Probably just want grab a bigger share of that US CHIPS for America act. I don't think that the US company will buy Chinese semi equipment?
Intel Buying Chinese Developed and Manufactured Tool Despite Alternative Suppliers And US Subsidies
Probably just want grab a bigger share of that US CHIPS for America act. I don't think that the US company will buy Chinese semi equipment?
Intel Buying Chinese Developed and Manufactured Tool Despite Alternative Suppliers And US Subsidies
Great great great great news.
Still think the first batch of immersion machines whatever they come out is to meet the needs for chips of sanctioned companies like Huawei.based on your assumption, then, yes, 15 SMEE SSA800.
But then you'll also need iLine, KrF, ArF systems from SMEE, too to make the 28nm chip in our example.
But you don’t have to use the most advanced machine for every exposure. I I guess 5 DUVi maybe enough for advanced node.It varies between DRAM, Logic, 3D-NAND and also between nodes. If we are talking SMIC's new Beijing fab, then it'll need something like 1 million to 1.3M exposures per month. So number of tools required would depends on how many exposures SMEE SSA800 can pump out per month (this would depend intrinsically on the SSA800's wafer-per-hour capability, its uptime, & the fab's utilization).
I don't know what SSA800's specification is so can't speculate.
For ASML, I found this on semiwiki.com:
View attachment 93690
If we take 2019 data, take median number, then I estimate up to 10 ASML immersion systems needed.
For SMEE, again, I don't want to get into trouble and accused of being biased. So allow me to trouble you to do your own speculation & math based on how fast you think SSA800 is relative to ASML's tool. ; )
I was talkng only about the critical layers that require immersion. There are like three times more layers that needs iLine/KrF/ArF that are not part of my estimate.But you don’t have to use the most advanced machine for every exposure. I I guess 5 DUVi maybe enough for advanced node.
Ok. I believe your calculation.I was talkng only about the critical layers that require immersion. There are like three times more layers that needs iLine/KrF/ArF that are not part of my estimate.
Damn, ACM is on fire. If they can expand into etch products like speculated that would mean China has 2 etch suppliers - both AMEC and ACM.
In transistor density. Not the same thing as transistor performance. Latter is more important. Intel’s chase to maintain density at each node shrink actually ended up slowing it down on matching transistor performance.I kept reading many people say this on SDF. Intel has process equivalent to tsmc/Samsung 7nm for at least 3 or 4 years now.