China vs. USA – The Screw pt. V

China does not have much of an answer.

• The outgoing administration will leave its mark with a further significant tightening of export rules and with the moribund state of its technology sector, China has very little with which to answer.
• The Bureau of Industry and Security (BIS) has issued docket 241126-0303 (see here) which makes several changes to what may and may not be exported to China.
  • First, Foreign Direct Product Rule (FDPR): which is being applied in conjunction with new restrictions on 24 types of previously unrestricted manufacturing tools.
  • The FDPR is a rarely used but potent tool of foreign policy which allows the US to restrict any equipment made anywhere in the world if it has any US-made technology in it.
  • It was the threat of this rule that forced TSMC to stop selling leading-edge wafer manufacturing services to Huawei some years ago.
  • Japan and The Netherlands which already have their restrictive regimes have secured an exemption meaning that this is really targeted elsewhere
  • A number of companies (including the US) have been making more machines outside of the USA to avoid the restrictions and it is this activity that is the real target of this amendment.
  • This will close this loophole and make it increasingly difficult for China to secure equipment capable of manufacturing chips at 20nm and below.
  • Second, Entity list: which is a list of companies to which a US-based company needs to obtain a licence to export their products.
  • 140 new companies have been added to this list most of which are in China, but some are also in Japan, Singapore and South Korea and are presumably suspected of helping Chinese companies to circumvent the restrictions.
• These changes appear to be relatively simple, but their scope is wide-ranging and quite unspecific which gives the US a lot of latitude in terms of how it wants to interpret the rules.
• A good example of this is high bandwidth memory (HBM) which is not referred to explicitly in any of the published materials and has raised a lot of questions.
• CXMT is a Chinese company trying to develop HBM to support the training of Chinese AI but has not been added to the entity list.
• I suspect that the view of the administration is that other restrictions such as the limitations on equipment below 20nm will prevent CXMT from making competitive HBM, but the US has little interest in preventing CXMT from trying as long as it is confident that it will fail.
• Given that all HBM is made on nodes way below 20nm, I think it extremely unlikely that CXMT will succeed in making HBM that can challenge Western versions or meaningfully help advance the development of Chinese AI.
• However, preventing CXMT from trying could hurt sales of US companies which the administration will be keen to avoid where it has no benefit.
• The lack of specifics on the docket appears to have caused a lot of confusion but I think that this gives the USA the ability to use these new rules with much greater latitude meaning that there are fewer loopholes.
• China has responded by restricting the sale of Gallium, Germanium, and Antimony which are used in semiconductor manufacturing as well as in batteries, solar panels and so on.
• This is one of the “Three Rs” that RFM has identified as the actions that China can take to fight back, but it is unlikely to have much long-term effect.
• This is because while China refines almost all of these materials, it does not control the reserves in the ground.
• This means a price spike from restrictions will incentivise countries sitting on their reserves to start (or in many cases restart) production.
• Hence, in the long term, this strategy reduces the world’s dependence on China for these materials and will hurt Chinese companies that currently refine and export these materials.
• This combined with the moribund state of the Chinese technology sector means that it remains unlikely that China is close to inventing a must-have technology that will give it the leverage it needs to push back against these restrictions.
• Hence, I continue to think that these restrictions are broadly effective and with the latitude that the new rules confer, the US will have an easier time of enforcement.
• 2025 looks like it is going to be another difficult year for China with its own economic problems deepening and more effective restrictions.
• This combined with China’s priority of control over free enterprise means that the Chinese technology sector will remain hobbled both in terms of its ability to develop technology as well as attract investment.
• This will accelerate the decoupling which is now well underway and divide the technology sector into two distinct and incompatible pieces.
• The Balkanisation of the global network inevitably means less growth for the technology sector in the long term.