Chinese semiconductor industry

Status
Not open for further replies.
D

Deleted member 24525

Guest
Slightly off topic, but does anyone know how Huawei has been able to maintain their base station business after being cut off from tsmc? As of 2020 their base station chips were at 7nm, and as of last year so were those of xiaomi and zte, both of which are buying from tsmc. It's possible Huawei is being supplied by smic but I'm skeptical they have the spare capacity. I've seen rumors that they're making due with advanced packaging but nothing concrete.
Mainly I'm just concerned about China's vulnerability to cutoff from 5g base station chips, since that would have pretty immediate implications for the development of smart manufacturing. This is something I've been concerned about for a while, but I haven't been able to find any meaningful details on how feasible it is to use older nodes for that purpose or how intense the power consumption drawbacks are from doing so. How Huawei has adapted may help answer that.
 
H

huemens

Junior Member
Registered Member
Boltzmann said:
Slightly off topic, but does anyone know how Huawei has been able to maintain their base station business after being cut off from tsmc? As of 2020 their base station chips were at 7nm, and as of last year so were those of xiaomi and zte, both of which are buying from tsmc. It's possible Huawei is being supplied by smic but I'm skeptical they have the spare capacity. I've seen rumors that they're making due with advanced packaging but nothing concrete.
Mainly I'm just concerned about China's vulnerability to cutoff from 5g base station chips, since that would have pretty immediate implications for the development of smart manufacturing. This is something I've been concerned about for a while, but I haven't been able to find any meaningful details on how feasible it is to use older nodes for that purpose or how intense the power consumption drawbacks are from doing so. How Huawei has adapted may help answer that.
Click to expand...

I think two possibilities are:
1) Using existing stockpiles. Base stations are not as numerous as consumer devices. So their stockpiles of base station chips are probably going to last longer than smartphone chips.

2) Using chips made with older processes. Base stations can afford to use less power efficient chips than smartphones because base stations are usually connected to grid power rather than tiny batteries like smartphones.
 
tphuang

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
Boltzmann said:
Slightly off topic, but does anyone know how Huawei has been able to maintain their base station business after being cut off from tsmc? As of 2020 their base station chips were at 7nm, and as of last year so were those of xiaomi and zte, both of which are buying from tsmc. It's possible Huawei is being supplied by smic but I'm skeptical they have the spare capacity. I've seen rumors that they're making due with advanced packaging but nothing concrete.
Mainly I'm just concerned about China's vulnerability to cutoff from 5g base station chips, since that would have pretty immediate implications for the development of smart manufacturing. This is something I've been concerned about for a while, but I haven't been able to find any meaningful details on how feasible it is to use older nodes for that purpose or how intense the power consumption drawbacks are from doing so. How Huawei has adapted may help answer that.
Click to expand...
i posted about it on this thread, you can look back on it, but most likely SMIC is supplying them
 
horse

horse

Colonel
Registered Member
Boltzmann said:
Slightly off topic, but does anyone know how Huawei has been able to maintain their base station business after being cut off from tsmc? As of 2020 their base station chips were at 7nm, and as of last year so were those of xiaomi and zte, both of which are buying from tsmc. It's possible Huawei is being supplied by smic but I'm skeptical they have the spare capacity. I've seen rumors that they're making due with advanced packaging but nothing concrete.
Mainly I'm just concerned about China's vulnerability to cutoff from 5g base station chips, since that would have pretty immediate implications for the development of smart manufacturing. This is something I've been concerned about for a while, but I haven't been able to find any meaningful details on how feasible it is to use older nodes for that purpose or how intense the power consumption drawbacks are from doing so. How Huawei has adapted may help answer that.
Click to expand...

It is the stockpile.

That is what I read about the 7nm chip, that there are two of them inside a basestation.

Just bring out a napkin, and we start writing on the back of it.

Before the bans went into effect cutting TSMC from Huawei, Huawei used to sell 250 million phones a year at its peak.

Those phones would need advanced chips.

Huawei would have been ordering at least 50 million per quarter from TSMC.

As we have heard, all of China was stockpiling chips. Huawei may have started their stockpile when ZTE got banned.

Let's say that there are two 7nm chips inside a Huawei 5G basestation, and the countries needs 3 million basestations. That is a high estimate, apparently they already install their 5G standalone network, and that was maybe 2 million basestations at most.

Let's say it will take 6 million 7nm to be able to build out the standalone 5G network. But let's take no chances, and go with a 10 million 7nm chip stockpile.

Then Huawei could stockpile about 10 million 7nm chip for basestations sold abroad.

So let's keep pushing up this estimate. Suppose 20 million 7nm chips is not enough, okay goto 30 million 7nm chips as the stockpile.

Huawei used to sell 250 phones a years, chips all from TSMC.

Now, we should remember, after they were cut off, Huawei was still able to sell a lot phones in that year, before finally running out of chips for phones.

So to make a long story short, realistically, after the ZTE ban, Huawei probably had a year and a half to stockpile any chip they wanted.

After the ban was announced, the grace period allowed TSMC to supply Huawei with chips to add to the stockpile.

If they needed 30 million 7nm chips for the basestation, and they were selling 250 phones per year, and there was a grace period before the ban would be implemented, well it is pretty obvious Huawei will not encountered any problems with their basestation chip, with their stockpile. If you're going to war, you need the ammo.

The problem with the Americans, is that they cannot count.

If we ask some American person, like Blinken or Sullivan or Harris, how many 7nm chips would the US need to build out the standalone 5G network inside the basestation for the continental United States? Those three stooges would never know.

Same thing here.

:D
 
horse

horse

Colonel
Registered Member
The Russians, even today, are still using the Ho Chi Minh Trail to obtain their chips. Figuratively speaking.

Then the Russians would build their rockets and missiles with those chips, and blast the Ukrainians, and foreign mercenaries.

The double whammy. They were able to obtain your technology. Then they kill your mercenaries with your technology.

Did Huawei actively seek chips made outside of China, and try to smuggle some for themselves?

Seems like the answer to that, is a resounding no.

That does not prove anything in particular.

It only should indicate that they are thinking not about guerilla war and the Ho Chi Minh Trail, geez, Ren Zhengfei is PLA you know, as they had other ideas in mind.

But, don't say that to the Americans like Moe Larry and Curly. Their narratives are suppose to be like religion, and sacrosanct.

:D
 
H

Hitomi

Junior Member
Registered Member
huemens said:
I think two possibilities are:
1) Using existing stockpiles. Base stations are not as numerous as consumer devices. So their stockpiles of base station chips are probably going to last longer than smartphone chips.

2) Using chips made with older processes. Base stations can afford to use less power efficient chips than smartphones because base stations are usually connected to grid power rather than tiny batteries like smartphones.
Click to expand...
I would like to add that the last minute surge purchases from Huawei even included unpackaged chips(that Huawei could complete later while leveraging China's strength in the chip packaging segment) to speed up the deliveries and lead to a much larger than normal stockpile.

I also read that they modified the base station designs to reduce the usage of the 7nm chips.

I would also imagine Huawei changing their maintenance SOPs to minimize chip swapping as much as possible such as ensuring the chips are truly unusable before dumping them.

Who knows maybe TSMC had some "defective" wafers that were etched with patterns that coincidentally matched those required by Huawei and the wafers were coincidentally sent to a recycling company that shipped it to the mainland and "recycled" them.
 
A

ansy1968

Brigadier
Registered Member
Hitomi said:
I would like to add that the last minute surge purchases from Huawei even included unpackaged chips(that Huawei could complete later while leveraging China's strength in the chip packaging segment) to speed up the deliveries and lead to a much larger than normal stockpile.

I also read that they modified the base station designs to reduce the usage of the 7nm chips.

I would also imagine Huawei changing their maintenance SOPs to minimize chip swapping as much as possible such as ensuring the chips are truly unusable before dumping them.

Who knows maybe TSMC had some "defective" wafers that were etched with patterns that coincidentally matched those required by Huawei and the wafers were coincidentally sent to a recycling company that shipped it to the mainland and "recycled" them.
Click to expand...
Or maybe buy an unfinished 7nm chips and Huawei do the packaging. ;)
 
T

tokenanalyst

Brigadier
Registered Member
Boltzmann said:
What is the technological significance of sensing and detection in the lithography process?

Also, the caption is kind of odd. The institute seems to describe itself both as an integrator of various lithography subsystems and also a specialist in sensing and detection. That seems like an idiosyncratic mix to me but I'm not very knowledgeable on the subject.

Still, it's definitely good to see dedicated institutions set up for something obviously EUV related.
Click to expand...
Look like they are focusing in the highly precise instrumentation, sensors and how to integrate those in lithography machines.
 
tphuang

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
horse said:
It is the stockpile.

That is what I read about the 7nm chip, that there are two of them inside a basestation.

Just bring out a napkin, and we start writing on the back of it.

Before the bans went into effect cutting TSMC from Huawei, Huawei used to sell 250 million phones a year at its peak.

Those phones would need advanced chips.

Huawei would have been ordering at least 50 million per quarter from TSMC.

As we have heard, all of China was stockpiling chips. Huawei may have started their stockpile when ZTE got banned.

Let's say that there are two 7nm chips inside a Huawei 5G basestation, and the countries needs 3 million basestations. That is a high estimate, apparently they already install their 5G standalone network, and that was maybe 2 million basestations at most.

Let's say it will take 6 million 7nm to be able to build out the standalone 5G network. But let's take no chances, and go with a 10 million 7nm chip stockpile.

Then Huawei could stockpile about 10 million 7nm chip for basestations sold abroad.

So let's keep pushing up this estimate. Suppose 20 million 7nm chips is not enough, okay goto 30 million 7nm chips as the stockpile.

Huawei used to sell 250 phones a years, chips all from TSMC.

Now, we should remember, after they were cut off, Huawei was still able to sell a lot phones in that year, before finally running out of chips for phones.

So to make a long story short, realistically, after the ZTE ban, Huawei probably had a year and a half to stockpile any chip they wanted.

After the ban was announced, the grace period allowed TSMC to supply Huawei with chips to add to the stockpile.

If they needed 30 million 7nm chips for the basestation, and they were selling 250 phones per year, and there was a grace period before the ban would be implemented, well it is pretty obvious Huawei will not encountered any problems with their basestation chip, with their stockpile. If you're going to war, you need the ammo.

The problem with the Americans, is that they cannot count.

If we ask some American person, like Blinken or Sullivan or Harris, how many 7nm chips would the US need to build out the standalone 5G network inside the basestation for the continental United States? Those three stooges would never know.

Same thing here.

:D
Click to expand...
hmm, the stockpile is pretty low at this point. Kirin chips ran out 2 years ago. Kunpeng and Ascend chips were pretty old and not competitive anymore, so that's why they had to new kunpeng and Ascend project (at least Hisilicon guy said they did). 5G base station chip project is one of the projects that the hisilicon guy said was on their list. May very well be their next project after Ascend and Kunpeng. Keep in mind that there is no major space constraint with base station, you can probably get 7nm performance level through 12nm with advanced packaging. The only one with their space and power consumption constraints is smartphone. So, that might be the last project they complete.
 
Status
Not open for further replies.
Top