Chip Winter or Time to Buy?

Data is everywhere, but is it ready for consumption? Morningstar developed the language of global investment data so you have the right ingredients to help you shine. Morningstar, where data speaks.

I don't know that you have to buy today, but if I look at the industry multiple, last Friday we hit a decade low in terms of PE multiples for semiconductor stocks. And so to me, It feels like the semiconductor stocks have reflected most of this weakness, you know, in terms of fundamentals.

Hello and welcome to the Barron's Streetwise Podcast. I'm Jack Howe, and the voice you just heard is Quinn Bolton. He's a semiconductor analyst who says that the stocks look mighty cheap. We'll hear about his favorite names and the one thing he thinks needs to happen before the group bounces back. We'll also talk with a semiconductor historian about how the US-China chip battle is likely to play out. Listening in is our audio producer, Meta. Hi, Meta.

Hi, Jack.

I was at the airport last night and I realized I left my phone charger in the hotel room and then I had to pay $40 for one of those airport store ones, which is wasteful because I'll never use it again. I have like a dozen faster chargers at home. Yeah, that's a bummer. And so I asked the woman at the counter, is this the cheapest charger? And she says, you could just use mine until your flight. And she pulls a pink charger out of her purse and just hands it to me, basically ambushes me with niceness. So I reflexively reach for cash. I figured I'd give her $20, which would reward her generosity and allay any worry that I assumed that she would have about me walking off with her charger. And I'd still be $20 ahead by not having to buy one. I said, "I'll tell you what—" but she cut me off before I could get my hand out of my pocket, and she said, "No, just pay it forward." And I said, "That's so nice. Is this your store?" Because I was working through all the permutations in my head, and I thought maybe she owns a long-term license for the store, and she's thinking, here's a business traveler who looks like he's no stranger to airport snacking. For the wholesale price of a charger cord, I can exploit his urge to reciprocate, and I can lock him into years of lucrative panini sandwiches and KitKat sales. But she said, I just work here. So I left the store, and I sat within view of her so that she could see that I wasn't leaving with her charger. But she didn't look over once. Wait, so you stole the charger? I did not steal the charger. I didn't even report her to her manager for, let's be honest, cutting into charger sales. Meta, I looked for someone to be nice to the whole way home. I thought maybe this driver, but he was terrible. He was reckless, had his high beams on the whole time, and I'm definitely not paying it forward to a hostile high beamer. So now I'm trapped in this kindness Ponzi scheme. I have to go out in the world today to find someone to be nice to. A man with my schedule, I wasn't planning on interacting with humanity again until probably later in the week. Sounds like AirPod Lady really screwed you. Yes. Let that be a cautionary tale for listeners. If there's one thing I know about Ponzi schemes, it's better to get in early than late. Better to launch your own kindness Ponzi scheme right away. Stick it to them before they stick it to you. That's brutal. Well, you have to be. Now then, Chips or semiconductors. They go by both names. Until recently, they were widely viewed as having outgrown boom and bust cycles. Back in the 1980s and '90s, demand was driven by personal computer sales, which were driven by the release schedule for Microsoft Windows. So there'd be a huge surge in chip demand and the industry would blow out production, and then demand would dry up and chipmakers would report big losses. Several things changed in the 2000s and beyond. The rise of smartphones created a massive new chip market. Soon silicon—that's an even cooler way of referring to chips or semiconductors—began gaining ground in cars and factories and spreading to previously dumb devices like refrigerators, even light bulbs. Computing power shifted from homes toward the cloud, where chip demand for things like data analytics, and artificial intelligence has seemed endless, and chip companies consolidated. If you invested in a broad S&P 500 index fund over the decade through last year, you multiplied your money 4 times, which is incredible compared with history. But if you invested in the Philadelphia Semiconductor Index, sometimes called the SOX for its trading symbol, SOX, Over that same stretch, you multiplied your money 12 times. Now look at this year. The SOX was recently down 43% for the year. That's double the decline for the broader S&P 500 index. Chips are getting hammered. Big swings in demand are back, but for two new reasons. First, the pandemic created a surge in demand for computing devices and a shortage of chips to power them, and And now we're seeing the reverse, better chip supply and fading demand. Second, in recent weeks, the US government has sharply scaled up export controls designed to deny China key chipmaking equipment and expertise. So sales to China are plunging, and that leaves investors trying to figure out how bad things could get for the industry and how much of the likely damage is already priced in. Is this chip winter or just a chip hiccup?

If I look at the industry multiple, last Friday we hit a decade low in terms of P/E multiples for semiconductor stocks. And so to me, it feels like the semiconductor stocks have reflected most of this weakness, you know, in terms of fundamentals.

That's Quinn Bolton, who covers semiconductor stocks for Needham and Company. He points out that the shares have gotten cheap. The stocks recently traded below 14 times forward earnings projections versus an average over the past 5 years of more than 18 times. Quinn says that a lot of the bad news for chips appears priced into the market, but also that there's no rush to buy. And he recently wrote that there appears to be, quote, little light at the end of the tunnel in the near term. I asked him about that.

I think what needs to play out before we can see a sustained rally is that the sell-side estimates have come down for semiconductor companies as well as some of the capital equipment companies. But you haven't yet seen the companies across the board taking estimates down. And I think that's probably the data point we need to see, that, that most semiconductor companies cut their forward guidance to lower estimates. And at that point, I think investors are going to be much more comfortable stepping in and buy semiconductors.

After I spoke with Quinn, a company called Lam Research warned investors that it would lose 2.5 $5 billion in revenue from U.S. export restrictions, and its shares the next day traded higher. Lam Research is named for David Lam, L-A-M, an engineer who was born in China and who worked for companies like Xerox and Hewlett-Packard and decided that in order to keep cramming more computing power into smaller chips, the industry would need better machines for etching circuits into silicon wafers. So we started Lam Research in 1980, and today it's a major Silicon Valley employer and one of a handful of companies that make the machines that are used to make computer chips. Others include California's Applied Materials, Japan's Tokyo Electron, and a Dutch company called ASML. The fact that Lam stock jumped on bad news suggests that investors expected even worse news.. It's a data point that should get bargain hunters watching what happens next. Quinn expects earnings estimates to fall, but says they don't have to be done falling for the shares to rebound. Just look back at a supply glut in 2018.

But if we look back to the 2018-2019 downturn, which was largely inventory driven, we think '22-'23 plays out very similarly where you saw that the stock sell off and once estimates started to correct, you saw the multiple beginning to expand and the semiconductor index in 2019 rallied almost throughout the year, even as estimates were coming down. And so we're starting to see the initial cuts. You've seen that in some of the PC and semiconductor names. My guess is by the end of 2022, we'll start to get the capitulation across the rest of the semiconductor industry. And so as we get into year-end, I think that's going to be an interesting time to start looking at this group. And adding to positions.

I asked Quinn for a couple of his favorite stocks for investors who might think about buying sooner rather than later. He likes a mid-sized company called MaxLinear, ticker MXL, which is active in infrastructure. It trades at 7 times this year's projected earnings, which Quinn calls way too low given the quality of the business.

We think the company is gaining share across Wi-Fi and cable and fiber segments. And so we think those share gains can help buffer the company from this broader industry slowdown.

Quinn also likes slightly larger MACOM Technology Solutions Holdings. If you ask me, they could have fit another solutions and at least two more holdings into the name. The ticker there is MTSI, has a relatively new management team, trades at 18 times earnings. Quinn says earnings there are less prone to an industry downturn.

We see a number of new products coming to market that can help buffer that company's sales from the broader industry slowdown. We also like MACOM because about almost half now, about 45% of its business comes from the defense business. And we find defense spending tends to be not, not influenced by the broader economic cycles.

For investors who want to make a broad move into chips later this year, if we get that capitulation sell-off that Quinn mentioned earlier, there are exchange-traded funds like iShares Semiconductor, ticker SOXX. The long-term outlook for semiconductor demand, Quinn says, remains bright.

If you look at the amount of processing you need to compute artificial intelligence, you just need more processing, more memory, um, and importantly, to get better performance out of those processors, the die sizes are getting bigger. And so I think continued strength in the data center just will drive demand for silicon across both advanced processors and memory, and that's going to be a good long-term driver for the industry. And then I think the other clear area of strength is automotive and the trend towards autonomous vehicles and electrification, and the amount of semiconductor content going into cars is clearly going up to support those trends, and, and those trends will be playing out over literally decades.

So if the choice is Chip Winter or Chip Hiccup, I guess it might feel like winter in the months ahead, but might look like more of a hiccup decades from now. Meta, if I make a surprise career change to weatherman, I'm definitely calling myself Chip Winter. Not Chip Hiccup? I don't think so. If I make another surprise career change from weatherman to rodeo clown, I'll do Chip Hiccup. Now coming up, we'll circle back to China and talk with a chip historian about where the battle for semiconductor supremacy goes from here. That's next after this quick break.

Data is everywhere. When orchestrated properly, it sings. At Morningstar, we analyze and enrich data, making it actionable and powerful for you. Morningstar, where data speaks.

Welcome back. Meta, we're talking about chips or semiconductors or silicon, and sometimes you'll also hear integrated circuits or processors. These terms are used interchangeably, although they don't quite mean the same thing. Semiconductors like silicon are used to make the wafers that hold the integrated circuits that form the chips. And in the UK, they're called crisps. Exactly. Now let's quickly circle back to China. Where does this showdown on chips take us? I spoke recently with Chris Miller, a history professor at Tufts University and consultant on geopolitics. And the author of a new book called Chip War: The Fight for the World's Most Critical Technology. He says that investors think of chips in terms of smartphones and personal computers and devices and data centers, but governments think of them in terms of weapons.

Right now, the US government is trying to drive a wedge between what China has access to in terms of chips and what the US government has access to in terms of chips. Doing so, in the US government's view, requires not only severing military research or militarily relevant chips, because actually it's really difficult in most cases to distinguish what's military versus what's non-military. Most ships are both.

That reminded me of something I read about video game consoles and North Korea. When the PlayStation 2 launched more than 20 years ago, the Japanese government worried that the machine's sophisticated chips for modeling 3D worlds could be used for real-world applications, like missile guidance. So it made Sony apply for a special export permit. A story a couple of years ago on PC Magazine recounted a history of fears about weaponizing gaming consoles. There was a dubious report about Saddam Hussein hoarding PlayStation 2s, and a quite real project at the U.S. Air Force Research Laboratory that turned more than 1,700 PlayStation 3s into what was then one of the world's most powerful supercomputers. Okay, back to Chris.

And so the conclusion has been that an effort to just stop chips going to the Chinese military has failed. And as a result, the US wants a greater bifurcation of the chip supply chain. So there's a non-China or an allied chip supply chain that involves the US, Taiwan, South Korea, Europe, Japan, and then there's what gets into China, which is several generations behind. And so that's going to be the US policy, I think, for the foreseeable future. It's going to have a major impact on Chinese firms, but also on US firms, which for a long time have treated China as just another market. And now they're coming to terms with the fact that it's not just another market. It's another market that faces huge restrictions for these military reasons.

China spent vast sums to subsidize its semiconductor industry,, and a lot of that money went to US companies that supplied advanced chip-making machines. So now those same companies stand to lose significant revenue in China as the announcement by Lam Research has shown. But who holds the stronger hand in the China-US chip war?

If you look at the tools needed to make chips, the machine tools that are able to move materials on silicon at almost an atomic level, these machines are among the most precise, heavily engineered, and expensive tools made in human history. Just for example, the most advanced extreme ultraviolet lithography tools, without which you cannot make an advanced chip, cost $150 million apiece and are made by a single company in the Netherlands. So there's a number of places in the chip supply chain where there are just monopoly positions that have been in place for decades.

The particular company Chris is talking about is called ASML. We've talked about it in the past on this podcast. When we spoke last year with the CEO of Intel about his efforts to get back on top in chipmaking, securing machines from ASML was a key part of his plans. Chris says it's not just difficult for China to secure chipmaking machines for now, it's impossible. Making its own could take a decade and many tens of billions of dollars. The US ban extends beyond US companies to overseas ones that want to do business with the US?

If you look at where most of these critical companies are located, they're largely in the US. There's ASML in the Netherlands, though ASML's production processes rely on US technology and take place partly in Connecticut. There's Tokyo Electron, as you mentioned, in Japan. And then there are a number of key suppliers to these companies, like Zeiss in Germany supplies the flattest mirrors ever produced in human history that are inside of ASML lithography machines. That's a German company, but it has some of the critical technology that goes into the Dutch machines. All of the companies, all the suppliers that are really critical are based either in Japan or in Europe. They're not in the US. So it really is US allies that dominate this.

That makes it sound like the US has a strong position in the chip war, and the government is spending tens of billions of dollars to incentivize chip production at home. But Chris says not to expect the US to dominate that business anytime soon.

And most of these funds will go to subsidizing leading-edge production, so leading-edge logic, leading-edge memory in the United States. And so I think in 5 years we should expect we're going to have a number of new ultra-advanced facilities in the US run by TSMC, run by Samsung, run by Intel, run by Micron on the memory side, which will somewhat reduce America's reliance on importing advanced chips from abroad, from South Korea, for example, and from Taiwan, not completely. And a completely onshore supply chain is really a fantasy at this point, it would cost vastly more money than we're willing to spend. But the CHIPS Act will reduce the concentration in Taiwan to some extent.

Chris mentioned Taiwan and TSMC, or the Taiwan Semiconductor Manufacturing Company. Which he calls one of the most impressive companies in world history. They've pushed chip manufacturing technology forward. We ought to admire them, Chris says, but the rest of the world has reason to be nervous because 90% of the world's most advanced chips today can only be made in Taiwan. If you wondered why there was so much attention paid to the U.S. Speaker of the House visiting Taiwan in August, This is part of why China views Taiwan as part of China, even though it doesn't exercise control over Taiwan. The U.S. stance is sometimes referred to as strategic ambiguity. But what's less ambiguous is that the U.S. has also agreed to supply a vast stockpile of weapons to Taiwan to defend itself in the event of an invasion from China. For Taiwan, China has long been a key customer. A survey in Taiwan last year found little support for either full independence or full unification with China. Overwhelmingly, respondents said they'd prefer to keep things as they are. The U.S. government, not so much, at least not with regard to Taiwan's lock on the advanced chip market.

The concentration is enormous. Most of that is TSMC, and there are risks. To the fact that this is geographically concentrated on one island, earthquake-prone, and more worrisomely than that, war-prone, given the fact that China's military is growing in power every year.

Thank you, Quinn and Christopher and Airport Lady and Chip Winter and Chip Hiccup. And thank you all for listening. Meta Lutzoff is our producer. A newly wedded Jackson Cantrell will be back producing next week. You know, Meta, You got married and had a baby since this podcast started in early 2020, and now Jackson got married. I take full credit. I'm basically an audio household formation catalyst over here, and you're all welcome. You're a financial Cupid. Now I feel like you're just calling me chubby. See you next week.

Data is everywhere. When orchestrated properly, it sings. At Morningstar, we analyze and enrich data, making it actionable and powerful for you. Morningstar. Where data speaks.