Odd Lots - Stacy Rasgon on How the Global Chip Crisis May Be Getting Even Worse

Episode Date: September 16, 2021

We've been talking about chips on Odd Lots for almost a year now. Thanks to a unique combination of events and constraints, capacity to make more semiconductors is incredibly tight. One industry that's lost out significantly is cars, as automakers are still cutting production due to an inability to source chips. On this episode, we speak with return guest Stacy Rasgon, a Managing Director and U.S. semiconductor analyst at Bernstein to discuss the current state of the industry, and why things are still so messed up.See omnystudio.com/listener for privacy information.

The news doesn't stop on the weekends. Context changes constantly. And now Bloomberg is the place to stay on top of it all. Hi, I'm David Gurra. Join us every Saturday and Sunday for the new Bloomberg this weekend. I'm Christina Rafini. We'll bring you the latest headlines, in-depth analysis, and big interviews. All the stories that hit home on your days off.

And I'm Lisa Mateo. Watch and listen to Bloomberg this weekend for thoughtful, enlightening conversations about business, lifestyle, people, and culture. On Saturday mornings, we put the past week's events into context, examining what happened in the markets and the world. That on Sundays we speak with journalists, columnists, and key political figures to prepare you for the week ahead. Join us as soon as you wake up and bring us with you wherever your weekend plans take you. Watch us on Bloomberg Television.

Listen on Bloomberg Radio, stream the show live on the Bloomberg business app, or listen to the podcast. That's Bloomberg this weekend. Saturdays and Sundays starting at 7 a.m. Eastern. Make us part of your weekend routine on Bloomberg Television, radio. and wherever you get your podcasts. Hello and welcome to another episode of the Oddlots podcast. I'm Joe Wisenthall.

And I'm Tracy Allaway. So Tracy, you know, obviously all year we've been talking about inflation and supply chain issues and shortages. But one of the first areas that we really started talking about was the had to do with the disruptions and semiconductors. Yeah, I think this was like. our entry into global supply chain issues. We started talking about semiconductors and this idea that there was a shortage of these chips that were needed for all sorts of things. And I think, you know, if you are of a certain age or generation, you tend to think of semiconductors as something that goes into a computer. But of course, nowadays they show up in phones, in, you know,

things like refrigerators and, of course, cars as well. Just almost every major appliance now seems to have some sort of chip embedded in it. Right. And we did like a six episode series, I think, earlier this year about chips. But the chip story, even though we haven't talked about as much lately, it really hasn't gone away. And I think I just saw last week. There was a story about some chip manufacturer in Malaysia, reducing production. They're pretty regular stories. You mentioned automobiles about car companies still not being able to get an adequate supply of chips and thus being forced to cut production. So the issues have not gone away by any strides. Yeah, I think Toyota made a massive downgrade to its expected production. But yeah,

so this is something that I've been wondering about. So on the one hand, yes, we have some COVID outbreaks in Asia that have disrupted semiconductor factories or manufacturers, particularly in Malaysia where they manufacture. My understanding is they manufacture some components for semiconductors or something like that. So, okay, you have the Delta variant, and that's not production to some extent. But on the other hand, you know, as you said, this issue has been going on for quite some time now, like more than a year. And the big question is, why aren't we getting to a better place in terms of production? At this point, the semiconductor companies know that they need to boost production. They know there's lots of demand out there. So why isn't that happening? Well, hopefully we are

are going to get some answers today on the show, why it is that we're now in the middle of September. And we still don't have something resembling like a smooth semiconductor supply chain. We're going to be going back to our first guest that we had on the semiconductor series. I think he actually appeared twice already. The last time was in March. We're going to be speaking with Stacey Razgan. He's managing director and senior analyst at Bernstein covering U.S.

semiconductors. He's going to hopefully answer all of our questions. So, Stacey, welcome back to Adla. Oh, it's great to be back. Thank you for having me again. So are we right? I mean, it appears not only are the issues persisting. Like in some, in some measures, things seem to be getting worse. Yeah. And so we had hoped that by now things would be getting back to some semblance of normalcy. And you're right. They're really not. It's actually getting worse, not better in many parts of the market. So we still are seeing, you know, production shortfalls. We're still seeing shortages. You know, we've been hoping, obviously, that, you know, COVID would be, you know, you know, on the decline at this point, it is not, it is still causing disruptions. You mentioned Malaysia and some other areas. So, yeah, we're not back to normal yet.

We're still seeing some of the same issues that we've been seeing for quite a while. So what exactly is driving those disruptions then? So, you know, you mentioned some COVID outbreaks in parts of Asia, but it seems to be, you know, you might see some factory closures, some isolated factory closures, but it seems like the issues are sort of more endemic than just, you know, an outbreak at a particular factory. Yes, it may be helpful to go back a little bit and talk about what caused these shortages in the first place. And I talked about this.

I think I can't remember what stuff. Maybe it's the second time that I was on. But if we look at the automotive sector first, that's where we're seeing some of the biggest issue. We've seen the biggest issues all the whole time. And you're right. We're still actually seeing production cuts. You mentioned Toyota.

We also had recently GM and Ford and a bunch of others. And it's been going. so they're still seeing shortages. You have to remember, the issue there was initial supply chain whipsaw. In the beginning of COVID, factories got shut down. People were locked in their houses. They weren't driving. Demand, at least initially, for a small amount of time, plummeted.

And remember, the auto vendors canceled all their orders. Demand came roaring back. Obviously, the plant started opening up. They went back hat and hand to try to get capacity to get production, and it just didn't exist. And it was like, well, get in line, right? It'll take time. And remember what I said, you know, even if you're starting, even if you have the capacity available, it takes months to make these chips from a dead start.

And so that caused, you know, cascading effects. I mean, it still has, right? And, you know, the other issue in the other end market is just demand is very strong. And that was part of the reason we saw issues in auto as well. I mean, the capacity that they were using got backfilled in many cases by other stuff. But just in general, we've seen really, really strong demand. I mean, much stronger than I think anybody. thought was going to be possible in things like PCs and like anything that was sort of work

from home, study, from home, play from home. I think you also need to couple this with the fact that the industry itself, I mean, you know, remember semis, it cost billions of dollars to add capacity. And, you know, historically, they, you know, they added capacity to demand that they saw. And when demand was very, very strong, much stronger than the, I just didn't have the capacity, it was available. And because that demand is still strong, you know, we're still shy, we're still short, we're still tight. And so now you have to layer on on top of all that, though. You know, remember this stuff is never smooth.

And when you have these kinds of problems, they do cascade and they cause more issues. And it's just a big mess. I mean, it still is. And so it hasn't, we haven't had time to actually bring enough capacity online to alleviate things. And the demand, you know, in most parts of the market, we can talk about where people are worried about, you know, demand may be starting to have to roll over a bit. But in general, demand is still strong.

Supply is still tight. and we've still got problems. It hasn't alleviated yet. And now you throw the resurgence of Delta on top of this. And again, you mentioned Malaysia. Well, they do a lot of Malaysia. They do what's called back-end. It's a lot of what's called packaging.

So they don't necessarily, they don't have like semiconductor factories in Malaysia. They're not making like chips. Once you actually make the chip in the factory, you have to what's called packaging. You have to put the electrical leads and everything on it such that you can take that chip and actually put it on a circuit board.

So they do that kind of stuff, assembly and testing and other things in Malaysia. And a lot of those issues were impacted. A lot of that stuff's done in Southeast Asia. And a lot of those things were impacted by COVID. So that's just something else, icing on the cake, on top of everything else. It's not getting better.

Yeah, that was one of the things I remember we talked about in one of our semiconductor episodes with Willie Shee at Harvard Business School. And obviously, just like in addition to the sort of technical challenges, is any given ship has a highly globalized supply chain. And so as you mentioned, maybe it gets manufactured somewhere in Taiwan, but then some sort of packaging happens in Malaysia. And so there must be, to some extent, issues that are, the sort of generalized supply chain issues that are affecting more or less everything

must have some having some impact on semiconductors specifically. It is. And so you're right. So semis are extremely global. I forget how many different borders the average semiconductor crosses like in its lifetime from initial production to actually

getting sold to the end user, but it's a lot right. And you're right, they can be it can be a U.S. company that's using a Taiwanese foundry to make it where it gets packaged in Malaysia and then it goes to a distributor in China and like it's all over the place. So there's a lot of that and there are a big logistical channel

the challenges that we're seeing right now. Shipping costs are actually getting much higher. We've seen port shut in China and other places as they're trying to control the spread of Delta. We've seen container ships like stacking up. I think maybe, I can't remember if the Suez Canal was blocked the last time we chanted, I can't remember.

But we're seeing tons of logistical challenges. A lot of semiconductor companies in front of the other companies, too, are starting to talk about this as well, like when they talk on earnings calls and everything. We're starting to see some of those costs go up and some of these supply chains are, it's just a big hot mess. I mean, the whole thing, which I guess shouldn't really be a surprise. I mean, again, you know, we had a global, potential global catastrophe, you know, that was causing as part of me is, like, amazed that it's been as resilient as it has been, frankly.

But it is, it is still causing problems. And it's funny, like, obviously the question I get the most is, how long is it going to last? I wish I had an answer. I don't, you know? And it's really funny when you think about the stocks, right? And the companies, I mean, they're still putting up monster numbers. The sector right now in terms of revenue is on track. to grow probably 25% this year and blow past $500 billion for the first time in history.

And every quarter of their companies are putting up better and better numbers. And I would say investor conviction is getting less and less, like the higher the numbers go. Because, again, we've all seen this movie before. We see this massive strong demand. Nobody really knows how much of it is real and how much of it is phantom. We talked about double ordering and everything else the last time it was here. You know how this is when things were uncertain, they order more. And so, like, it's funny, the stocks haven't done much of anything, you know, even as the numbers have it continued to go up and up and up, because people still don't know how sustainable this strong demand is.

So, Stacey, this is something that I wanted to ask you about. So, you know, you mentioned this cascade effect early on. And you didn't mention it specifically, but there is this bullwip effect that seems to be at play where everyone was expecting demand to drop. companies cut orders. So the semiconductor manufacturers sort of backed off from production and then things turned out better than people had expected and there was loads of demand and the semiconductor companies have been

trying to catch up ever since. But I guess my question is like one, is there something in particular about semiconductors and the manufacturing process that makes them more sensitive to the bullwip effect? This casket

idea. And then secondly, one of the things that comes into play when there's a shortage, a supply shortage, is everyone starts stockpiling, right? So it's hard to tell what's real demand versus what's just people hoarding components because they think they're not going to be able to get them. So what's your sense of that at the moment, real demand versus stockpiling? Yeah, yeah. So let's address the first question. They're like, why are semis the bullwip, right? Yeah, it's there at the back of the supply chain. So again, you think about it. I'll, I'll take a microprocessor for a PC just as a simple example, you know, like Intel or AMD or whoever makes the processor. And then who does that get sold to like a Taiwanese ODM, maybe it's called

original device manufacturer who builds like a notebook computer. And then they stick it on a plane or a boat, right? And they sell it, you know, they send it off to HP or something like that. And then HP, you know, sells it to Best Buy and then Best Buy sells it to you. Or like, however, there's probably other steps in there as well. But there's a lot of points within that chain where you can get inventory building or bleeding or anything else. And the fluctuations tend to magnify as you work from the front of the chain back. I mean, this is a very well-known thing. You see it all the time.

And the semis are at the back of that supply chain. So small fluctuations, even small fluctuations in end demand, like at the customer level, can propagate backwards and have correspondingly larger impact, like the farther back you go in the chain and the semis are just at the back. And remember, I think we talked about this last time, but, you know, that the semi-com companies have to plan in this environment. And their actual, my opinion, at least, is that semiconductor company management teams,

their actual visibility into what their end customers are truly doing is precisely zero. They have no idea. They do the best they can. I'm not blaming them. It's not good or bad. It just is. And the best companies out there are willing

to accept this and deal with it. But their actual visibility is not great. By the way, companies are trying to do some stuff to get to your second question around potential double ordering and stockpiling. And so, yes, this is a standard kind of behavior, probably not just in semis. I mean, this is normal human behavior.

Like when you can't get something that you need, you tend to order more. And again, this is a phenomenon known as double ordering. And I mean, a simple example, let's say you're making widgets. And you need, you know, 100 semiconductors, you know, whatever, right? And your vendor says, okay, I can't. I'll give you 10. I can supply 10.

And I'll give you the other, you know, 90 that you need. I'll give you those in 50 weeks. So your next step is to order a thousand semiconductors from wherever you can hope to cobble them together on the hope that you can build up the hundred to the eating and then you cancel all the other orders, right? And so what semiconductor investors tend to watch for, this is a phenomenon phenomenon known as lead times, like how long does it take to get the product after you order it and those lead times are stretching out? What investors tend to watch for is when the lead times start to pull back in because that's usually when the cancellation start to happen. And so what a lot of companies are doing right now is they're actually doing things like putting in

like long lead time orders, non-cancelable orders, for example. Generally in the space, historically, that you didn't do that. There were no penalties to canceling orders. Companies now, some companies are starting to put things like non-cancelable orders. Other companies are parsing their orders. It's like, well, I know this customer is ordering 100 parts, but I think they only need 30, so I'm going to only ship them 30, right? We're seeing some of that as well.

And so the companies are doing the best they can in this environment to try to manage through it. Is their stockpiling going? Undoubtedly, their stockpiling is. I don't see how there couldn't be. Can you measure it? No, it's hard. And I'll give you an example. There are some cases where, again, you go look at the auto vendors, even if they're

stockpiling, and by the, they may be stockpiling so much, but we still are seeing actual production cuts. So clearly, you know, even if they're stockpiling in some parts, clearly they don't have enough of what they need to build the cars that they want. So we're still seeing some primary examples here of situations where actual shortages, the shortages still seem to be real because we wouldn't be having actual production cuts if they didn't have actual shortages. I'm June Grosso, inviting you to join me for the Bloomberg Law podcast. Every weekday, we help you make sense of the legal stories that shape the nation and the world.

Listen for complete analysis of the biggest court cases, the latest actions from Congress and regulators, and the legal moves driving the markets. From corporate law to constitutional law and from state courts to the Supreme Court. At Bloomberg Law, we go beyond the day's headlines. We speak with top attorneys, judges, scholars, and policy experts to break down what the rulings really mean. We do this every weekday, then bring you the best conversations in our daily podcast. Search for Bloomberg Law on YouTube, Apple, Spotify, or anywhere else you listen.

On the East Coast, listen as you start your day. And on the West Coast, catch up in the evening. That's the Bloomberg Law Podcast. With me, June Grasso. Subscribe today wherever you get your podcast. Obviously, what you sort of described is going on in a range of industries. Are we going to hit one day? And could this be a source of investor concern where, like, we wake up and there is just a crazy glut of chimp because so many entities or is it so, is capacity so constrained generally that it's just hard to imagine there ever being a point where there's,

a huge over supply. I mean, never, never say never, right? And again, the normal practice is, yeah, you'd wake up one day and you'd see a glut. That is normally how this would happen. And again, like, this is why investors are worried. And this is why, even though the numbers are going up, the stocks aren't really going up all that much. multiples are coming down. And people are worried, you know, that we're, you know, getting closer to peak, right? And I guess mathematically, like by definition, we're getting closer to peak every day. At the same time, we could be, quote, and close to peak for quite some time, right? Because as long as this stuff's working on,

the other thing you need to remember with this is, even though demand is really strong, I mean, like, you need everything. So that's the thing is like the auto value, I'll pick on the auto vendors again. Like, I mean, they can be getting 95% of everything they need, but if they don't have the 5% that they need, they can't ship the car, right?

And you can have a situation where, like, you have auto vendor A, you know, they have 95% and they're short 5%. And auto vendor V has 95% of what they need and they're short 5%. But it's a different 5%. right? So that may be some of what we're seeing right now.

So, like, I don't know. Like, yeah, at some point, we may get a glut. And now at the same time, you can also argue that, like, post-COVID, some of this demand is real. And so, like, I'll look at PCs, for example. PCs right now have been incredibly strong. I mean, we were doing 250 to 260 million PCs a year pre-COVID in, like, in 2019. 2020, we did 300 million.

2021, right now we're on track to do probably close to 350, which would sort of match the peak in the industry, which is in 2011. And there's a big conjuristic PCs rate are PC's going to grow next year. They're going to be down. Like, whatever. But even if they're down, they're probably not going back to 250. Right? Probably not. Like, I could argue that there's a

structural case. We need that like, whatever the normalized run rate of PCs ought to be, it's probably higher than it was pre-COVID. And you can make that argument about a number of end markets. And so even if we get a glut, some of this things, we probably had pulled forward in digitization and everything pre-COVID or since pre-COVID. And we've

got more people doing remote. Remote work's going to be a thing from now on, like, whether or not people start to go back to the offices more. I could argue that some structural level of men, things should be higher. You know, in terms of capacity, I think Tracy, you asked during the intro, why isn't the semiconductor industry added capacity? It just takes time, right? They're all trying. I mean, you can go look at, like, the semi-cap guys, like, like, Wayfer Fab equipment, WFE is going through the roof right now. And you've got big forecast from, from big companies, you know, throwing out big numbers in terms of what they're planning on spending of few years. Capacity probably will get added. It just takes time to get added. So the worry,

obviously, is that capacity comes online right when the man starts to roll off. And I don't know what it's going to look like. I guess we'll, this year will probably be fine still in terms of the fundamentals of the numbers. We'll see what 22 and 23 look like. Those are the kind of timeframes we're probably talking about in terms of supply coming on. And we'll see how well it is to, well matched it is to demand like when it does it. So I wouldn't expect this to make a massive difference to the supply demand imbalance. But what are the chances that you start to see customers, the people who are actually ordering chips, just back away or decide they're going to design more analog stuff, you know, in order to avoid this problem? And I'm thinking specifically,

you know, last week Bloomberg ran an article about air conditioning manufacturers dropping copper components because the price of copper was so high. And they're all starting to, not all of them, but some of them are starting to switch to aluminum. So I'm wondering, like, could you see a similar thing with semiconductors, or is it just, you know, pretty impossible, given the nature of the types of electronics we're talking about? What do you mean, like, switching away from semiconductors to something else? Yeah, I guess, like, I guess going analog, right?

There were, like, a few car companies that were saying they're going to, like, drop some of the complicated electronics in order to produce. Well, we are seeing some of that. So some of the car, we've seen like cars for them. I'll make it up. but like shipping without, you know, the electronically controlled rearview mirror or stuff like that, right? We're seeing some things because remember, you know, if you're going to ship a car, you kind of need the full kit. If there's things you could leave out where the car is still saleable and then add them later, like you could do.

We're seeing some of that. The other thing we're seeing with like auto vendor sometimes is they're building the cars like every, they're building the car complete except for the stuff they have and they're parking it. So that when they get the part they need, then they can ship. I actually suspect we're seeing this in PCs as well. Well, my math that we published, as my math suggests that as strong as PCs are right now, I think CPU shipments, the microprocessor's shipments are even stronger.

And we've heard like HP has sort of talked about, on earnings calls about building things like strategic inventory. I suspect they're building kits, you know, and sitting on the shelf and so as soon as they get the final parts that they're missing that they can ship. So we are seeing some of that. So you can call that stockpiling in some sense, right? It's hard to know how widespread it is and everything else,

but I mean, it's a normal behavior that we might see. In cars, we are seeing, to your point, some of these things are getting left out. But the thing with cars, you can't just like, we can't go back. You know, like my first car was a 1978 Toyota Corolla. It didn't have very much in the ways of semiconductors in it, right? We're probably not going back to something like you can't. And in fact, one thing we're seeing from the auto vendors, the limited supply of semis that they do have, they're actually selling higher-end cars, right?

They're actually prioritizing them because the profits are higher. right so they're they're saving the limited semiconductors they have to actually make as most money as possible with them and so in some cases we may be seeing like mix actually going up rather than rather than down oh yeah i've wondered about that uh with with with various industries the the degree to which like okay if you have a shortage of components overall uh i remember we were talking about this a little bit we actually did an episode uh several weeks ago like on bathtubs and appliances but i've been wondering the degree to which end users of technology are raw materials are changing their mix and prioritizing the production of the higher margin lines. Yeah, it's hard to know. I do think in auto, we've seen that. In other end marketing, again, I could mention graphics cards. I mean, people, you can't even buy a graphics card right.

They're just selling everything they can make right now. But yeah, I do think we're seeing some of it. I do say anecdotally, we are seeing it. How much is, you know, obviously, on top of everything else, 2021 has seen this incredible crypto boom, how much does that contribute at the margins to tightness in the semiconductor space? And how much as a professional semiconductor analyst, does that force you to get to know this sort of new industry? Well, it's not that new anymore. We've seen a number of crypto, crypto booms and busts over

the years. Yeah, fair enough. I don't think like, so you can talk about like the crypto impact on shortages in a couple of different ones. They're just like, example, shortages of graphics cards, and again, go try to buy one. The other is like taking up capacity at the foundries. And I remember during 2018, you had a lot of the Bitcoin miners, guys like Bitmain that make custom chips to mine Bitcoin. I think at one point, they were a 10% customer of TSM, like, for a heartbeat. Right?

So that was like back during like the 2018 bubble. I don't think we've seen anything near that much for the guys that make the custom mining chips this time. I don't think TSM has been prioritizing those guys in the earliest months. But in general, I mean, we've just got tightness all around at the leading edge. And certainly that's been part of the problem with graphics cards. It's been hard to get. And obviously the demand has been very strong.

I'd also say, like, even beyond, for graphics cards, even beyond crypto, the gaming cycle this time is much stronger. Obviously, we had COVID and people were trapped in their houses and they wanted to, the video games were a good way to pass the time. So they wanted to buy the cards. And the products from both Nvidia and AMD this time were actually much better than the prior cycle, which was kind of like, you know, second half of 20. 2018. And so there's actually a lot more, I think, demand from gamers for the parts this time

versus last time where I do think a lot of that incremental demand was from the cryptocurrencies. And so what happened in 2018, 2019, when the miners stopped buying, the gamers didn't really want to step up and pick up the slack. And so obviously we saw some fairly big shortfalls, you know, the beginning of, end of 18, beginning of 19 for guys like in Vidia. This time, their new cycle, it's called, it's called Ampere, is the code name for it. I do think actually gaming demand is quite strong for it. They can't get the parts. So one hope for Nvidia's is if and when the miners stop buying that the gamers will

start to pick up the slack. And we'll see. Like we're not there yet in terms of a demand is still like off the charts. I think for both Nvidia and AMD for games for graphics cards. So let's talk about I think actually the first time we had you on who was about like the fall of Intel or Intel stumble. That was like the first theme. And of course it had a pretty rough 2020 and they had fallen behind.

So then they bounced. Then they hired the new CEO, Gelsinger, and the stock rallied, and people like this idea of like, oh, they're going to like reinvest in production, because that's the core issue that we talked about. They're falling behind on productive capabilities. Now the stock has had a pretty rough year once again. So what is the current issue? What are the big concerns right now facing Intel? You bet it's a few things. So you're right. With the day they hired Pek-Elsinger, I think the stock was up like 20%. And I get it. Like, I like Pat. Pat is the guy they probably should have hired two years ago, right? And so he's the right guy for the job. Now, that being said,

he's not a magician either. And he has to play the hand that he's been dealt. And the hand that he was dealt is pretty tough. And I think I said this the first time it was on, but the problems that Intel's having, they didn't just develop, like, over the last couple of quarters. These have been 10 years in the making. And they're not going to get fixed in a couple of quarters. They will probably take five or 10 years to fix. And so they've done a few things, actually, since we I think that was last November when we chatted, and so they've done a few things. Number one, we talked a lot about their difficulties at seven nanometers. So they're claiming that seven nanometers is now fixed. They're claiming the issue was they weren't using EUVE, what's called extreme

ultraviolet lithography, which is a very advanced type of patterning technology to make very small features. And I guess that is their public stance on what the solution was they're going to be much bigger using EUV, and that's going to fix the problem. It's also good. by the way, clear probably why they didn't want to use it. It's very expensive. Their CAPEX this year is going to end up to something like $20 billion, and that's not foundry. That's their own core business, but they don't really have a choice. So they claim that's fixed.

They've also now, I don't know, a month, maybe two months ago, they put out a road map. And this road, they did a few things. Number one is they adjusted their nomin. Remember we talked about the node nomenclature and how like Intel nodes didn't correspond to TSMC nodes in terms of process. So they've changed their nomenclature to align. Okay.

So what they were talking about seven nanometers before, That is now Intel 4. Okay, fine. Whatever. I'm not going to knock them because they weren't gaining anything by being a pedantic about it. So, I mean, okay, fine. But then the other thing is they said they're going to do kind of like five evolutions in four years. They're going to do, they're doing 10 now and they're going to go to, I think, I can't

remember what it's called 10 enhanced super fin or something. And then they're going to go to Intel 4 and then Intel 3, which is like an enhanced 7. And then they're going to go to something called 20A. And 20A is like in back in like 2024, 2025. And 20A technology will have a couple of new innovations. It's going to have something called a gate all around transistor structure, which we can talk about what that is, but it's a new type of structure on the transistor to get higher performance and everything.

And so that's one thing. The industry, by the way, is wrestling with that technology. Samsung has at the roadmap. Intel's now putting it on the roadmap. And they're going to have something called backside power, which again we could talk about. I'll have some innovations there. And what they're basically saying is by that timeframe,

by that sort of like 2024, 2025 timeframe, when we introduce this 20A process using this gate all around architecture, we will have caught up and then surpassed TSM at that point. So they put out this roadmap that says in like four, four or five years, we think we can close the gap with TSM and 2025 and beyond we will surpass it. So a couple of things with that. Number one, even though it sounds good, it on paper it's a very aggressive road map like I said it's sort of four iterations in five years

and their they're you know trajectory and execution over the last like decade you know this would be a big change so more power to that they can do that but the problem is like even if they can execute on it and it's not a done deal they basically came out and told everybody we're still going to be behind for the next four years right so the window is still for everybody else to attack us is still open for at least like four years so that was the was the dark side of what they were saying so even if they execute on it perfectly the window is still open for four years or more. And if they don't execute on it, they're screwed. But he had to come out and put a credible, like a semi-credible, like statement out that this has we can close. They got because his alternative was to lay down and die.

Right. Because if they can't catch up by that 2025 or whatever, they'll never catch up. That's it. Right. So that's one thing that they're doing. The other thing they're doing is they're making more use of outsourcing. And again, when they're doing a lot of stuff in-house. They're doing some more stuff at TSM, and so they're booking capacity, I guess three nanometer capacity at TSMC. And so we'll see what kind of products they make.

They're actually doing some graphic stuff over there now and they're going to be doing some other things. And then thirdly, they're going to be focusing on specialized packaging, what's called chiplets, where I'm disaggregating the chip into various types of functionalities and making each of those smaller pieces,

potentially different places, and then combining those together. And this is something where Intel thinks they have an advantage. And they've got very good, this is packaging. They've got very good packaging technology, but so do others. And so we'll see how all this works out. So that's sort of the plan, right? And then they're going to be building out all this foundry stuff. And to do that, they're going to be getting subsidies, right? So this is the other thing. So that in general, the strategy like simplified is to try to stabilize the roadmap, try to put out some sort of credible path to catching

up, and thirdly to beg for money from the government. Which is not a dumb idea, right? If there's any time right now to go get money from governments to build out sending to their capacity, now is the time. A lot of short daily news podcasts focus on just one story. But right now, you probably need more. On Up First from NPR, we bring you three of the world's top headlines every day in under 15 minutes because no one's story can capture all that's happening in this big, crazy world of ours on any given morning. Listen now to the Up First podcast from NPR.

are. So this is something that the Biden administration has been, you know, specifically calling out and targeting the idea of making the U.S. stronger in semiconductor manufacturing. Like, how much of a difference could that make to future supply demand balance? Well, they probably, I mean, right now it's $52 billion, I think, is the number on the table, which frankly is a rounding error. Again, TSMC is talking about one single company. They're going to go spend 50 billion. sorry, $100 billion over three years, one company. So the U.S. right now is saying $52 billion, which is manufacturing and R&D and a bunch of stuff

over like a five-year period for the entire U.S.-based industry. So, I mean, it's kind of gives you an idea of the scale. It's probably not going to make much of a difference. But it's a start. You've got to start somewhere. And then, like I said, we'll see, like, how they get spread out between Intel and Samsung and TSM. It's funny, you know, Intel, like they did a paid article in Politico a little while back where, I mean, Gelsing was

basically saying don't give the money to foreign companies, give it to us. Because like TSM, like they're going to build it. They're building this factory in Arizona. It's five nanometers. By the time they bring it online, that's going to be lagging edge. And they're going to leave all the leading edge IP offshore. And so you should give the money to us effectively with saying that, which is, I personally think if we're really interested in building up the U.S.

like supply chain in semis, you need to make it resilient. I don't think you can bet on one company. I think you need to spread the wealth around. And guys like TSM and Samsung, they already have customers. Like, Intel's talked about. They're going to work with quality. Alcom and some others, like in like five years, like maybe we'll see, right? TSM and these other guys have customers today.

So I'm of the opinion. It'd probably be better to spread that wealth around. And we'll see if $52 billion is where it stops. Like, I don't know. Like I think if we really want to make a difference, we need more. But it's a start. So we'll, you know, we'll see. You know, there's some other issues with Intel.

We talked about the stock. I mean, this strategy sounds credible. So you could ask us why has the stock been like such a dog? The problem is like he's, you know, they've been running around like trying to drum up excitement for what this vision could look like, and I get that. But I'm of the opinion they're about to, they have to drop a hammer on us because you start thinking about this, this stuff, like what's going to, what's going to happen in terms of capital and capital intensity is going up, free cashos going down, gross margins next year are way too high. They need to come down.

They're talking about, they think PCs are going to grow next year. I personally think that they're saying that because they're screwed if PCs don't grow next year, right? I think numbers need a need a fairly sizable reset. And we've got, they've got an analyst day coming up in November when everybody's sort of widely expecting them to give us like the model, like what is the vision for what this thing looks like during the transition as we go from where we are now to where maybe we're going in five years or whatever. And people are really hesitant to do anything with the stock in front of that analyst day because we don't actually know how low things like gross margins and free cash flow and everything are going to go. But you could make the case that they need to go,

that they need to go a lot lower than they're getting modeled right now. So this sort of brings back, you know, even prior to, so, okay, during the last 18 months of COVID and so forth, chip industry demand has been extremely strong, and that's contributed to the tightness. And one of the themes that was discussed even prior to COVID was that the chip industry seemed to be getting less cyclical overall. And so, you know, for years, the 80s and 90s, I think chips were seen as this classic boom bust industry. And then for years going, you know, I don't know, the last decade, there's just been the secular growth in semiconductors overall because they get put into like everything from cars to refrigerators to computers, to PCs, to chairs and so forth. Is there any, is that fundamental theme expected to continue that regardless of what happens, the secular growth of chips is going to continue? Or are there concerns that at some point some of these boom bus cycles might return? I don't think that those two things are mutually exclusive.

And so I'll give you my own view. So number one, let me put the near term cyclical concerns aside for a moment. Long term, I'm actually more positive on semis coming out of COVID than I've been in a long time in terms of that secular growth thing. I am. And look, you know, the sector, I mean like last year was $440 billion. And off the 2020 base, even if semis just grew 5% a year over the cycle, you'd have a trillion dollar industry by 2035. And by the way, this year semis are probably not going to grow 5%. They may grow 25% this year off of that 440%.

And if they did, we'd get pretty close to $550 billion this year. And so we'll see what things do, right? But you're going to have a lot of growth this year. And so even if we go into a downturn next year, who knows? Maybe we could, right? Sure. You think about that runway, we're likely still to be well over $500 billion would be, my guess, even in a downturn scenario, which is much higher than we were.

And long term, this idea of like is the industry someday a trillion dollar. I think absolutely it will be a trillion dollar. Whether it's 2030 or 2035 or like what have you. I think a lot of the trends were seeing are real. This whole idea of content increase and digitization and everything else, I'm a firm believer. And more so than I have been. So I'm pretty positive long term on semis. Now, the near term cyclical, this is kind of interesting because you kind of mentioned

like semis used to be really, really cyclical and then they got less cyclical. What we used to see, we used to see these big supply cycles where, like, supply would exceed demand and pricing would follow when you get these really big swings. And the last really major supply cycle we had was the tech bubble in like 2000, 2001. Mostly since then we had what we would have called inventory cycles. Again, semis are the back of the supply chain and you get fluctuations in demand that can propagate backwards and you get bigger swings. And usually, an inventory cycle happens when like the customers of the channel starts to bleed or build inventory. and so they they reduce purchases of semis for a little bit and they bleed out inventory or they increase purchases and they build inventory. And so semis overship or undership demand.

But it's much shorter. They tend to last a quarter or two or three and they show up as kind of like resets. You sort of reset the base either up or down rather than these big swings. You look what's going on right now. We're having a good old fashion supply cycle right now. We haven't had one of these for a long, long time. So I don't know how this Aladdin. Like historically they tend to end badly, you know.

And so we'll see. Although I could also argue that, like I said, maybe some of the secular drivers of the industry are, and they, you know, if you go back to when we used to have big supply cycles, I would argue the secular growth in the industry was pretty strong back then too. So maybe you have that to try to offset it somewhat. But, I mean, we've got a supply cycle going on right now. We haven't had seen one of these outside of memory in a long, long time.

But you can believe that and be nervous about the near-to-medium-term cyclical and still have a positive long-term view in terms of, like, secular growth of the industry. They're not mutually exclusive. Stacey, any other big sort of thoughts, things that we should watch for next? I mean, you mentioned the Intel Analyst Day. You know, like, what else is on your radar? We should do like a whole one of these calls.

Somebody just probably on semi-cap on wafer fabrication equipment. That's a whole other part of the industry, right? Yeah. And I want to do one on, you know, the one that we never did in the original series was the ASML. Uh-huh. And so maybe we got, at some point, we got to do that one, too. That's like the one big company we've never really done. I would be thrilled.

Yeah, I would be thrilled. Again, I don't cover ASML. I have a colleague of mine that is, but we can talk about it. And like the semi-cap industry in general, again, it's even farther back in the supply sense. You do get, you get cyclicality there. But again, like that's another one of these areas, like long-term, I'm like that, that long-term secular story on semi-cap, capital intensity going up, which has implications

across the value chain, I think is hugely interesting. That'll be the next one we do. You let me know when I'll be there. We can talk about it. Sounds good. Stacey, thank you so much for coming back on Adla. You bet. My pleasure.

Anytime. That was great. Thanks, Stacey. Tracy, I always like talking to Stacey, obviously. You know, I think one of the interesting things that, you know, he pointed out, which I hadn't really thought of or appreciated before, was like just how far deep in the cycle the chipmakers are and how little visibility they have into end demand.

And so on top of everything else, you can really see why so deep into this crisis, things aren't really smoothed out yet. Yeah. And also, of course, now you have this issue of double ordering and how much of the demand is real versus people stockpiling. It feels like that's just going to cloud it up and make it much more difficult in the future. And, you know, it kind of reminded me of, this is very, like, geeky, odd lot stuff, but it reminded me a little bit of the corporate bond market where when there's a big, sale of like a bond that everyone wants to get. People normally put in like padded orders so that they can get quite a lot. So, you know, they might not actually need 500 million or something like that, but they'll put in a massive order hoping that they get some of that. You know, I kind of

wonder like maybe that started because there was one really hot deal and then it just became normal behavior in the corporate ball market. I kind of wonder if this is a permanent shift for customers who are ordering semiconductors. Yeah, you should write about that, Tracy. That's a very quintessential Tracy analogy, the connection between the bullwhip effect and semiconductors and book building in the big corporate bans. You got to write that put. Maybe I will.

I think the odd, like, I'm imagining the Venn diagram of people who read about semiconductors and corporate bond market. It'd be a quintessential Tracy oddlots post. But no, I mean, I did think that. That was really interesting. Also, like, this idea that, like, you know, there used to be free cancellations or the idea that, you know, and also that the manufacturers then try to game it the other way so that if, you know, if a client makes a $100 million order, they might only make $50 million on the assumption that, oh, this is really just a $50 million order

in disguise. And so, yeah, you can see why that on top of everything else. And then, of course, the emergence of the Delta variant, the fact that chips crossed so many borders during their manufacturing just for like the production itself and packaging and put it in component. I'm not surprised or recording this September 13th that we're still seeing so many issues. I mean, one thing I will say is that I think the virus situation is improving in Malaysia. So maybe that will help and maybe at a minimum as these sort of Delta outbreaks start getting under control in Asia. Hopefully that'll allow people to see how much of the current tightness in the market is driven by COVID disruptions versus the bullwip effect that's been ongoing

for like more than a year now. Yeah, exactly right. Well, we definitely got to get Stacey back for the semi-cap ASML episode and, you know, this story is not over yet. No, it definitely isn't. All right. Shall we leave it there? Let's leave it there. This has been another episode of the Odd Lots podcast. I'm Tracy Allaway. You can follow me on Twitter at Tracy Allaway. And I'm Joe Wisenthall.

You can follow me on Twitter at The Stallwork. Follow our guest on Twitter, Stacey Razgen. He's at S. Razgen. Follow our producer, Laura Carlson. She's at Laura M. Carlson. Follow the Bloomberg head of podcast, Francesca Levy, at Francesca Today. And check out all of our podcasts at Bloomberg under the handle at Podcasts. Thanks for listening.

You can get the news whenever you want it with Bloomberg News Now. I'm Amy Morris. And I'm Karen Moscow here to tell you about our new on-demand news report delivered right to your podcast feed. Bloomberg News Now is a short five-minute audio report on the day's top stories. Episodes are published throughout the day with the latest information and data to keep you informed. Yes, there are other products like this from a variety of news organizations. But they usually rerun their radio newscasts through. the day. That's not what we do. We create customized episodes that can only be heard on Bloomberg News Now. And we don't wait an hour to publish breaking news. When news breaks, we'll have an episode up in your

podcast feed within minutes. So you're always getting the latest stories and developments. Get the reporting and the context from Bloomberg's 3,000 journalists and analysts we're all over the world. Listen to the latest from Bloomberg News Now on Apple, Spotify, or anywhere you listen.