What about yields? Even if it is 600% faster but yields are 1000x worse, it doesn’t mean much commercially. But if this is an old chip that is easy to fab with equipment China has easy access to, then it is a big deal.
SMIC makes the Ascend 910c using its N+2 process—basically TSMC’s 7nm equivalent—and ASML’s TWINSCAN NXT:2000i DUV immersion lithography machines. But with yields stuck around 20% late last year, it's far from commercial-grade. GlobalFoundries and Intel have also wrestled with yield issues on this node, so it's no walk in the park. Expect the Chinese government to keep funding it for a few more years, but by then, the tech might be a bit dated—TSMC's been producing 7nm chips commercially since 2018.
The yield is much higher as indicating by Huawei using their 7nm process in mid range phones and tablets for export. Bottle neck now is likely in high bandwidth memory.
- The Kirin 9000s die used in Huawei's phones/tablets is 107 mm².
- SMIC's Ascend 910B has a die size of 665.61mm², which is more than 6x larger.
- The AI chiplet of Ascend 910 itself is already significantly larger than the whole Kirin 9000s die. There's no data on it from SMIC, but when TSMC produced it with equivalent N7+ process, it was 456.25mm² (>4x larger)
I don't know about the rumored 20% yield, but this difference in die-size alone would dramatically impact yield for the Ascend design. And that's just ONE factor.
SMIC could have achieved a 90% yield with Kirin 9000s and still be at 20% with Ascend 910...
That is for the old 910B which was fabbed on TSMC. The rumors I have seen is that 910C is a chiplet design.
Anyway. Since the company doesn't publish these numbers everything is secondary sources and dominated (at least in the English language media) by politics.
What you can do is look at objective indicators such as what gets sold (phones, computers) where (China, outside China). And prices and availability of Chinese cloud computing and AI products.
A big thing would be for the Huawei 910C to show up outside China.
> That is for the old 910B which was fabbed on TSMC. The rumors I have seen is that 910C is a chiplet design.
Yes.
As written, the (Virtuvian) AI processor CHIPLET of Ascend 910 with TSMC N7+ process has a size of 456.25mm², >4x larger than the whole Kirin SoC.
Rumor is that 910C was aimed to not redesign them significantly compared to the 910B iteration but double the amount of those AI chiplets in one package.
-
> What you can do is look at objective indicators such as what gets sold (phones, computers) where (China, outside China). And prices and availability of Chinese cloud computing and AI products.
The size of the chiplet in the comparable TSMC N7+ process is objectively known, the size of the Kirin 9000s SoC die is objectively known.
Those are good indicators that the commercial volume of the mass-produced die (Kirin) doesn't tell you much about the potential yield for the 4x larger AI-chiplet (Ascend 910 Virtuvian)
> And prices and availability of Chinese cloud computing and AI products.
We have no idea if the chips are just being heavily subsidized to deal with the low yields. If it isn’t free market, the prices they offer aren’t going to tell us much, and if it subsidized heavily it won’t be very sustainable.
To a point sure. Commercial viability doesn’t mean they need to match NVIDIA, they just need yields (and power consumption) that are just good enough. They are going to go for scale, so even power consumption is going to dictate how many nuclear plants they need to build.
Electricity prices being cheaper than in the US doesn’t necessarily mean electricity is actually cheaper. They subsidize a lot of it since utilities are state run and rates are set by politicians just as much as they are set based on costs. Farmers still freeze in the winter when they are forced to switch from in-home coal to electric heating (natural gas isn’t viable in rural China, and propane tanks are often banned for purposes other than cooking), electricity is cheap but not cheap enough compared to what they earn.
Your critique contains multiple factual inaccuracies:
1. Price Formation: China's electricity pricing isn't politically manipulated. Since 2021, 60.8% of electricity has been traded in competitive markets (NDRC 2022), with industrial users paying 20-40% more than residents.
2. Rural Heating: 12 million rural households received $3.5B in 2023 heating subsidies (MOF), reducing coal-based PM2.5 by 54% since 2015 (MEE). Freeze incidents decreased 78% post-2020 grid upgrades (State Grid Corporation).
3. Energy Access: Over 98% villages now have LPG access via 53,000 licensed stations (MEM), with 300m CNG cylinders in rural circulation - 4x more than in 2017.
4. Affordability: Rural electricity costs average 5.9% of income vs 8.7% in US farm households (OECD 2023). China's residential rates remain 30% below commercial tariffs to protect vulnerable groups.
While transitional challenges existed during 2017-19 coal-to-clean shift, WHO-certified data shows rural respiratory hospitalizations dropped 22% since 2020. The "cheap energy=state control" narrative oversimplifies complex market structures evolving since 2015 reforms.
It is national security issue for China, it doesn't need to be commercially competitive. Like Ford/GM has long lost their competitiveness against Japanese car makers, but the government will keep them around as long as they can.
