The chip industry has been one of the biggest investment stories this year.
The iShares PHLX Semiconductor exchange-traded fund (SOXX) is up 35%, despite the pandemic and economic shutdowns. Helping drive the gains: a wave of deals, spurred in part by the elevated stock prices of the acquirers.
We checked in with C.J. Muse, 50, Evercore ISI’s widely followed chip analyst, to see what happens next. After receiving his M.B.A. at Columbia University, Muse worked as an investment banker at Lehman Brothers before becoming an equity-research analyst in 2000, first at Lehman Brothers, then at Barclays, and since 2014, at Evercore ISI. Muse does his job well: He’s the top-ranked semiconductor analyst in Institutional Investor’s annual survey. He resides in New York City, where he plays tennis and takes care of his new dog when he isn’t tracking the semiconductor industry. Keep reading for more.
Q: Why has SOXX been climbing?
A: C.J. Muse: One of the most important factors to recognize, coming into 2020 and the pandemic, is that much of the semiconductor industry had corrected in the latter part of 2018 and all of 2019. During the global financial crisis, revenues for the industry dropped 40% from peak to trough, and then, within three to four quarters, were back to peak. In this cycle, we were 20% off the prior peak. That gave us confidence that the impact to semiconductor fundamentals would be muted. This year, semiconductor revenues are probably up 5%, which is pretty amazing in light of the high correlation with global gross domestic product.
Q: What kind of growth are you expecting?
A: Our base case is semiconductor revenue growth of 5% this year and 10% or even higher next year. The consensus estimate for next year is below that. Driving accelerated growth in 2021 and beyond is, first and foremost, the secular story with the cloud. We also have the 5G cycle. Not only will that lead to greater smartphone units, but there’s also an underlying content story. If I think about secular growth for the cloud, and smartphone units tracking down 10%-plus this year, and then going into the 5G cycle, I don’t know why smartphones can’t grow 10% to 15% in 2021.
The third leg of the stool is broad-based analog, with auto and industrial recovering after seven straight quarters of undershipping. Those three drivers are what’s going to enable and add growth in ’21. A number of drivers should enable continued relative outperformance. Covid-19 has exacerbated the digitalization going on across almost every industrial vertical.
Q: Is a second wave of the virus of concern?
A: Another wave and/or a slowdown in a reopening of economies would be a negative. At the same time, businesses and individuals have accelerated their adoption of the digital world. You’ll see relative outperformance because of how critical semiconductors are.
Q:Trade is a big issue for this sector. What does Joe Biden’s assumed presidential victory mean for China trade and semiconductors?
A: The election had a big impact on my group. Results from Qualcomm (QCOM) and the election definitely led to the SOXX going dramatically higher. U.S.-China relations will be less transactional and normalize a bit. There will be pressure for China to improve on intellectual-property theft and other concerns. But it does mean China offering regulatory approval of merger-and-acquisition transactions. The market was concerned that the restrictions on shipping equipment to Semiconductor Manufacturing International (SMICY) could move to all of them. That would be removed as a concern.
A stimulus package would be good for GDP and therefore semiconductor growth. Finally, the Chips Act to fund local U.S. semiconductor manufacturing will be a huge focal point. It’s in the Senate Appropriations Committee.
China wants to build their own semiconductor industry. This is a wonderful industry for the U.S., and we need our government to support it, like all the other foreign governments do. I would hope this is a mission-critical focus by the new administration.
Under a Biden administration, the big winners are equipment suppliers like Lam Research (LRCX), Applied Materials (AMAT), and KLA (KLAC). We could see pressure on Texas Instruments (TXN), with planned higher corporate-tax rates favoring foreign-domiciled chip makers like NXP Semiconductors (NXPI).
Q: M&A is this year’s story. Why?
A: One, there are only so many remaining high-quality assets. Scarcity value, despite the uncertainty around Covid-19, has played a role in the willingness to pursue M&A. Two, China is clearly looking to build its own domestic semiconductor industry and, looking out five to 15 years, will be a viable threat. Scale is critical for success. Three, stock prices moved up considerably in the past 12 months.
In July, Analog Devices (ADI) offered to buy Maxim Integrated Products (MXIM), using 100% stock. We weren’t out of the woods yet on Covid-19, and that was novel. Subsequently, we’ve seen Nvidia (NVDA) agreeing to buy ARM Holdings, Advanced Micro Devices (AMD) agreeing to buy Xilinx (XLNX), and Marvell Technology Group (MRVL) agreeing to buy Inphi (IPHI), all using their stock as currency. The probability of M&A being approved by Chinese regulators is now higher—good news for these transactions to close.
Q: Moore’s Law spurred growth in the past. But technological advances are getting expensive.
A:The driver of incremental leading-edge capacity is high-performance compute (chips used for the cloud). If you think about the business models for Google, Amazon.com (AMZN), and Microsoft’s (MSFT) Azure, which are competing heavily in the cloud, I don’t think they care about the cost of silicon because the return on invested capital is so vastly better than in the consumer market. That’s a very important distinction that will enable investments in leading-edge silicon, despite the fact that capital intensity is accelerating as much as it is. Other techniques are being adopted to get around the challenges of Moore’s Law, such as advances in packaging. Design also plays an important role. Nvidia, the leader in graphics processing units, is utilizing GPUs for artificial-intelligence machine learning. And Nvidia isn’t at the leading edge—they’re typically a node or two behind, but have the best product out there. So it’s not just transistors and transistor size that matter.
Q: Are there any companies left to buy?
A: If you look at the semi landscape, we now have five big equipment companies. U.S. regulators stopped Applied Materials from buying Tokyo Electron and Lam Research from buying KLA. So, semi equipment is done. The four major compute platform names in semis are Qualcomm, Nvidia, AMD, and Intel (INTC). The DRAM industry has consolidated to three names. In NAND, there are still six players after Intel’s NAND sale to SK Hynix (000660.Korea).
Elsewhere, we’re seeing two major analog players emerge in Texas Instruments and Analog Devices. We could see further consolidation among leading analog/mixed signal names like Marvell, NXP, Microchip Technology (MCHP), ON Semiconductor (ON), Power Integrations (POWI), and Monolithic Power Systems (MPWR). The most interesting growth areas in semi land are cloud, networking, and automotive, where Marvell is probably the most interesting candidate. After that, the more interesting question is if major players in one silo venture into another. Who might have an interest in M&A in general? Qualcomm has been noticeably silent. But Qualcomm’s CEO is viewed as a value investor. I suspect he’ll be very focused on price.
Q: What should investors own?
A: Nvidia is creating the AI platform. They’re dominant in training and AI for the leading cloud players. Less appreciated is their platform-based approach targeting all the major enterprise verticals. If they’re successful with it, there’s no doubt in my mind that they’re on track to increase earnings from around $15 next year to $25 by 2025. My price target is currently $600, but if that view is right, then this is a $1,000 stock looking out over the next three-plus years. (Nvidia recently fetched $532.)
NXP has one of the best exposures to automotive at 50% of revenue. Revenue will grow at a 12% compound annual rate. Plus, NXP is a content story, led by their positioning in radar, battery management systems for electric vehicles, and digital clusters for automotive. Next year, investors will start to expect $10 of earnings in 2022. We have a $160 price target (recent price: $146.) If $10 is real, our price target is probably conservative.
Q: Anything else you like?
A: Teradyne (TER) has a number of interesting levers. And we expect another strong year from Apple (AAPL). Teradyne has excellent leverage to Qualcomm, an auto-industrial story that will recover nicely in ’21, and an industrial-robotics business that will grow nicely next year. We think they earn $5 in 2021 and $7 by 2025, which supports the $110 price target today and $140 over time. (Teradyne recently traded for $100.)
Another is Micron Technology (MU). While 2020 was disappointing, with Covid-19 weighing on end demand, supply/demand should move into balance. Very likely you’ll see shortages in 2021, meaning multiple quarters of upward pricing trends for DRAM that will lift the stock. Our current price target is $75, so that’s 50% upside. The assumed multiple is pretty low, so we could see upside there, too.
Q: What about equipment stocks?
A: ASML Holding (ASML) is domiciled in the Netherlands, and isn’t impacted by U.S equipment embargoes. They have a monopoly with their EUV (extreme ultraviolet lithography) technology. Earnings are going to accelerate, and we believe 20 to 24 euros ($23.60 to $28.32) a share in 2025 is achievable. We have a current price target of €375, and long-term upside of €500, if our €20-plus earnings number is accurate. (ASML traded on Thursday for €352.)
|For more news you can use to help guide your financial life, visit our Insights page.|