Just when we were nearing a feeling that it seems to be safe to go back into semiconductor waters, along comes Putin and bam – another kick in the derriere for semiconductors. As if it was not enough that the last administration sent the U.S. economy to the bottom of the ocean with its moves against China, now the Ukraine situation has thrown salt on the wound.
And there is more. There is the fear that China is mulling over invading Taiwan. If that happens, we can really kiss any semblance of a semiconductor “recovery” goodbye for the immediate future.
Of course, there are other variables that influence how the semiconductor industry reconfigures. One is the possibility of a recession. Another is unforeseen geopolitical conflicts not currently on the radar screen (Middle East, for example). A third is natural disasters. Any of these can further upset the semiconductor industry recovery cart.
Perhaps the biggest unknown revolves around Asia. Most of us know TSMC produces the lion’s share of the world’s semiconductors. In fact, Taiwan controls more than 90 percent of the global output of one of the more advanced semiconductor technologies. These are particularly necessary for advanced weaponry used by both China and the U.S. So, the issues here are not simply supply chain for commercial and consumer use, but also for national security, military, and things like AI.
If China invades Taiwan and takes control of its chip manufacturing segment, this will be the third strike against the U.S. semiconductor industry (China; Russia – by interrupting the Ukraine’s production of core materials; and Taiwan).
The outlook for the U.S. to acquire semis is still shaky, even though there has been a loosening of the offshore supply chain and some semiconductor supplies have begun to flow again. Almost all semiconductor manufacturers have plans to either bring some, or all, production back to the U.S. or partner with someone who is. But it will be a lengthy and expensive process, even with the government throwing money at it. We all know the pace at which government funds flow in such cases. As well, the physical manufacturing infrastructure will take time to build. It will be several years before we can build enough of an onshore supply infrastructure to make a dent in the demand. The U.S. is not expected to ramp up production significantly until 2024, more likely 2025-6.
Of course, there is product we can get from our allies. But they are also suffering from supply chain issues. They need them as badly as we do. After all, when they buckled under the pressure from the Trump administration to shut out China, that left them in the same boat as the U.S.
Our allies in the Pacific rim (Japan, South Korea, and Taiwan) supply nearly 70 percent of the global supply. TSMC makes up about 22 percent overall and Taiwan is a lynchpin for the more advanced semiconductor technologies. These are the critical components, particularly necessary for advanced weaponry used by both China and the U.S. If they are gobbled up by China, Asia’s output will drop by about one third. And, with Taiwan in Chinese hands, the U.S. and the rest of the free world lose access to the critical technologies it supplies to our security sector while China will have access exclusively.
The Ukraine situation has put a damper on raw material obtained from there (and it is looking more and more like a protracted war). Certain raw materials – argon, krypton, xenon, and neon critical to semi manufacturing are now, or will be, in Russia’s hands; Ukraine produces 70% of the world’s neon, or did until recently. Even if Russia does not take over Ukraine, it cannot supply materials as long as the war goes on. And there is no excess supply of neon gas anywhere in the Western world. As well, Russia produces 40 percent of mined palladium. These are just a few of the current challenges in the resource space.
So, as one can see there are there are a number of uncertainties facing the chip segment beyond moving manufacturing back to the U.S. In addition, finding alternative sources for raw materials is easier said than done. Again, everybody is in the same boat. There is competition from the rest of the world for such product; chip makers, globally, are under intense pressure from the ramp-up of demand as a fallout from the pandemic.
If the scenario of Taiwan falling into Chinese hands comes about, that puts Russia and China in a power position unparalleled in modern history. China is already ahead, and will be even further ahead by the time the rest of the world gets anywhere near back up to speed in advanced semiconductor technology. One noted analyst said that by the time we are back in the semiconductor game, China will be 10 years ahead of us in advanced semiconductor technology.
Assuming the status quo remains for the immediate future, the best-case scenario will drag the semiconductor crunch out another year or two, as supply chains ramp up. If not, the worst-case scenario may see an additional two to four years added to that as well as an extended technology catch-up cycle. So, if Russia takes over the Ukraine’s resource and China takes over Taiwan, the free world will be grubbing for semis for years to come. And the ripple effect of that is hard to predict but it will certainly have a significant, negative effect on the U.S. and other nation’s economies (on top of the rampant inflation).
The semiconductor industry’s awakening
However, there is hope. The once secure and ostrich-minded semiconductor industry, which has plodded along for decades with little interruption or variance in its procurement and production model, is having a rude awakening. They are used to the JIT and uninterrupted supply chains. However, that is history for the time being. They are now being dragged, kicking, and screaming, into a tumultuous, fast-moving technological world where the supply is constantly challenged. Now, chip firms are having to manage the fill-levels of an array of materials buckets, each with changing dynamics, each one of which could lead to a chip shortage in a particular market segment.
Demand is not proving easy to get a handle on. Intense post-COVID demand, combined with unexpected societal and geopolitical events affecting the market, can quickly ripple through the entire supply chain. That leads to a level of insecurity on the purchasing side. That, in turn, can lead to over-purchasing and hoarding by the vendors and other chip users.
This makes it difficult to understand what the actual levels of demand are, leading to a see-saw cycle of under- and over-capacity.
The semiconductor players are learning, however. They are becoming more rational about how to deal with a changing marketplace and ecosystem and adopting a longer-term model. That translates into a supply chain devoid of buying cycles–no purchasing peaks followed by periods of lower demand, during which the supply chain can focus on replenishment. Demand for more chips from more buyers from many more industry vectors will create an even more challenging marketplace.
What the “new” chip industry normal will look like
Given all of this, what does the crystal ball say for this segment? Mostly new transformation strategies.
The first of which is to be able to deal with disruption in planning. That means having a clear understanding of all the elements in play. Companies must have a bottom-up understanding of what their different parts are doing because transformation has to address all the components simultaneously. Any issues with one vector will have ripple effects across all company and customer chains.
Another is to widen the supply chain. Gone are the days when back-room deals are made that are a win-win for a few big players. Semiconductor companies will have to be open to new segments, partnerships, and new ways of building products. Vectors such an AI, edge computing, internet of anything and everything (IoX), autonomous vehicles, and more will all have to be understood and customers will be a much wider group of users. As well, agility will be the word of the day.
Finally, while this has been evolving for some time, it is about to hit light-speed. The standard model for chip sales is one and done. Once the chip was out the door, it was out of the manufacturer’s sight. New models of revenue, like subscription-based services, moving to software services especially, and partnering from cradle to grave with customers, will be opportunities to level the speed bumps of raw material supply chains. Those which stick with the one and done mentality will become as extinct as the dinosaur.
The semiconductor industry has to look at this evolution as an opportunity rather than a challenge. The pandemic and geopolitical events have been a lightning rod for change. As we move forward, whatever momentum is built up must be capitalized on. The old adage of being able to pivot on a dime is one which the semiconductor industry must embrace.
Ernest is a Senior Member of the IEEE, the Executive Editor of AGL’s Applied Wireless Technology, staff editor of AGL’s eDigest. Other credentials include being the Editor of the IEEE regional newsletter, Adjunct Professor, Colorado State University College of Engineering, the Vice-Chair, of the Denver IEEE ComSoc Chapter, a technical advisor for the Telecommunications Industry Association (TIA) Standards Development Committee, and former Technology Editor of Semiconductor Engineering’s IoX and security channels. He has been the editor of several other high-tech publications over the years as well.