SMIC isn't close to the same node as TSMC. A competitive high-end SOC could not be manufacured by them.
Yet I think it's important to keep some perspective: this Chinese/US spat is not a championship where a winner will be declared at the end of the summer. There's no reason to believe that 50 years from now either China or US will have collapsed. Xi Jinping dictatorial inclinations [1] are a threat of course, as he might decide to consolidate his power at the expense of China. But US recent [2] trend of politicians running on the hate of the other half of the US population is also a tremendous existential threat.
In any case, assuming that both US and China still stand 50 years from now, the only way for TSMC, Samsung or even Intel to stay ahead is, well, to stay ahead. It's a never-ending marathon, and in that regard funneling Huawei money into SMIC sounds very counter-productive.
[1] Moreso than his predecessors IMHO, with the notable exception of Mao. And indeed, Mao was very good at keeping China poor, underdeveloped and starving.
[2] Is it really recent? I'm French, during my two stays in the US (first time 3 years, second time 5 years) I felt a major polarization of the population, but it might have been there all along and I simply didn't notice.
I think this doesn't paint an accurate picture. TSMC were "high end" sans Intel. Since you cant Fab with Intel, TSMC was the best any Fabless Vendor could get, now it just happen TSMC is also the best in industry.
Compared to SMIC, they aren't even close to TSMC, nor Samsung, or even Global Foundry. In Tech lead, Volume and yield. To put things into perspective, they will have to accelerate at 2x the speed of current industry's rate for the next 5 years in other to be level with TSMC.
SMIC's 14nm is still no where near Samsung, or even Global Foundry ( GF ). And GF has been shipping 14nm in volume for nearly 5 years now.
When Intel refused to make Apple's ARM processor for iPhones Apple went to Samsung didn't they?
Now Huawei cash is going to finance SMIC. It's cash being funneled at a competitor, supported by the Chinese government. Which means that if Huawei is not competitive because of SMIC, SMIC is going to get further money/intelligence from the Chinese government to become competitive. I don't see how TSMC, Samsung or Intel win from that situation, at least in the long term.
I can be completely wrong of course, and maybe China will be in a similar situation to that of the Soviet Union during the cold war: without access to competitive electronics. But SMIC is not that far behind, or are they? They can do 14nm, so they are at TSMC level of 2016. 4 years late? Without ASML support though, it might be harder to catch up.
I think TSMC having to drop Huawei is more a side effect of American sanctions than anything being done by design. Again I can be wrong. US sanctions are very far reaching and problematic for a lot of companies (see: Airbus, BNP Paribas, etc.), it doesn't mean that these companies in particular are targeted by the US. Unintended consequences.
Meanwhile, they have 14nm.
There's enough public knowledge about EUV to clone the machines if starting with unlimited financial resources and no legal constraints.
If it happens in 5 years what will ASML do then ? The west’s monopoly on this sector would be finished.
At this point China has no choice but to develop the tech or be forced to not have processors. Time will tell.
Given that hard power is being used to punish China they will use every means at their disposal from massive investments and manpower to poaching talents and even espionage.
If North Korea can make nukes then this shouldn’t be impossible for a power like China.
I have a sneaking suspicion that they are much further ahead than they are revealing in public which explains their confidence and refusal to back down. Time will tell.
Espionage is usually the first thing, not something of last resort. Pretty much all major players are doing it all the time. Doesn't even take much incentive.
The west isn't quite as free with giving the stolen information to the private sector (though the US really likes giving their domestic companies bidding advantages) - but since we're talking about China here, there's pretty much a guarantee of anything and everything finding a way to their private sector.
The Soviet Union didn't collapse because a lack of capability. They collapsed because it was too expensive to forge their own way, when most of the world could pool capital via free trade.
China is big. But the rest of the world is bigger. And despotic regimes inclined to trade with an embargoed state aren't typically the most flush economic partners.
SMIC which is their Intel raised 7 billion dollars last week from investors inside China. That’s on top of what the government is putting in. People underestimate the resources at chinas disposal.
It's benefited from 20 years of optimistic free market access.
No doubt they'll find trading partners, but it's not going to look anything like the open doors they've been used to. And they're going to start finding that their domestic and international choices now have trade consequences, rather than the look-the-other-way, laissez-faire approach the world has taken thus far.
If one thing the EU should worry about not turning into a Hotel Europe where people only go to visit instead of producing stuff.
I don't see why EUV would be impossible for them to replicate. The list of times when we (and I too) thought that they could not catch up in a certain technology is long, and I won't make that bet anymore.
The resources of the Chinese state are incredibly vast. If they decide that they absolutely must get something done, that will most likely happen. It would be a Manhattan Project level initiative, but even more important.
One of the best things I feel like I've learned in systems eng is that this stuff is just too disorganized and difficult. It's not enough to steal all of the plans - you're not going to make sense of them and they don't have the rejiggered edition Jimmy's WAR from last Tuesday.
IP theft occurs because employees are owned for years at a time and are able to slowly push useful tidbits across the proprietary barrier.
if you no longer care about pissing off the US and EU (because they're up in your business anyway) then maybe it doesn't matter so much if you violate those patents.
Same thing as with Taiwan's "special trade status", Trump thinks that revoking their status will get China to leave them alone, because he has a very simplistic bullying-oriented view of international politics, but it's widely predicted to have the exact opposite effect as intended and lock in PRC control. Or trying to bully Iran with further sanctions even though they're already in compliance with treaties.
The approach is "carrot and stick", if you only use the stick then you provide no incentive for future compliance, people realize it's more effective to just start preparing for the inevitable future beatings and set yourself up to mitigate as much of the pain as possible.
This often leads to countries building up their own domestic industry to make themselves impervious to sanctions - that's how it's worked with Iran, after 50 years of sanctions they are more independent and less susceptible to the effects of sanctions than ever. That's how it's working with China, this is far from the first thing we've tried to choke off, there was the spat with memory too and they just ended up building their own plants.
And it often leads to further "escalation" and violation of norms. If you're going to get shit on anyway, why would they respect ASML's intellectual property?
If I were Pakistan and in desperate need of financial assistance, I would look to China instead of USA, simply because of the former's consistency (despite all the strings attached to these investments).
Predictability is highly underrated, in international politics and elsewhere.
and now you know why China is investing in Pakistan, it is not because they are good hearted wanting to help Pakistan
It is destabilize the Region by putting a wedge between the US and Pakistan,
The US keeps Pakistan and India stable, if the US loses it's influence over Pakistan well that will be very bad for the region.
Not sure that is really an advantage or defense. One would hope more than just "reliable" would go into a complete foreign policy
Have I missed major developments, or have you confused Hong Kong and Taiwan?
If the American plan fails and the Chinese are able to produce high quality chips without ARM IP and without relying on TSMC. Then this would have been a strategic blunder of historic proportions.
In reality ASML doesn't need that export restriction to stop shipping to China. Their EUV TwinScan production capacity is behind schedule, with or without the pandemic. And with factory closing and now not working in 100% capacity the delay will be even longer.
The next two years order are already booked by TSMC and Samsung Foundry, and looking at Intel's roadmap I would not be surprised if 2023 is also filled. Samsung are looking at EUV for their NAND and DRAM.
All ASML has to do is to reply SMIC they are all sold out. Which itself isn't a lie and also an Industry standard excuses for not shipping to a competitor.