As the author of “Chip War: The Fight for the World's Most Critical Technology”, Chris Miller usually has more insights on chip issues. Yesterday, he participated in a virtual panel organized by Nikkei Asia and shared his thoughts on various topics. Here are some key points:
1017 Chip export control: The updated export control to AI chips underscores the fact that when the US government thinks about why semiconductors are important, training AI systems is one of the most significant use cases they perceive. That's a primary focus of US policymakers on AI training systems and the potential defence and intelligence ramifications of those systems.
Huawei's progress in 7nm chips: What's the volume that Huawei is going to produce of phones with 7nm chips over the next 12 or 24 months is going to be a key question going into 2024. Haven't learned a whole lot new about the key question. Apple is the most interesting of the other competitors that is going to face some pressure from the fact that Huawei is back in the high end of the smartphone business in a serious way. It will be watching very closely as to what those quantities look like. We're gonna find out in 2024.
AI militarization: Many different militaries are trying to deploy AI to defend systems. The US might be a leader here, but it's actually very difficult to know how to measure that or know how to assess which countries are in charge because we're still in the early stages of what will be a major change in the way that militaries operate. That's why we see political leaders and military officials all over the world are focusing on this area.
Research that Chris will focus on for the next year: Trade issues posed by the investment around the world in lagging edge semiconductors, particularly what are the implications of the build-out of capacity in China for lagging edge. He believes there will be more discussion over the next year about whether a similarly tough approach, like the EU's investigation of overcapacity in China's EV industry, will be applied to lagging edge semiconductors.
Future of development of the AI industry: A main winner is going to be companies who are focusing on providing the hardware that AI requires. He expects over the next half-decade or decade, many new companies to emerge in this space, focusing on providing the processing that's needed for inferencing on edge networks.