China vs. USA – False dawn.

There is more to EUV than a patent.

• There is a very big difference between filing a patent on the reflection of Extreme Ultraviolet light (EUV) and producing a working EUV machine that can be used to make chips in volume with good yields at 7nm and below.
• Despite this difference, the market seems to think that China is one step closer to circumventing the US restrictions on the export of semiconductor technologies to China by producing its own equipment.
• However, I think China is no closer to producing a working EUV scanner than it was 6 months ago and with every day that goes by, it slips further and further behind because it is frozen at 20nm.
• The area where Huawei has filed its patent for EUV is a very good example of just how difficult this is.
• Huawei’s patent refers to using interference patterns of EUV light in order to focus the beam in the right place and etch the tiny pattern required to create the circuit.
• EUV’s wavelength is so short that passing through glass will absorb the light which is why mirrors are used to direct the beam.
• However, what Huawei does not explain is how it made the mirrors to direct the light to a sufficient level of accuracy.
• These mirrors are so precisely manufactured that if one scales them up to the size of the earth there will be no imperfection larger than 10cm.
• This is a level of precision that almost nobody can achieve but this would imply that Huawei has already solved that problem.
• Precision manufacturing is something that Chinese companies have great difficulty with and I am not convinced that Huawei has suddenly solved this excruciatingly difficult problem.
• Also unexplained is where Huawei got the light source from.
• ASML’s unique light source is manufactured by its subsidiary Cymer in the USA and the 13.5nm light is created by bombarding tin particles with lasers until they become plasma at which point they emit the required light.
• The technical problems that were overcome to get this light source to perform consistently were legion in number and took many years to solve.
• If Huawei has managed to make or buy one, I suspect that it will be very unreliable and not suitable for use in a machine used to make semiconductors.
• Furthermore, it took ASML and Intel 30 years to perfect this technology and bring it to market which is a feat that China is unlikely to repeat any time soon.
• Hence, I think that this will have no bearing on the semiconductor restrictions that have been placed on China in the short r medium term meaning that the current predicament remains unchanged.
• This will lead to the Chinese semiconductor companies focusing on lagging edge at 28nm – 45nm where there is plenty of business to be done but it is very competitive.
• While this will be seen as not good news for China, I am optimistic with regard to an economic recovery.
• The CCP’s reputation has taken a dent from the Covid Zero policy, and I suspect that it intends to mend this by delivering a solid economic recovery that will benefit all Chinese citizens.
• This is why I think that the Chinese technology sector (especially the internet) could end up as the best-performing sub-sector of the technology industry in 2023.
• I am playing this through a position in Alibaba as well as a position in the state-owned banks that are still paying a 7% dividend yield while I wait for the recovery.