The revenue of SMIC's SN fab in first half of the year is USD 800 million. Assume sales price of single 14nm wafer is USD 4,000, the output estimated is close to 35,000 wafers per month. If we take 7nm into consideration, it safe to say the output is close to 30,000 wafers per month.
Also, the biggest customer of SMIC's 14nm process is Phytium.
@tphuang
Bro, if you have the link to SN revenue, could you pm me?
I looked up their Q2 earnings presentation. They don't really break things down there. I just see that their revenue went up a lot from 2021 Q2 to 2022 Q2 without a lot of addition wafer capacity, so SMIC is obviously moving to more advanced nodes.
SMIC but have quite a few 14nm customers, because my calculation is that current demand from Phytium is only about 1000 wafers per month. 1000 wafer * 250 dies/wafer = 250000 chips/month -> 3 million per year of D2000/E2000 chips. Maybe they are stocking up quite a bit of them.
I wouldn't be surprised if they are at 30k wpm or even higher now (just because last report from 2021 was 15k wpm I think? and SN2 factory construction was still ongoing at start of this year), but my guess is they were probably closer to 20k at start of this year. So, if they averaged about 25k wpm in first 6 months and 14 nm is 4k per wafer and N+1 is 8k per wafer (their N+1 is probably between 7 and 10 nm technology). 100k 14nm wafers and 50k N+1 wafers would be $800 million. I think they've really started to mass produce N+1 chips already. Waiting to hear announcements from Phytium, Zhaoxin and other firms.
I think by the end of this year, maybe they can be producing 15k 14nm wpm and 20k N+1 wpm. And once HLMC/ICRD start 14nm production, SMIC probably will slowly move off 14nm as their clients request more advanced processes.
