The Cost of Deglobalization will be High
An industry whose commercial structure is based heavily on a global scale discusses the unthinkable: deglobalization. Will this ever work and what will the ramifications be? Deputy Chief Editor Heinz Arnold (Markt & Technik) and Georg Steinberger, freelance Digital Supply Chain Expert at Sourceability, approached this topic at Embedded World.
Heinz Arnold: In the last 2 years, the European industry – mainly automotive and industrial – mainly suffered from an extreme shortage of components made at larger geometries, which to some extent continues. How will the situation and the market develop?
Georg Steinberger: We see the supply situation improving and a weakening of the order situation, which is no surprise after two years of record demand and orders having been pulled forward. In other words, an inevitable market correction. Thankfully, it seems that Europe is not going into an overall recession. And mid-term the demand created by the transformation of the transportation market towards e-mobility and a few other factors will remain high. Sure, the overall semi market sees a downturn currently, mainly driven by weak smartphone and PC market data, but they could come back later this year. So will China after ending the zero-Covid strategy. At the moment, the name of the game is the working of built-up inventory in the market.
Arnold: How did the last shortage impact the relationships between the different players in the supply chain – manufacturers, distributors, EMS, OEMs?
Steinberger: There was a huge pressure on everyone in the channel to rectify the supply chain disruptions caused by the pandemic and the subsequent demand surge across all industry sectors – particularly the computer and communications industry. A lot of challenges were in distribution and at EMS companies. The clear winners were manufacturers, at least until now. Proof: their financial reports. This is not meant as a judgement, as their financial risks are high as well. The lesson learned in my opinion is that in order to mitigate the risks in the supply chain – eliminating would be too much asked for – we need more transparency in the supply chain across all players. From a manufacturer's perspective, the dependency on one player – TSMC – is too high, but due to their excellence at leading-edge semi manufacturing, there is no easy remedy.
Arnold: How do you think more transparency can be achieved?
Steinberger: We live in the age of digitization, but many companies still do their planning isolated, based on the limits of contracts and with a lack of visibility. Imagine a platform where all components that are available would be visible with all necessary data, not just technical but anything that is supply-chain related and can give you a clue on future availability and potential risks. This could make planning easier and take out a lot of anxiety which leads to double ordering, panic and so on. The prerequisite is that the big market participants embrace this and make their data available, so that customers see the full market picture. Sourceability has created such a platform (Sourcengine) and associated tools (Datalynq for case management and risk analysis and Quotengine for quoting optimization) to try to achieve that. I am confident that many customers are waiting for such tools.
Arnold: The growing tensions between the USA and China put the business models of US companies as well as western countries in general in jeopardy since most of them are generating a considerable amount of revenue in China. How can these relations be disentangled and what will be the impact on the supply chain?
Steinberger: As you can see in the heated discussions in Europe, there is no clear answer, but the “de-globalization” is starting to happen. The United States is pushing hard to cut China out of the global electronics supply chains. It is more globalization minus 1. At the same time billions are invested to create more strategic independence in semiconductors – in Europe, in the US, in Japan, in South Korea. Where this ends, we do not know yet. What we do know is that the cost for this endeavor will be very high, for taxpayers, for industries, for individuals. The electronic components business in my opinion needs global scale to be economically successful. With leading-edge fabs costing north of $20 billion USD, regionalized markets or a global market minus 1 will not be able to help amortize the investments.
Arnold: It seems that governments in regions such as USA, Europe, and Japan are trying to coax Taiwanese foundries to invest in their regions in order to secure production capacities. Europe also seeks to attract leading semiconductor manufacturers. Will these activities lead to building new semiconductor networks or initiate new semiconductor wars?
Steinberger: The irony is that fabs alone mean nothing without an ecosystem that starts by creating intellectual property that can be put on silicon. Having a Taiwanese foundry in Europe does not mean it produces for Europe. Foundry customers are other semiconductor manufacturers who use foundries for additional capacity or technologies they cannot produce themselves. Very rarely do you find an OEM using a foundry or investing directly in wafer capacity. So, the fab output goes where demand is and so far, this demand (at leading-edge) was not in Europe. The US has a different reason to get leading-edge capacity built in the US, 40% to 50% of semiconductor IP (leading-edge) is in their hands.
Arnold: How will this change the market? What will the new semi world look like?
Steinberger: The conflict is actually bigger than the components market, so a prediction of the outcome is terribly difficult. It is a giant systemic conflict in which Europe will have to choose sides. At the moment, neither the US Chips Act nor the European Chips Act will be stoppable, nor will other initiatives of a similar kind. New fabs will be built and ecosystem initiatives will see enormous funds. Giant risks and also giant opportunities. If Europe wants to prevail in this game, it has to stop internal grabbling and create a long-range industry policy that starts with education and has to imagine its participation in future markets beyond its current automotive focus.
Arnold: Talking about automotive and e-mobility, a significant part of semiconductor growth is considered to come from automotive, e-mobility, and energy markets due to growing electrification. Could you imagine that expectations are exaggerated? Is there a gap between expected and real-world demand?
Steinberger: The train on e-mobility is out of the station, the transformation from fossil combustion to EV is on the way. And each EV needs significantly more components than a comparable traditional car. Replacing over the course of 30 years 1 billion fuel-based cars will mean more semi revenue from automotive than ever before. The question remains who will win (which car makers, US, Europe or Chinese) and how individual traffic will develop in general. My guess is that we will run into resource issues that are by far bigger than the semi question. And not to forget: EVs that are more connected, autonomous, whatever, will need much more computing intelligence in data centers, too. Automotive outgrowing the computer and communications market? Maybe not.
Arnold: Do you think automotive and industrial will be able to capture 30% of the global semiconductor demand? Further, do you think they could replace computers and communications as technology drivers?
Steinberger: Technology drivers who demand leading-edge process like 1nm must have big enough platform sizes and performance requirements. As industrial is normally small to medium volume, they will remain fast followers. In automotive, 80% of cars will be economic and only 20% will show complex architectures that could need leading-edge. But in total automotive will not exceed 150 million units per year (currently ca. 85 million). Smartphones: 1.5 billion; computers & servers 300 million; tablets 150 million. Now throw AI and Web 3.0 into the equation and you will see a surge in computing and communication hardware over the next years that may dwarf the automotive market.
Arnold: Where do you see Europe’s opportunities? What does it mean to strive for strategic independence?
Steinberger: Strategic independence in the semiconductor industry is in my opinion difficult to achieve – at least for Europe. It would take a generation, cost 100s of billions of Euros and would have to encompass a tectonic shift in education as well as reimagining completely new markets. We are less than 10% of the world market, so we are missing a huge piece of the pie. Still, I appreciate what is happening on the fab front in Europe, we just need more around it, maybe that will help drive new markets, too.
