch to advanced technology.

, especially 7nm, because customers originally using 7nm, such as Apple , AMD, NVIDIA, Qualcomm , MediaTek , Broadcom , etc., will migrate 7nm to 5nm, 4nm and future 3nm.

has led to few people using the 7nm process, either using more advanced processes such as 5nm, or staying at 28nm, resulting in TSMC's 7nm process capacity utilization , which fell below 50%.

For this, TSMC has no choice but to accept the reality. Recently, Wei Zhejia, president of TSMC, said that the Kaohsiung plant will definitely be built, but the 7-nanometer process will be delayed, and the 28-nanometer process will be the first.

That is to say, TSMC is still trying to expand 28nm. As for 7nm, it will not be tossed first, so as not to build it and no one will use it, wasting financial, material and resources.

TSMC will expand the 28nm process again, who will have the greatest impact? Of course, it is the manufacturer that is still working on 28nm, especially the SMIC .

In 2021, the total global 28nm wafer foundry revenue will be about 7.2 billion US dollars, and TSMC alone will take two-thirds of the revenue, which is about 54 US dollars. and SMIC took about 10% of the 28nm, which is about 700 million US dollars, less than one-seventh of TSMC.

Now TSMC has to expand 28nm, and even 7nm has to be postponed, focusing on 28nm, then it is bound to grab the job of China Chip International, after all, SMIC is currently expanding 28nm, and its latest technology is only 14nm, 28nm is still a relatively advanced technology.

So it is conceivable that the competition on 28nm will be very fierce. I wonder if SMIC is ready?