China vs USA – Problematic signals

SMIC’s expansion highlights China’s difficulty in semiconductors.

• Semiconductor Manufacturing International Corporation (SMIC) is expanding its capacity, but its plans betray its limitations meaning that its 10nm aspirations are probably vapourware and that it will never compete in the big leagues.
• SMIC has announced that it will spend $8.9bn on a new fab in Shanghai that will be able to produce 140,000 wafers a month but it will be 2-3 years before production begins.
• Most tellingly of all, the Fab will be manufacturing at the 28nm node where the technology is a decade old and is already far out on the lagging edge.
• The reason for this is clearly that SMIC is unable to purchase the equipment that it needs to make semiconductors at more advanced nodes due to the embargo placed on it by the USA government.
• Consequently, this means by the time the new fab is able to make semiconductors, the current shortage will have long been alleviated and SMIC will be selling its fabrication services into a well-supplied, highly commoditised market at 28nm.
• This means that the financial returns earned on the $8.9bn invested are likely to be far lower than they would have been had the fab been ready to go today.
• I think it likely that SMIC has used today’s trading environment or something close to it, to forecast the financial performance of this fab and as a result, I think it will be sorely disappointed.
• I suspect that this might be the reason why SMIC’s chairman (Zhou Zixue) resigned a couple of days ago instead of the health reasons that were cited.
• By contrast, its competitors are currently manufacturing at 5nm (Intel 4) and will be moving onto 3nm (Intel 20A) by the time this facility is up and running.
• This goes to underline just how serious the disadvantage that China has when it comes to semiconductors.
• Furthermore, this disadvantage is unlikely to be alleviated until there is some kind of rapprochement between the two sides (extremely unlikely) or integrated circuits switch away from silicon to something else.
• Both of these possibilities are years away in the best instance meaning that the semiconductor problem for China is going to persist leaving its Made in China 2025 strategy in tatters.
• Consequently, in the crucial sector of semiconductor manufacturing, I think that China has no hope of gaining independence.
• However, in other areas such as AI, autonomous driving, and quantum computing, it is a completely different story.
• RFM research has found that in these areas, China is at the very least a major challenger and may even have a better offering in some cases.
• This means that the leverage that the USA has over China has an expiry date after which the competition will be on a much more level playing field.
• However, I think that this will require the Chinese government to stop interfering heavily in its technology sector and the economy more widely because at the moment it is doing more damage to its own aspirations than the USA can ever achieve on its own.
• It would be ironic if the long technology rivalry that RFM research predicts was settled in the USA’s favour with the USA merely watching from the sidelines.