sarker
It isn't happening, and if it is, it's a bad thing
No bio...
User ID: 636
sarker
It isn't happening, and if it is, it's a bad thing
5hr ago
After domestic chips become better than H200 (or rather, domestic systems + power subsidy become competitive with Nvidia clusters) and there's wide adoption of Huawei CANN and Cambricon NeuWare, I predict that they will relax controls on imported chips, maybe replacing it with a simple tariff. Your model suggests that they will tighten controls. It's an empirical matter, we will see in a matter of 2 years, most likely.
Hardly. If Chinese chips become better than the best Nvidia has to offer, there is, as you say, little reason to ban imports since approximately nobody will want to use them. EVs are a good example - how many Chinese people buy American EVs? I'd assume the number is approximately zero since American EVs are about 5x the price for a comparable product.
Of course, if Chinese chips are better than Nvidia's best, the prospects for Nvidia market share become quite slim. You seem to want it both ways - Nvidia exports to China will ensure Nvidia's dominance in China (lock in! Efficiency! Yields!), but also China is going to ensure that domestic chips are the market leader.
sarker
It isn't happening, and if it is, it's a bad thing
5hr ago
Sure, let's throw that in to the consumption number.
That brings us to 46% for China and 65% for the US based on the numbers above, once we apply the increase from the text you quoted. Still, the gap is fairly significant.
sarker
It isn't happening, and if it is, it's a bad thing
6hr ago
How is it not supposed to help? Money is good, market share is good.
We were just discussing long vs short term thinking. Money and market share are the short term returns, the question is about the long term.
In a more relaxed geopolitical environment, they may not even stop 100%, just like they're still buying high-end Western CNC machines, despite being able to make functionally similar ones now. Nvidia will still have better yields, lower unit prices, higher power efficiency, more polished software, likely for decades. It's not like they have a philosophical commitment to not buy at all, they are simply focused on becoming strategically invulnerable to export controls after years of American gloating about how they'll fall behind and die, starting with Huawei.
We're not in a more relaxed geopolitical environment. We're in an environment where China is doing everything it can to stimulate domestic chip production including by banning imports of certain chips in the first place, which sounds like a philosophical commitment to not buying at all to me.
If you believe that China is interested in being invulnerable to export controls and you believe that Nvidia use can lead to lock-in, it follows that China's strategy would be to avoid the lock-in in the first place, which means that there is no long term market for Nvidia chips in China.
The main argument against this has nothing to do with Chinese plans to stop buying chips later
Correct. I brought up the other argument in a previous thread where you responded with a non-sequitur about orange man having bad trade policy. Here I'm limiting myself strictly to your claim that chip exports would be beneficial in the long run for Nvidia.
sarker
It isn't happening, and if it is, it's a bad thing
6hr ago
This feels like a misdirection. The price level of China vs the US doesn't matter for the question of how much of Chinese GDP is household consumption. In each case the ratio can be calculated in local currency without any need for PPP adjustments.
The article you linked is (apparently, based on your excerpts) discussing correct PPP factors based on household expenditures, which is really not the same question at all.
sarker
It isn't happening, and if it is, it's a bad thing
9hr ago
I still don't understand how exporting chips to China is supposed to help the US in the long run when the Chinese long run plan is to not import chips.
sarker
It isn't happening, and if it is, it's a bad thing
9hr ago
Chinese consumption is not that low, read this.
It's paywalled. What does PPP have to do with the fraction of Chinese household income that goes to consumption?
I don't even have a "problem" with their system in the context of this conversation, and it seems a bit facile to resort to accusations like that when we're discussing a narrow empirical question about Chinese consumption as a fraction of GDP which you were the one to dispute.
@aquota said it's 40%, you said that's bullshit propaganda and provided a paywalled source to support that which, it turns out, suggests the true number is 46%.
Now you're saying that 40%-50% is perfectly normal, actually. Well, maybe, but then why call the original claim imperial propaganda when it's pretty close to correct?
More options
Context Copy link