Description

Introduction

At least in the stock market, the semiconductor sector has been on fire. Since the recent bottom in late 2018, the SOXX index has appreciated over 175%. The sector’s equity market performance is all the more puzzling if you consider the fact that the sector’s revenue actually declined 12% in 2019, with most large diversified semiconductor companies guiding below consensus virtually every quarter Q4 2018-Q1 2020 such that industry revenue estimates declined consistently the first 15 months of this 25 month and counting rally. Basically, the industry spent 2019 in a demand shortfall and inventory drawdown, only to be hit again in Q1 2020 by the coronavirus lockdowns and pause in industrial production. However, demand for goods that include semiconductors (computers, datacenter computing, telecom/networking equipment, home electronics, industrial equipment, etc.) held up better than in a normal recession due to the unique factors surrounding coronavirus pandemic and the low inventory levels heading in. Still, industry revenue only grew 5% in 2020, remaining below 2018 levels and even the rebound in revenue growth expected at the beginning of 2020 pre-coronavirus. Everyone on Wall St. expects the industry growth to pick up markedly in 2021 and companies to resume a beat/raise cadence or at least consistently meet expectations. Q4 2020 earnings reports indicate the industry has entered a strong cyclical upswing, and in parts is even capacity constrained.

Obviously, the scorching semi industry rally in the face of (until recently) poor fundamentals means the average semiconductor company valuation has shot through the roof over the last two years, reaching levels unseen since the dot com bubble. How do you invest in a cyclical sector that is (1) in the early stages of a cyclical upturn that (2) everyone paying attention is expecting and (3) has recently appreciated much faster than fundamentals such that (4) it has the highest valuations (and presumably highest embedded expectations) in recent history? I think you want to find companies with still-reasonable valuations, strong balance sheets, product cycles that will enable them to outgrow the industry and reasonably set analyst expectations so that you have several possible sources contributing to potential return- absolute growth, estimate increases, ability to return capital in dividends and buybacks, multiple increases and even the potential benefit of being purchased in an accretive M&A transaction. Looking across the semiconductor industry, I believe that Silicon Motion (SIMO) best fits the bill, and based on the three-year $1B revenue goal outlined in their Q4 2020 earning call, whose stock could reasonably double in the next two years without getting a crazy multiple or becoming a meme stock.

Company Overview and Recent Results

Silicon Motion makes flash controllers and solid state drive products. Flash controllers provide the interface and control functionality between the raw flash memory (mostly made by Micron, Samsung or Hynix) and the rest of an electronic system. The biggest markets for SIMO’s flash controllers are in PC/storage solid state drives and the mobile memory modules in Android phones. Note that in Apple phones the internal application processor has an integrated flash controller (so SIMO and its competitors do not serve the iOS market) but an integrated controller is not in current Android application processor designs or roadmaps. The company also makes controllers that go into flash USB drives and memory cards, although this is a shrinking part of their business. Finally, SIMO makes solid state drives (i.e. buying the flash and using its controllers and software for an integrated solution) for specialized industrial and data center usage, including for Chinese hyperscalers Baidu and Alibaba. Revenue breaks down to approximately 55% SSD controllers used primarily in PCs, 24% mobile storage primarily in smartphones, 18% SSD solutions and 3% legacy USB and memory card controllers. SIMO’s primary competition comes from Phison in Taiwan and Marvell as well as the internal development teams of the flash makers.

SIMO’s revenue performance has not been that much worse than the industry, but revenue had declined from a peak in 2016 due to losing a mobile memory customer. Basically, the flash industry goes through periodic interface standard changes which open up opportunities for design wins and losses. A few years back the most advanced flash made a transition to the eMMC standard, which was followed a couple of years ago by UFS. SIMO won the lion’s share of eMMC at Hynix, but Hynix went with an internal controller for UFS. SIMO’s mobile embedded memory business peaked in 2016 at $180M (35% of revenue), gently declined in 2017 and then shrank substantially in 2017 and 2018 as the Hynix-related revenue shriveled away to an immaterial part of revenue. At the same time, SIMO won Micron’s UFS business, causing mobile embedded flash controller revenue to bottom in Q1 2019 and return to high teens growth in 2020. During the interim SIMO has experienced steady growth in its core PC SSD controller business as virtually all laptop PCs and many desktops began using SSDs instead of spinning disk HDDs. The company has had up and down performance in its SSD solutions business with it flattish in 2018, down in 2019 and up substantially in 2020.

The company performed well overall in 2020, which makes its performance and low valuation puzzling. Top line grew by 20% overall, although due to strange comps the growth slowed as the year went on from 49% in Q1 (lapping their worst quarter in the mobile memory controller transition) to -6% in Q4 (lapping an anomalously strong quarter in Q4 2019). All three major product lines (mobile embedded memory, SSD controllers and SSD solutions) grew double digits for the entire year. SIMO was one of the few semi companies to come through the unexpected coronavirus-caused recession with revenue and EPS expectations for 2020 relatively unchanged over the course of the year, while its stock was slightly down against the SOXX going up 50%. There seem to be some non-fundamental reasons for this- the company is HQed in Taiwan so does not make it into all of the U.S.-based tech and semi indexes and despite the fact that it is a fabless IC maker it tends to trade with commodity memory suppliers. This disconnect has resulted in a value stock succeeding in a growth industry at around 2.5x forward revenue (with 50%-ish GMs), 7x EBITDA and 12x my 2021 earnings estimates with around $12.00 per share in cash on the balance sheet. Semiconductor companies in today’s market frequently sport multiples 2-4x these levels.

Silicon Motion typically pre-announces revenue results at the same time it gives its conference call information, so there was no suspense around 2020 results in its February 4th Q4 earnings call. However, the company uncharacteristically gave a robust three year growth outlook and stated its goal of $1B in revenue in 2023. The company also gave 2021 guidance of $650M to $750M in revenue for 2021, substantially higher than then-consensus of $590M, and further stated that its 2021 guidance was limited by foundry capacity, not demand.

Forward Business Prospects

I believe that SIMO is attractively positioned and each of its major business segments will grow in 2021 and beyond:

• SSD Controllers: In recent years SIMO has clearly gained share in PC SSD controllers and supplies the SSD operations of flash makers and third party assemblers. While 2020 benefitted from unexpected PC market strength, within PCs attach rates for SSDs continue to increase. Thus, even if the PC market slows down, I believe the PC SSD controller market will grow at at least 10% in 2021. SIMO has gained share in recent years but even if their share remains stagnant this segment should grow 10%+.

• Embedded flash controllers: This segment had declined three straight years before rebounding around 17% in 2020 as Micron’s UFS business ramp offset the decline from eMMC at Hynix. The largest end market for embedded flash (and controllers) is smartphones, and 2020 was actually a down year for smartphones due to coronavirus and turmoil at Huawei. 2021 should see a strong rebound in smartphone units, continued growth with Micron’s UFS business and the addition of another major flash UFS customer beginning in Q3. Thus, I think that the embedded controller business will be the largest % and collar contributor to growth in 2021, up around 31% or $40M.

• SSD Solutions: SSD solutions will likely be a lower contributor to growth, but SIMO is set to grow from one major datacenter program to three at its largest customer Alibaba, which along with steady enterprise/industrial demand should make this segment a steady grower.

Return Potential

I believe that SIMO is one of the few semiconductor stocks that could appreciate 100% without heroic assumptions or bubble-era valuations. Investors should benefit from several potential sources of return:

• Absolute growth: I believe the company will grow 25% or greater in 2021, faster than the overall semiconductor market after outperforming the semi market with 20% growth in 2020, and eventually investors will notice. Meeting the $1B revenue goal would mean growing 20%+ in 2022 and 2023 as well.

• Beating consensus: The company has a good track record of guiding conservatively, and indicated a high level of visibility on its 2021 guidance. Nevertheless, the St. consensus is $681M, below the midpoint of the $650M to $750M guidance. Similarly, the St. consensus for 2023 currently calls for just over $900M in revenue, short of the $1B goal.

• Returning capital: Much of the semiconductor world has stopped buying back shares given economic turmoil and high valuations, but SIMO remains cheap and has around $12.00 per share in cash with a repurchase authorization. The stock also delivers a near-3% dividend which has plenty of room to be raised.

All of the above has the potential to result in multiple expansion to something closer to the rest of the semi universe. A $125 stock (up 100%) would still only be around 4x revenue, 12x EBITDA and 19x earnings if the company can hit its $1B 2023 revenue goal (which implies something like $7 of EPS), still below industry average. Given the disparity in multiples SIMO would also be an attractive M&A candidate for someone like ON, although their 50%ish gross margin might limit their appeal to companies focused on increasing GM.

Disclosure: I and the fund that I work for are long shares of SIMO. We may buy or sell shares of SIMO at any time without notice.

Catalyst

• Revenue growth towards $1B goal in 2023
• Beating consensus estimates
• Raised dividend and potential repurchases