Chinese semiconductor thread II
- Thread starter vincent
- Start date
Fix isn’t hard at all. Just
at the start of the year, I had an article that said they sold 3 million chips.
And now, they say they've sold 6 million chips.
That's 3 million in 8 months.
Only a portion of that is for cockpit, since you need high end MCUs for Lidars and various control unit, suspensions and such. But cockpit SoCs likely still represent a good chunk.
Looks like they had 342k cockpit chips installed in first half of the year. now, installations come after sales, so their numbers should come up further by the end of the year.
But the big one to watch is BYD. When are they going to be using their own chips? BYD can replace all the suppliers in a year once it's comfortable with its own chips.
I think self-reliance is different from normal market competition.
Automotive chips are not a technological choke point ("stuck neck", as Chinese say), nor is a problem for Chinese fabs manufacturing them, they are almost all 28nm and above. Maybe only the ADAS chips need more advanced nodes.
If market is dominated by foreign firms, it is because local firms are still not competitive enough, but is a pure market driven thing: they will gain market share when performance/cost will be better for their customers. If for some absurd and impossible case US bans automotive chips and forces EU/Japan to do the same, in China it would be just cheers (干杯) all over the place....
Completely different the case of advanced SME tools, where self-sufficiency is a must and is needed to defend China chip sector from US blockades.
So regarding automotive I would not be obsessed with self-sufficiency, because in (the very unlikely) case of need China is already self-sufficient for those chips, instead I would just leave the market to do its thing....as in the old good days...
Right now Chinese foundries are operating at high rates close to capacity. For example SMIC and Hua Hong are operating at like 80% capacity. Most of these foundries have Chinese customers as well. So it is not like there is huge capacity just sitting there waiting for the market to take it up.
Automotive chips also have huge development cycles because of test and validation. So the sooner they start working on these the better.
This will change in four years after the capacity expansions are finished. Assuming the US does not interfere with foundry fab expansion plans again.
Automotive chips also have huge development cycles because of test and validation. So the sooner they start working on these the better.
This will change in four years after the capacity expansions are finished. Assuming the US does not interfere with foundry fab expansion plans again.
There is a ignore function. Just set tokenanalyst to ignore and you have your "macro" thread lol
Lithography is the critical equipment in Wafer fabrication but is totally useless with the rest of the tools, Ok, you bought the ASML NXT-2100, now you need coating-developing tools that can deposit photo resist with good uniformity, deposition tools for the different materials with enough uniformity and throughput, furnaces for oxidation and ion implanters for doping, RPT tools to annealing, Chemical Mechanical Polishers for wafers without damaging the made structures, etching tools that can handled the High Aspect Ratios of modern transistor, precision cleaning tools, metrology tools for yield monitoring, overlay metrology and CDSEM.
And then you have to take into account the materials like photo-resists, gases, liquids and their delivery subsystems, vibration isolation for the lithography machine. the lithography machines is useless without good photoresist, more useless without good overlay metrology and even more useless without good etching tools. And every single of those tools need their subsystems like RF-Matching Networks and maglev pumps.
Only thinking in the lithography machine is the trap that a lot pundits fall into.
And then you have to take into account the materials like photo-resists, gases, liquids and their delivery subsystems, vibration isolation for the lithography machine. the lithography machines is useless without good photoresist, more useless without good overlay metrology and even more useless without good etching tools. And every single of those tools need their subsystems like RF-Matching Networks and maglev pumps.
Only thinking in the lithography machine is the trap that a lot pundits fall into.
Last edited:
I agree in general, but today, in September 2024, litho is the only remaining weak link in the chain, because all the other pieces already fell into place.
This is a testament to the impressive, miraculous improvement of Chinese SME ecosystem.
Just 5 years passed since , when US fired the first big salvo on Huawei, a real Pearl Harbor moment for China. At the time SME ecosystem in China was scattered and filled with gaps. At the time the few local SME companies usually had subpar products kept at prototype/ small batch stage, due to very limited possibility of testing in real production. AMEC, NAURA where small players both in revenues and in product range and performances.
What happened in these 5 years will remain in history books!
Nobody in the West, first of all US administration, would have bet China was able to build a full local SME ecosystem as it is today. Only Japan did something similar but along the '80s and '90s of last centrury: 20 years against 5 of China!
Sure, some tool and material are still not at the advanced level, but they are anyhow ahead of litho and they will remain ahead of litho: when SMEE will be able to do 14nm the surrounding ecosystem will be ready, and so on for 7nm and lower...
So today lithographic machine is really the focus of the tech war, the last hurdle to overcome. Even US officials realized that: their full effort now is to stop ASML, they don't care anymore of the other types of machines / materials.
Pinejie Semiconductor (Hangzhou) Co., Ltd. completed hundreds of millions of yuan in financing
Recently, Pinejie Semiconductor (Hangzhou) Co., Ltd. (hereinafter referred to as "Pinejie Semiconductor"), a direct investment enterprise of the venture capital group, announced that it has completed financing of hundreds of millions of yuan. This round of financing was led by Shanghai Semiconductor Equipment Materials Industry Investment Management Co., Ltd. and followed by Nanjing Innovation Investment Group. The funds will be used for supply chain construction.
Founded in September 2018, Pinejie Semiconductor is a leading brand of China's third-generation semiconductor power devices. It mainly sells power device products such as silicon carbide MOSFET, silicon carbide SBD and gallium nitride HEMT. The company has the most complete silicon carbide power device product catalog in China. Silicon carbide MOSFET and silicon carbide SBD products cover various voltage levels and current carrying capacity, and have passed AEC-Q101 automotive-grade test certification. It can meet various application scenarios of customers and provide customers with stable and reliable automotive-grade silicon carbide power device products.