36 comments

[ 3.4 ms ] story [ 88.5 ms ] thread
If war happens in few years, I really wonder how markets would react to this? Algos might go crazy. It would probably create chip supply chain problems bigger than covid.
[flagged]
People who think it as merge are undoubtedly mostly Chinese.

> most of them have good relationship with China

I can enumerate more than half of the neighbours of China who has territorial disputes with it.

Besides South Korea and Japan, do India, Vietnam, Bhutan, and Philippines have good relationships with China? They all have significant disputes over fundamental territorial issues.
oh yeah, that's why china is doing war games around Taiwan in which it simulated attacks with bombers and boarding ships right now. Very peaceful and optimistic indeed. And somehow americans are the ones warmongering.
Brother, China only just moving ships

The US literally is arming Israel to commit a genocide in Gaza

If you want to play the super hero, start by not being a coward

And they are doing the same with Taiwan, South Korea, Philippine and Japan

Blood on your hands, you choose!

China is also arming Russia.
Wrong, you were not focus enough during the meeting:

https://www.bbc.com/news/60571253

China is not providing Russia with weapons like China is not providing the USA with illicit fentanyl. Which is to say, they definitely are, but have officially said they aren't.
(comment deleted)
Invading countries usually refer to it as a happy merger
This is one of the reasons why the USA is both increasing domestic chip production and boosting Taiwanese defenses.
Won’t be anywhere near fast enough or big enough on either count.
> The switch applies to the latest EUV machines out of ASML.

How long before China is able to make its own EUV machines?

This is discussed in the book Chip War at length. I believe the answer was that if it were possible at all, it's going to be at least half a decade behind. There were a lot of other supply chain considerations which would likely apply as well.
a lot earlier than ppl think, and there are a lot of room to innovate on transistor structure as well even using existing litho machines
They can make 7 nm chips in quantity. And one or even 2 years ago there were fotos of a few exemplars of China made 5 nm chips floating around. Concerning smaller structures, I have the gut feeling that they instead want to go for optical chips.
From the articles I read, they are making 7 nm on old dutch (AMSL?) equipment and their yields aren't great (they get quantity by just throwing money at the yield problem).

> The chip was fabricated by China’s Semiconductor Manufacturing International Corporation, or SMIC, a silicon foundry. (The Commerce Department says SMIC makes chips for the Chinese military, a charge the company denies.) SMIC produces the Kirin 9000s using Deep Ultraviolet lithography equipment acquired before the latest American sanctions. Huawei’s claim to have produced 7nm circuits was verified by the Canadian firm TechInsights and by Tilly Zhang of Gavekal Dragonomics.

> To begin, the DUV process “success rate”—that is, the percentage of chips coming off the manufacturing line that actually work as designed—may or may not be up to the industry norm. Neither SMIC nor Huawei are releasing much data about the new chip or its production process. As Gavekal’s Zhang has noted, SMIC may have pushed the DUV process as far as it will go, reaping only 40 percent or fewer usable chips. Huawei's competitors, such as Taiwan Semiconductor Manufacturing Co., have been producing 7nm chips for over five years; their success rate generally tops 90 percent.

(https://www.defenseone.com/ideas/2023/11/size-isnt-everythin...)

It isn't commercially viable yet, but it can be subsidized. They do talk about using their own lithography equipment soon.

Thanks for the info. Butare you sure they couldn't or cannot ramp up their yield? They bought a lot of TMSC knowledge in the last years.
It’s the lithography equipment they are using, it isn’t designed for 7nm and the only reason we guess their yields are bad is because lots of people have tried already. TSMC also buys from the same Dutch company, the knowledge/equipment to build commercially viable chips is incredibly distributed: if China took all of TSMC’s factories and people in tact, they would still be unable to do that well because the supply chain is intentionally distributed through multiple US allied countries.

China needs to do everything, they will need to produce and master their own lithography and so on.

Not convinced. E.g. the DUV machines for sure are build to allow flexible multiple patterning. Some progress there may ramp up the yield significantly.
Surprised to see a picture of a High NA EUV machine. Would have thought even that would be guard even if one can’t really see the interesting bits
CIA doesn't trust this, will use military to blow up TSMC in the event the Chinese invade. But the Chinese will just keep slowly taking over the US government, making this policy of questionable use. LOL.
> CIA doesn't trust this, will use military to blow up TSMC in the event the Chinese invade.

So, it's less a 'Kill Switch' against China and more of a 'Please Don't Kill Us Switch' against US?

It's hard to find a machine that does not have a remote control nowadays. Also I'm sure there are tons of RSA key activated software running on that machine.

But the real kill switch would be people who know how to deal with such level of complexity.

3 gorges dam has a similar kill switch..
> TSMC Has a Kill Switch in the event of a Chinese invasion

Did they tested it ? /s