Description

Silicon Motion (SIMO) is a decent company trading at an inexpensive price. SIMO is a fabless semiconductor company that designs solid-state drive (SSD) and mobile flash (eMMC) controllers, as well as enterprise SSD solutions. The replacement of traditional hard disk drives (HDDs) by SSDs and the growth of mobile has been happening for a while now, and SIMO has benefitted bigly (I even hear some people, some very smart people, say yugely). Revenue and EBIT 5-year CAGRs are 20% and 27%, respectively.

Flash memory is a complementary good for SIMO’s controllers and flash supply is currently tight (and prices firm) as the industry moves capacity to 3D flash. After a long run of growth, SIMO guided 2017 revenue flat to up 10%. Investors weren’t too enthused by this guidance, but if your time horizon stretches beyond 2017, SIMO looks like a good deal at a 13x P/E. Flash supply won’t always be tight and 3D flash will lead to further declines in cost/bit and end prices for SSDs. This will further hasten the replacement of HDDs by SSDs, and increase demand for SIMO’s controllers.

Business Overview

SIMO has four business segments: SSD controllers, eMMC controllers, enterprise SSD solutions, and a legacy bucket.

SSD Controllers

This segment has been growing rapidly as SSDs replace HDDs. Growth in 2016 was ~175% to over $170mm. There are ~500mm consumer drives shipped each year and SSDs are now about ~25% of the market, or 125mm units. SIMO’s SSD controller market share is roughly 30% and its main customers are Intel, WD/SanDisk, and Micron. I would highlight Intel as a customer because they are extremely aggressive on price (look at the retail price of Intel’s 600p SSD) and SIMO is the sole-source merchant controller supplier. Intel is spending $5.5b to outfit its Dalian fab in China to produce 3D flash for SSDs. Intel wants to drive adoption of SSDs for strategic reasons and SIMO will grow alongside it.

After 9 consecutive quarters of sequential growth, SIMO expects SSD controller sales to experience a seasonal decline in Q1. However, the company still expects this segment to grow 20-25% to over $200mm in 2017. I feel good about this segment growing nicely over the next 3-5 years. The inexorable takeover of the consumer drive market by SSDs will be hastened by 3D flash restarting Moore’s Law cost scaling. SSDs will continue to get cheaper and more consumers will buy SSDs instead of HDDs.

eMMC Controllers

Hynix is SIMO’s largest customer. Hynix buys a lot of eMMC controllers from SIMO, pairs the controller with Hynix flash and DRAM, and sells the package to smartphone manufacturers. This segment had sales of over $170mm in 2016 and SIMO expects growth of 5% in 2017, in line with overall smartphone unit growth.

Obviously, the customer concentration with Hynix is risk factor, but SIMO has been a good partner and I find it unlikely that SIMO loses the eMMC business. At SIMO’s 50% corporate gross margin, Hynix would gain ~$80mm in gross margin dollars if they were to completely switch to an internal eMMC controller. I think Hynix probably has higher priority projects. Also, Hynix was reported to be developing an eMMC controller in 2013 after the acquisition of Innostor’s eMMC controller unit… and here we are four years later with no internal solution.

SIMO more likely loses share at Hynix as the next standard in mobile embedded storage, UFS, rolls out. SIMO is still supplying an UFS controller to Hynix, but Hynix also has an internal UFS solution. It appears Hynix will use a mix of SIMO and internal controllers. This potential share loss is mitigated by a few factors. First, the transition from eMMC to UFS will take some time. Industry forecasts call for UFS to be 45% of the embedded mobile storage market in 2020, leaving eMMC still with the majority share. Second, SIMO is engaged with another flash manufacturer for UFS. The new partner is likely Micron and the UFS product should be in production this quarter. The UFS transition gives SIMO an opportunity to diversify away from Hynix. Finally, there is potential growth for eMMC outside smartphones in set-top boxes, smart TVs, auto, and IoT devices.

Investors have focused on the customer concentration with Hynix, but on closer inspection, it is not that big and scary. SIMO will lose share at Hynix over time as UFS rolls out, but this negative is balanced out by a number of positive factors.

Enterprise SSD Solutions

SIMO bought a Chinese company called Shannon Systems for $57.5mm in July 2015. Shannon sells enterprise SSD solutions (PCIe flash and SSDs) in the Chinese market. At the time of the acquisition, SIMO expected Shannon to generate $14-$18mm of sales for 2015. Shannon ended up doing $25mm. In 2016, sales more than tripled to $80mm. This hypergrowth was driven by a large project with one of China’s big hyperscale Internet companies, Alibaba.

I think Shannon acquisition will turn out nicely for SIMO in the long run, but insufficient flash supply will put a crimp on this business in 2017. SIMO guided the Shannon business to be down 15% to 20% in 2017, a sharp turnaround from the 15% to 20% growth they previously expected. The discussion around this got a little bogged into technical details during the Q4 call, but the bottom line is that flash is in short supply. When this happens, the flash manufacturers will allocate to the highest margin opportunities, such as their own SSDs, instead of selling bare flash wafers to Shannon and Alibaba. Unfortunate for SIMO, but such is life when you don’t have your own supply.

When flash supply loosens up in the second half of 2017, Shannon’s projects can move forward again. Shannon works with its hyperscale customers to build highly customized SSDs solutions, so there is limited competition. The Chinese government also encourages the big Chinese internet companies to purchase from Chinese technology providers. I think Shannon ends up being a fairly good business in the long run.

Legacy

The remainder of SIMO’s revenue is made up of $90mm of expandable storage controllers (things like USB flash drives) and a $45mm hodgepodge of multimedia SOCs and RF transceivers. Not much to say here except this segment will decline around 10% annually. Maybe they can sell off the RF business.

Valuation

SIMO trades for 13x 2016 GAAP earnings and consensus has earnings flat in 2017. SIMO has nearly $8/share in cash. The stock yields 2%. Once flash supply improves, earnings growth will reaccelerate. I think SIMO can earn $3.60 GAAP EPS in 2018, or 15% growth over 2016. I think earnings in 2020 will be even higher as flash penetrates further into consumer drives, mobile/embedded applications, and the enterprise datacenter. I think SIMO gets a better multiple once we get past the 2017 speed bump and the company diversifies away the Hynix risk. A 16x multiple on $3.60 2018 GAAP EPS gets us to $58/share.

Catalyst

Flash supply increases later this year.
Investors realize Hynix risk is a bit of a paper tiger (unlikely).
SIMO diversifies away from Hynix over time (more likely).