USA vs. China – No loophole

Nanoimprint lithography unlikely to solve anyone’s problems.

• Canon has announced that it will begin selling an alternative technology for advanced semiconductors that I suspect is unlikely to see a lot of traction in the market or help China circumvent the restrictions that have been placed upon it.
• After many years of development, Canon is ready to start selling a lithography machine that it says can compete with ASML’s EUV allowing semiconductor makers to go below 7nm without using EUV for the first time.
• Nanoimprint lithography is not a new idea and has been viewed as a potential alternative for many years although it has never managed to win any real traction in the market.
• Instead of printing circuits onto wafers with light, nanoimprint lithography uses a physical mask that imprints the pattern onto the wafer in order to make the circuits.
• This sounds great in practice, but no one has really managed to get it to work at high enough volumes and high enough yields to make it a commercial competitor to ASML.
• Presumably, Canon thinks that it has managed to solve these problems to a degree that it can mount a credible challenge.
• However, I suspect that the current environment around leading-edge semiconductor equipment has also been a factor in the decision to launch now.
• This is because demand for leading-edge semiconductors is very strong at the moment thanks to both geopolitical tensions with China and the seemingly endless demand for AI.
• ASML is currently enjoying lead times of 18 to 24 months on its equipment meaning that there will be a higher-than-usual willingness to look at alternative technologies.
• The fact that China is blocked from accessing EUV technology may also have created an incentive for Canon to commercialise an alternative that may be exportable to China.
• This might have worked under the old Japanese export regime, but the recent changes are likely to make exports to China even if the technology can compete extremely difficult.
• In 2022, Japan’s semiconductor equipment export regime only required licences to be obtained to export equipment to China but this has been changed such that all exports now require a licence.
• The regime has changed from presumption of denial to presumption of grant, and it will be extremely simple for Japan to deny any exports it likes to China with no further change of the rules.
• Furthermore, the US is looking to again tighten its export controls to make it even more difficult for China to circumvent the restrictions that are already in place.
• This will include actions targeted at exports to China via other countries and tighter restrictions on exports of AI training chips.
• The net result is that accessing advanced manufacturing in semiconductors is likely to get even more difficult for China and I still do not think that it will be able to develop economical leading-edge manufacturing within its borders.
• This means that it is likely to focus on older nodes at 45nm – 28nm where there is still plenty of business to be done but it does mean falling further and further behind.
• As a result, China is determined to achieve as much technology independence as it can, meaning that it will create its own standards for new technologies.
• Hence for technologies like autonomous driving, 6G, robotics, The Metaverse, and Artificial Intelligence, I expect that there will be no global standard leading to a bifurcation of the technology sector.
• From an economic perspective, this is never a good idea as two incompatible networks generate far less utility than one large one even with the same number of users.
• Hence, while the struggle continues, the long-term growth of the technology sector is likely to be the main casualty.