U.S. equity market returns over the last several years have been driven by the performance of a narrow group of large cap information technology and Internet stocks that have come to be known as the FAANGs, for Facebook, Amazon.com, Apple, Netflix and Google (now known as Alphabet). Our investment teams have shared concerns about the crowding effect that the outlier performance of this group has created. While these companies possess unique business models and maintain strong long-term growth potential, we believe more attractive current opportunities may be found in looking beyond the FAANGs to other areas of technology, such as semiconductors.
Select Chip Stocks Could Buck Late-Cycle Pressure
Mark Andreessen (founder of Internet pioneer Netscape) once remarked that “Software is eating the world.” This has certainly held true as Amazon.com, Facebook and Google have disrupted several industries. However, it is also true that the tremendous growth in software functionality has only been enabled by tremendous growth in semiconductor performance. Historically, semiconductors have benefited from the power of Moore’s Law, the 50-year-old prediction by Intel co-founder Gordon Moore that the number of transistors on a given circuit would double every 18–24 months or double approxinately every two years.
That remarkably prescient forecast has driven tremendous economic value for consumers, as evidenced by the fact that the average smartphone now has more computing power than the entire initial Apollo space mission. However, for many years much of that value wasn’t extracted by the semiconductor industry, as the pace of innovation led to a “race to the bottom” mentality, with higher volumes being offset by price declines. In recent years, however, Moore’s law appears to be slowing, with leading edge semiconductor manufacturers running up against both physical limitations (the size of semiconductor gates is now just a few atoms in width) as well as economic ones (after falling for decades, the cost per transistor in some cases appears to be rising). Counterintuitively, this end of Moore’s law dynamic has led to a more favorable industry structure, with better pricing power for semiconductor suppliers and more stable margins. On top of that, semiconductor companies are rationally adjusting to lower long-term growth rates, which are leading to lower investment profiles and higher margins across the industry.
However, while we continue to be positive on the long-term drivers of the industry, semiconductors remain susceptible to inventory cycles, which can impact their near-term fundamentals. We are cautious in our overall view of the semiconductor cycle, although we believe there are several semiconductor stocks that have a combination of defensive, counter-cyclical, and secular growth opportunities that will allow them to outperform even in the event of a cyclical correction in the space.
Historically, semiconductor stocks outperform early in the business cycle (such as 2003–2004 or 2009–2010) and again later in the cycle, after undergoing cyclical corrections (such as 2006 or 2014). The Philadelphia Semiconductor Index (SOX) has outperformed the S&P 500 Index by over 40 points over the past 12–18 months (Exhibit 1).
Exhibit 1: Semiconductor Outperformance Follows Cyclical Patterns
This outperformance has been driven by substantial growth. We estimate that the overall semiconductor market will grow 17% in 2017, well above its long-term average of 3%-4%. While we share in the excitement of multiple new technology trends that can drive excess growth (artificial intelligence, the Internet-of-things, or self-driving cars), they still represent a small part of a very large and mature industry. Looking back, it’s clear that even during large product cycles, such as the explosive growth of smartphones or cloud computing, the overall semiconductor market has still grown in line with global GDP (Exhibit 2).
We see that trend persisting, and similar to 2010, when the semiconductor market massively outgrew global GDP, we see growth rates falling significantly in 2018 and 2019 as customers digest excess inventory.
Exhibit 2: Semiconductor Growth Closely Correlated to Global GDP
|Global Nominal GDP||10%||-5%||9%||11%||2%||3%||3%||-6%||1%||4%||5%|
So how does one position for a cyclical downturn? Semiconductor companies tend to have a lot of operating leverage given their large fixed investments, chiefly in engineers and manufacturing facilities. As a result, when sales growth turns down, it tends to drag down margins as well. We believe that “fabless” semiconductor companies such as Qualcomm provide a more defensive earnings stream, in Qualcomm’s case particularly when factoring in their large licensing business.
We also see some growth pockets that we believe can be counter cyclical. For example, Apple is expected to launch three iPhone models in the next few months, which we believe will drive an upgrade cycle across the Apple installed base. Broadcom, a supplier of RF filters in the iPhone, is expected to have 40% more content in the new models compared to the iPhone 7/7+, which should allow the company to sustain growth even in the face of an industry downturn. We also see Broadcom’s upcoming acquisition of Brocade Communications Systems, and the company’s new initiative to return more cash in the forms of dividends, as being supportive of the stock.
Finally, in the semiconductor capital equipment space, we are positive on ASML. Our thesis here is based on customers adopting the company’s new Extreme Ultraviolet (“EUV”) technology, which can fit even smaller transistor sizes on semiconductors. Given ASML’s 100% share in this new technology, and the enormous benefits that continuing Moore’s law provides, we believe ASML will be able to outperform its equipment peers.
Overall, we see near-term cyclical risks for the semiconductor sector, but believe that the fundamentals of the industry have markedly improved, which makes us constructive longer term.