Geopolitical And Economic Outlook For Chips And Equipment

Experts at the Table: Semiconductor Engineering sat down to discuss geopolitical and economic changes and how they affect the chip industry with Jean-Christophe Eloy, CEO of Yole Developpement; Risto Puhakka, president of VLSI Research; Carolyn Evans, chief economist at Intel; Duncan Meldrum, chief economist at Hilltop Economics; and Rozalia Beica, head of the semiconductor business at AT&S. What follows are excerpts of that discussion, which was held at the Electronic Components and Technology Conference.

SE: How are we going to solve the current capacity shortfall? Initially, it was only supposed to affect 200mm wafers, but it’s now across the board.

Puhakka: First of all, shortages are truly real, and you cannot spend your way out of this. You need the entire supply chain to deliver deliver components, and some of them are in short supply while others are available. But if you’re missing one component, you can’t produce a car. It really comes back to the equipment suppliers. We saw some of the equipment markets like assembly and packaging growing 40% this year, and that may be a conservative number. The equipment supply chain is stretched all the way to the limit. With scanners, we’re looking at a one-year lead time. But our expectation is that by the end of this year supply will meet demand.

SE: What impact does this have on heterogeneous designs and materials?

Beica: The materials are there, even if the equipment is not. We are seeing a lot of growth in IC substrates, especially driven by servers and high-performance computing, but it’s also across a lot of different markets. The lead times for equipment are increasing, with many companies putting in orders. The whole supply chain is impacted by high demand for substrates, and it is affected by shortages.

Eloy: There are a lot of things working in parallel. There is high growth in servers and high-end logic devices, which are moving to more volume production, which means there is a lower volume of wafers for other things. There is a lot going on, including a technology transition in packaging. Those combinations mean that the level of growth at the moment is linked to price increases and what companies can produce. An equipment delay of 12 months is really incredible, and it is affecting investments. And because there is no equipment available, you may not be able to increase the amount of equipment you have until 2023. The positive side is a growing recognition that, from foundries to subtrate to devices, semiconductors are important for the whole economy. That means there is a lot of funding available for R&D, and funding to enable the industry to grow. This is really good for an older business, because it’s a long-term view for different geographies and from different governments. Semiconductors were really not news in Europe and a lot of countries before this. The business impact is significant, but the poltical impact is really, really important for this industry over the long term.

SE: Given the supply constraints, is the industry experiencing the kinds of double- and triple-ordering of parts and supplies that we saw back in 1999 and 2000?

Meldrum: As an economist I look at behavior and long cycles, and I would be very surprised if we aren’t getting some of that behavior going on in the background, particularly given the disruptions in the supply chain. I doubt it’s the kind of risk we had in 1999. The industry is a lot smarter this time, and we have fewer players. But it’s something we should all be cognizant of. As we start traveling again, those of us who bought three, four, or five PCs for various members of our households are not going to be buying for awhile, so there’s a downdraft ahead — but probably not for a year or two at the earliest, because we have so many economies around the world that have yet to make really good progress against the pandemic. There are enough risks out there that the underlying strength of the industry should hold up, even if we do get a downdraft with multiple orders that eventually get canceled.

Puhakka: Looking at the inventory data, so far we haven’t seen a substantial increase in inventory. The overall inventory level is higher, but so is the business level. The inventory-to-billings ratio is actually pretty low. But when you talk to various points in the supply chain, there is definitely an interest in running it with higher inventory levels. So instead of just-in-time, now it’s a little more just-in-case. There are truly big supply chain disruptions that everyone has been dealing with, and the easiest way to alleviate that is with higher inventories. In the long-term, companies are run by CFOs or operations managers, so that’s going to be a battleground inside of companies.

SE: In the past, it was pretty simple to predict how the chip market would fare based upon sales of PCs and mobile phones. The market has become much more fragmented and global since then, and there are a lot more hot technologies that never existed before. What impact does that have on your market outlook?

Eloy: There is a strong movement in terms of applications that are really driving growth. So in addition to PCs and mobile phones, now we have to consider automotive and autonomous vehicles, which means more high end devices, as well as servers and high-performance computing. There are multiple markets driving growth, but they also are driving technical changes. Automotive is moving from very low-end to very high-end devices. It’s the same for advanced packaging. You have to look at all of those combinations. There are multiple markets driving advanced packaging for different applications, and for new devices that are being integrated because of some big transitions that are happening in fan-out and 3D. And all of these are happening at the same time. So now you also have to understand market transitions on multiple levels.

Beica: We see a lot of applications on the high-end servers, mobile, automotive, industrial and medical, and there are requirements coming from all these different markets, and applications within these different market segments. There is connectivity, 5G modules for the high end, high-performance computing, automotive going to more advanced nodes and requiring more advanced packaging. I’ve been in the packaging and semiconductor arena for quite a while, but I haven’t seen such high growth or as many opportunities before. There are many technologies being developed, and you can’t really say there is a standard technology. That provides a lot of opportunities, but too many companies also are coming up with their own technology or system integration to address the different requirements.

SE: On a macroeconomic level, how will inflation impact various markets and new technology adoption?

Evans: We don’t know for sure yet whether there will be inflation, and there’s a lot of debate about that right now. There are two broad camps. One includes the Federal Reserve, which is the central bank of the United States, and they would argue that we’re seeing some spikes right now. If you look at the most recent inflation data that came out, there were definitely price increases, and more than we’ve seen in awhile. But the Fed also would argue there’s a low base for comparison. So in 2020, prices were very soft because of the pandemic. We saw oil prices go negative for a period. At the same time, there are supply chains everywhere that are having problems catching up with demand. It’s not only physical goods being moved. It’s also services. You hear about restaurants that can’t get enough people to serve customers. And there are a number of different industries that are not able to meet demand. What leads to inflation generally is supply that can’t meet demand, which creates upward pressure on prices. The Fed and others say these are temporary spikes and that we’ll go back to a 2% or so target. There are others who say it’s going to go on for awhile, and that it’s not temporary. Maybe for some reason demand has changed in such a way that there’s a shortfall in supply, and we’re going to have this gap between demand and supply. And so therefore there’s going to be upward pressure on pressure on prices for a longer period of time. That context affects this industry specifically. We have to think about relative prices. So if the prices of everything else go up, and the prices of technology doesn’t go up, it looks relatively cheap. And if the price of technology goes up, you’re sort of on a level playing field with everything else. That’s one piece of it. Another piece of it is that if there’s a big spike in prices in a particular area, and consumer and business dollars go to that area, there may be less available to spend on technology. It’s going to depend on whether it’s a temporary spike, or whether it’s going to go on for a longer period of time, as well as where this inflation is happening.

SE: What’s the impact of a dual supply chain as a result of the trade war between the United States and China?

Evans: The general thinking is that a tariff is a tax, and it will distort behavior. Those who are affected by tariffs are going to change their behavior, perhaps to optimize it based on tariffs as opposed to other factors.

Puhakka: First of all, China is one of the top three equipment regions, along with Taiwan and Korea. But when you look at equipment shipped to Taiwan, if it costs $1 billion you know it will turn into wafers out very efficiently and very capably. When you ship that same $1 billion worth of equipment to China, it doesn’t turn into the same wafers out. So we asked what will happen with $10 billion of front-end equipment that’s going to be shipped to China this year, and we really don’t have a good answer. We see a lot of efficiencies in turning that into actual capacity. On the back end, it’s different because China can turn that investment into capacity very efficiently. The second thing is that we don’t see a Chinese supply chain related to equipment for semiconductor manufacturing, and we don’t see they’re making a lot of headway. It’s really the U.S., Japan and Europe dominating that suply chain. The Chinese supply chain is still enormously dependent on Western capabilities.

Meldrum: If you think about the world we were in pre-pandemic, or a few years before the pandemic, we had a supply chain that adjusted globally to a just-in-time type of model. The auto industry is finding out that doesn’t work right now with chips. But globally, we have materials being shipped all over the world. We have equipment being shipped all over the world. We have the capacity located more or less in optimal places. When we start putting tariffs on this, we distort that. If we look at broken supply chains and get excessive responses to try to rebuild them in economies where they no longer exist — and assuming we still have them in economies where they currently exist — we’re going to end up with a higher-cost system with more capacity than we may need years down the road. That’s the risk we run. It is a risk in that it makes the entire supply chains a little bit more expensive than they otherwise would have been.

SE: From a technology standpoint, what impact will domain-specific designs have on packaging and customization? We’re seeing differences by industry as well as by region.

Eloy: As you said, there are more and more things that are defined as customer-specific or application-specific, not as a roadmap for the industry. So that’s quite different than before. Second, if we are talking about China, and particularly packaging in China, the supply chain is moving out of China and into Southeast Asia. Vietnam is a haven for semiconductor at the moment, which is different than several years ago. The Chinese supply chain is not really there in terms of equipment. Investment is there, but it will take 10 or 20 years to develop the supply chain. But if you compare China now compared to what it was like 10 years ago, you can imagine what China will be in the next 10 years. It will be at the same level as Japan and Korea. Still, the next 10 years will be as challenging for the Chinese supply chain as it has been for the last 10 years. When it comes to packaging, a lot of companies are shifting away from a stalled out roadmap in order to design what the customer needs. And when the biggest companies like Apple and others change the way they work, supply chains will follow.