China vs. USA – A Drop in the Ocean

China’s new machines don’t add up to much.

• China has announced new semiconductor manufacturing machines that will help it eliminate its dependence on foreign equipment, but the reality is that these machines are so far behind the leading edge that I think that they will have very little impact.
• The USA’s current policy is to limit China’s rise as a geopolitical, technological and economic power by limiting its access to advanced semiconductors.
• This will have the impact of making it much more difficult and more expensive for China to develop new technologies right the way across the technology spectrum and not just in semiconductors.
• In effect, semiconductors have become a tool of US foreign policy.
• The Chinese Ministry of Industry and Information Technology (MIIT) has announced the existence but not commercial availability of two new semiconductor lithography machines from an unnamed manufacturer.
• One of the machines operates at 65nm with an overlay accuracy of 8nm and the other operates at 110nm with an overlay accuracy of 25nm.
• By comparison, the equivalent machine from ASML operates at 38nm with an overlay accuracy of 1.8nm.
• This is a step forward for Chinese semiconductor equipment manufacturing as the best currently available from Shanghai Micro Electronics Equipment Group (SMEE) is 90nm but in the grand scheme of things, this is a drop in the ocean.
• This is because US restrictions prevent the export of equipment to China that can operate at 20nm and below with leading-edge EUV at 7nm and below.
• Hence, Chinese companies can easily buy the equivalent machine from ASML or Nikon and get better overlay accuracy which is an important factor when looking for high yields.
• They do nothing to alleviate the problems that China has in procuring and servicing machines that can operate at 20nm and below.
• The real point of these machines is not necessarily to displace ASML and Nikon but is more about the mission to become self-sufficient in leading-edge equipment which would render the restrictions useless.
• This latest machine represents a small step on that road but there is still so far to go that I remain unconvinced that China will get there before silicon becomes obsolete.
• 21 years ago Intel was the first to announce 65nm with widespread adoption coming in 2006 and 2007.
• This means that China is still at least 21 years behind the USA and its allies, and it has yet to make any tangible progress on EUV.
• Development of EUV was started in 1990 and it took 25 years to get it to work reliably enough to be commercially viable.
• This is because the level of precision that is required to manufacture these machines is by far the highest in the world.
• Furthermore, the 13.5nm light source is fiendishly hard to make and involves passing droplets of tin through a carbon dioxide laser that heats it to the point where it becomes plasma at which point it emits the 13.5nm light required.
• Consequently, I think that if these new machines are commercially viable, they will marginally reduce the dependence that China has on foreign machines but not in any meaningful way.
• Furthermore, I suspect that this does not represent a breakthrough or a sudden acceleration of China’s progress and consequently, I still do not expect to see EUV machines made by China working at commercial yields for at least 10 to 15 years.
• Hence, I continue to think that for the foreseeable future, China will struggle to make advanced semiconductors at high yield and low cost.
• This means that the US restrictions which are expected to be updated next month will continue to be effective with very little impact coming from the availability of these new machines.
• This continued rivalry means that the days of global standards are numbered and, going forward, we are likely to see one standard outside of China and another, competing, and non-compatible standard inside China.
• This is bad news for everyone as two incompatible networks will generate much less value than one global network.
• Consequently, long-term growth for the entire technology sector over the next 10 to 20 years will be lower than it otherwise would have been.