Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
I predicted this exact copium over one week ago:
I just returned from a Huawei store in Shanghai. Huawei sales rep was deliberately vague when asked about 5G chips in Mate60, saying that I can "test the speed on my network for independent validation to be sure" (with a wink, wink, nod). I believe this "strategic ambiguity" is deliberate to troll the American government.... I approve.
5-Because this year, Biden and the White House Department of Commerce made a decision to allow SMIC to provide a license to manufacture 5nm chips for Huawei.
That would be even more impossible than Huawei buying ASML tools that can get second hand or from others, is political suicide in D.C. to help Huawei in any shape or form, their objective is pretty clear. they are trying to kill the company.
6. Can SMIC continue to make breakthroughs in process technology and carry out 4nm-level chip foundry?
Is not impossible, EUV and especially Low NA EUV is not a "red line" technology that without it smaller features cannot be patterned, there is a reason why some immersion tools from ASML was export controlled, because is all about the overlay, matching, dosage control and highly capable OPC techniques. Multiple patterning techniques are quite capable, especially when used with good deposition tools and selective etching tools.
7. Does Shanghai Microelectronics (SMEE) have the opportunity to develop EUV shadow exposure equipment?
I don't know if SMEE would be the integrator for EUV in China but there is an new alliance for EUV lithography taking shape in China right now, given the priority of this technology, China doesn't have the luxury to wait until SMEE get its act together and the fact that CIOMP and SIOM already tested a demonstrator technology, I think it could be done by a joint venture and with SMEE as a supplier.
That would be even more impossible than Huawei buying ASML tools that can get second hand or from others, is political suicide in D.C. to help Huawei in any shape or form, their objective is pretty clear. they are trying to kill the company.
6. Can SMIC continue to make breakthroughs in process technology and carry out 4nm-level chip foundry?
Is not impossible, EUV and especially Low NA EUV is not a "red line" technology that without it smaller features cannot be patterned, there is a reason why some immersion tools from ASML was export controlled, because is all about the overlay, matching, dosage control and highly capable OPC techniques. Multiple patterning techniques are quite capable, especially when used with good deposition tools and selective etching tools.
7. Does Shanghai Microelectronics (SMEE) have the opportunity to develop EUV shadow exposure equipment?
I don't know if SMEE would be the integrator for EUV in China but there is an new alliance for EUV lithography taking shape in China right now, given the priority of this technology, China doesn't have the luxury to wait until SMEE get its act together and the fact that CIOMP and SIOM already tested a demonstrator technology, I think it could be done by a joint venture and with SMEE as a supplier.
Could it Mate60pro uses domestic CMOS，rather than Sony's ?
Did you get a chance to play with the phone? How was it?
Look into which companies supply the satcom chips for Mate 60 Pro
first is RX60003EQK that @tonyget already posted earlier. According to this, it's similar to RX3606 RF chip and produced by Runtech (广州润芯)
This company is 50% owned by (海格通信)
who specializes in wireless communication, beidou navigation & aerospace areas.
Other chip that's in there is HTD1010 by what we think is Hwa Create
you can see they already advertise 2 mobile satcom SoC, both on 40nm process. One is civilian and the other is military grade. So, HTD1010 is likely a special version for Mate 60 Pro
It is most likely that both chips are needed for all the satcom process that Mate 60 Pro offers. Interesting enough, it supports both beidou and GPS
first is RX60003EQK that @tonyget already posted earlier. According to this, it's similar to RX3606 RF chip and produced by Runtech (广州润芯)
It comes with 2 high performing LNA and built-in loop RF filter. Can support FDD/TTD 2 format
This company is 50% owned by (海格通信)
who specializes in wireless communication, beidou navigation & aerospace areas.
Other chip that's in there is HTD1010 by what we think is Hwa Create
you can see they already advertise 2 mobile satcom SoC, both on 40nm process. One is civilian and the other is military grade. So, HTD1010 is likely a special version for Mate 60 Pro
It is most likely that both chips are needed for all the satcom process that Mate 60 Pro offers. Interesting enough, it supports both beidou and GPS
That is what I believe.
Announce the DUVi and the EUV together, for maximum attack.
If it is a raid, well, that is probing defenses. An attack, is like going for it. It has to be massive and overwhelming.
It is still tech war.
Mate 60 Pro means DUVi is in full scale production. And soon to be fully domestic. All foreign inputs may be jettisoned.
After all the hints about EUV research and improvements in China, now we get this new indication that what seems like prototype will be delivered shortly in the coming months (maybe they already delivered the test machine, we don't know).
This is the fastest we/they can go. We cannot go faster than this.
Makes perfect sense to announce DUVi and EUV together.
The former in production, the latter in prototype.
Please don't go China speed or else I'm losing my bet about EUV in 2025, while China on China speed will release EUV in 2024
There is a lot of discussions on how many 5G phones they can produce with domestic BAW filter.
Well, first of all, there is no way to know for sure that HW is not using BAW filter from Murata in its PAMiD. But let's say going forward, what can they actually make with domestic ones.
Well, Memsonic's owner 慈星股份 gave some hint today
Let's start with the basics
40 to 50k/ wafer and 70 to 100 RF filter per phone. Now I actually don't think you will end up using only BAW filters, but let's go with 70 here. And assume they way start off with lower yield so closer 40k/wafer, then you end up with around 600 phones of BAW filter per wafer
x 2000 wafers -> 1.2 million phones or 14.4m/yr
of course, if you end up only needing 50 BAW filter per phone (with remaining as SAW/TF-SAW), then that would be 900 phones of BAW filter per wafer (assuming 45k good filter per wafer)
x 2000 wafers -> 1.8 million phones or 21.6m/yr
if they ramp up to 10k/yr -> would be 70 to 100m/yr of phone they can make
of course, a lot of that would depend on how quickly they can get added to supply chain of domestic phone makers and L-PAMiD companies, but they could theoretically supply all of domestic market in a couple of years if they keep ramping up. Of course, I'm making the assumption that there will be more players than just Sai Micro & Memsonic here. Even if they end up being the largest player, they will like have < 50% domestic market share in this area
I also took a look at other sources.
This one is pretty good
Hefei 芯投微 (Corecat Micro)
started construction of plant to produce 6-inch SAW/TF-SAW/filter wafers
according to 旷达科技
That kind of gives you an idea of how many you can cut. 8-inch gives 1.8x the cutting areas as 6-inch. And maybe BAW are smaller than SAW.
But looks like 8-inch BAW -> around 40-50k filters per
6-inch SAW/TF-SAW -> around 20k filters per
There are quite a few SAW producers like Maxscned. So we can use this to figure out their production and such.
Remember, China currently needs 20B rf filter/yr. A number that's likely growing pretty fast.
Well, first of all, there is no way to know for sure that HW is not using BAW filter from Murata in its PAMiD. But let's say going forward, what can they actually make with domestic ones.
Well, Memsonic's owner 慈星股份 gave some hint today
All the questions are regarding how many phones you can make with it
Let's start with the basics
40 to 50k/ wafer and 70 to 100 RF filter per phone. Now I actually don't think you will end up using only BAW filters, but let's go with 70 here. And assume they way start off with lower yield so closer 40k/wafer, then you end up with around 600 phones of BAW filter per wafer
x 2000 wafers -> 1.2 million phones or 14.4m/yr
of course, if you end up only needing 50 BAW filter per phone (with remaining as SAW/TF-SAW), then that would be 900 phones of BAW filter per wafer (assuming 45k good filter per wafer)
x 2000 wafers -> 1.8 million phones or 21.6m/yr
if they ramp up to 10k/yr -> would be 70 to 100m/yr of phone they can make
of course, a lot of that would depend on how quickly they can get added to supply chain of domestic phone makers and L-PAMiD companies, but they could theoretically supply all of domestic market in a couple of years if they keep ramping up. Of course, I'm making the assumption that there will be more players than just Sai Micro & Memsonic here. Even if they end up being the largest player, they will like have < 50% domestic market share in this area
I also took a look at other sources.
This one is pretty good
Hefei 芯投微 (Corecat Micro)
started construction of plant to produce 6-inch SAW/TF-SAW/filter wafers
according to 旷达科技
they plan to produce 30k wpm and around 20k filters per wafer -> 600m/month -> 7B/yr
That kind of gives you an idea of how many you can cut. 8-inch gives 1.8x the cutting areas as 6-inch. And maybe BAW are smaller than SAW.
But looks like 8-inch BAW -> around 40-50k filters per
6-inch SAW/TF-SAW -> around 20k filters per
There are quite a few SAW producers like Maxscned. So we can use this to figure out their production and such.
Remember, China currently needs 20B rf filter/yr. A number that's likely growing pretty fast.
Last edited:
- Status