Navigating The Fog On Silicon Motion's Horizon

|
About: Silicon Motion Technology Corporation (SIMO)
by: Jenks Jumps
Summary

Silicon Motion reported 2018 fourth quarter and full-year results on January 30th. The results were mostly already known. But, the 2019 guidance was influenced by a foggy horizon.

Its NAND flash partners have limited visibility and are, therefore, providing six month procurement forecasts rather than twelve month forecasts.

Though the near-term is cloudy, the long-term market outlook still holds substantial potential and is projected to grow at a high-teen rate through 2026.

The most important advice to heed when driving in fog is to slow down and take extra time. That's exactly what Silicon Motion (SIMO) is doing – slowing down and taking extra time. Most bouts of fog are temporary. This industry's fog should be as well.

Quarterly Results

Silicon Motion is the world's leading NAND flash controller supplier. It reported 2018 fourth quarter and full-year results on January 30th. Because the company routinely provides preliminary results a few weeks prior, the quarterly report was largely a known rather than an unknown.

For the quarter, revenue was $123.4 million compared to the company's guidance range in its preliminary report of $120.5 million to $124 million. For the year, revenue improved 1% to $530.3 million compared to $523.4 million in 2017.

Though total sales improved only slightly, sales of 3D NAND flash controllers at higher gross margins grew while sales of lower-margin controllers decreased. Thus, GAAP earnings per diluted ADS improved to $2.78 for the year, a 33% improvement over the $2.09 in 2017. Non-GAAP earnings per diluted ADS improved 21% to $3.41 compared to $2.81 in 2017.

The Fog

With its release of quarterly results, Silicon Motion also provided initial guidance for 2019. The company explained its present view is somewhat foggy.

“Some of our NAND flash partners who typically provide us with a twelve-month controller procurement forecast have reduced their forecast period to just six months. We believe they are struggling with their own limitation in operational visibility due to demand uncertainty from weakening Chinese economic conditions, U.S.-China trade negotiations and other issues."

The weakening Chinese economy has been a hotly-discussed topic for months. Last November, Business Insider noted the situation was creating a “crisis of faith”. Policymakers in China want both Chinese citizens and the rest of the world to believe they are ready and capable of handling the slowdown. But, President Trump's expectations for changes to trading policies were not supposed to be part of those equations.

A recent article in the Los Angeles Times suggested the Chinese government has been censoring bad news. Though official growth numbers are around 6%, some reports have found growth in the low single-digits or negative. This creates a mountain of mistrust about the “official” line and many believe otherwise.

“Just about every indicator of economic health in China is down.”

At the heart of the trade war threat is the Chinese government's Made in China 2025 initiative. It aims to innovate and upgrade manufacturing capabilities in the country. As part of the strategy, forced technology transfers were to be required. A forced technology transfer “requires non-domestic organizations to license technologies and know-how to domestic partners”. The Chinese partners would gain (NYSE:IP) intellectual property and advanced technologies without having to develop themselves. The non-domestic companies would lose their IP but gain access to the world's second largest economy.

The two countries have agreed to a temporary truce and set a deadline of March 1st on negotiations. Barring an agreement, U.S. tariffs could increase from 10% to 25% and tariffs on additional goods could be initiated.

Between the two concerns, a slowing economy and the trade wars, products most likely to be impacted tend to be higher-end such as autos and smartphones.

Because of management's conservative nature and its limited visibility, Silicon Motion is guiding 2019 full-year revenue to be basically flat compared to 2018. The first half of 2019 is expected to be slower than the second half. The first quarter revenue projection of $97.5 million to $103.6 million reflects a significant decline compared to 2018 first quarter revenue of $130.3 million, the highest first quarter revenue in corporate history.

Throughout 2017, the industry converted flash storage from 2D NAND (planar or single layer of memory cells) to 3D NAND (stacked vertical layers of memory cells). The conversion resulted in tight supply which drove prices higher. As 3D NAND flash inventory increased in 2018, prices, in turn, fell. As prices fell, adoption and demand increased. NAND flash storage devices require controllers.

“A flash controller provides the interface between the host controller or processor and the flash memory device. It decides where in the flash memory to store the information. It then retrieves the information when requested. Flash controllers will also organise the way data is stored to achieve ‘wear levelling'.”

Silicon Motion's SSD controller sales increased 35% in 2018 compared to 2017.

The company expects NAND flash prices to continue to fall rapidly in the first half of 2019. In October, 2018, industry sources estimated prices could fall another 50% throughout 2019. With prices falling so rapidly, some of Silicon Motion's customers have opted to wait for lower prices. However, this short-term breather hardly changes the long-term outlook.

According to Market Prognosis, the global SSD market should grow at a high-teen rate annually and generate $100.44 billion by 2026. The adoption of SSD in smartphones, tablets, notebooks, gaming consoles and other “smart” devices continues to increase. From an enterprise perspective, SSDs are being deployed for cloud computing, e-commerce and high performance data centers.

Another primary target market for Silicon Motion's client SSD controllers are PC OEMs. Less than half of the PC units shipped in 2016 included SSD controllers. But, the company expects client SSD controllers will replace HDDs (hard disk drive) in the majority of units shipping in the next few years. It expects to own 40% to 45% of this share of the PC OEM market.

The Next SSDs

Silicon Motion reported its program pipeline is 70% fuller at the end of 2018 as compared to the end of 2017.

As mentioned above, 3D NAND is based on stacking vertical layers of memory cells. Major vendors, Samsung (OTC:SSNLF, OTC:SSNNF), SK Hynix (OTC:HXSCF, OTC:HXSCL), Toshiba (OTCPK:TOSBF, OTCPK:TOSYY), Western Digital (WDC), Micron (MU), Intel (INTC), are already or are very close to mass producing 96-layer versions.

Samsung just recently updated its 970 EVO SSD with a new version, 970 EVO Plus, to glowing reviews. The new SSD is based on 96-layer V-NAND, Samsung's moniker for vertical NAND. The company claims it is 40% faster than 64-layer, its data write speed improved 30% and its power requirements are less.

In its fourth quarter earnings call, Silicon Motion confirmed it is shipping an SSD controller for a 96-layer TLC-based SSD.

But, there are unknowns regarding further scaling.

“As you add more 3D NAND layers at a given cell density, the WL [Word-Line] staircase also needs to lengthen and takes more space. For example, in the case of a 32-layer NAND device, the WL staircase stretches out 20um from the edge of the cell array. For a 128-layer architecture, the WL staircase would extend out 80um [2]. Current WL staircase designs may be a key obstacle to cell efficiency and scaling of this type of 3D NAND architecture, due to this linear scaling effect.”

The quadruple-level (NASDAQ:QLC) stacked cell structure could provide a capacity boost because it stores four bits of data per cell as compared to TLC's three.

Micron and Intel announced QLC-based PCIE NVMe client SSDs in May, 2018. The trade-off for capacity is lower write endurance and lower write performance. In July, Western Digital and Toshiba announced the development of 96-layer QLC with storage capacity of 1.33 TB. Mass production was targeted for 2019. SK Hynix disclosed its plans in late 2018 for a QLC version of its 96-layer 3D NAND SSD with a capacity of 1 TB in 2019. Samsung announced QLC model names in late 2018 but has not released any specifications or dates. Several of the model names indicate the enterprise-level market.

The vendors are equally busy developing storage solutions for cloud computing, e-commerce and high performance data centers.

Mentioned above, Samsung already has enterprise-level 96-layer QLC models named. SK Hynix has enterprise-grade plans for the second half of 2019 for its 96-layer 3D NAND TLC SSDs. Western Digital and Toshiba's plans for 96-layer QLC extend to the enterprise. In the May 2018 announcement, Micron and Intel shared its roadmap for QLC SSDs.

Silicon Motion's Enterprise-Grade Status

Silicon Motion's consumer-side adoption of 3D NAND flash in 2018 paced ahead of the company's initial expectations. But, adoption on the enterprise side lagged the company's initial expectations.

As late as August 1st, the company was expecting decent sales to transpire in 2018. It didn't change its forecast until October. The migration to production had been pushed to the second half of 2019.

“We are happy to announce that our controller has passed final testing with two hyperscale customers. And these controllers are scheduled to begin initial production this quarter. However, initial shipments will be small. SSD solutions using this controller will still need to undergo live data center application testing, before they go up extensively in the second half of next year.” (emphasis added)

In the fourth quarter earnings call, Silicon Motion confirmed its most recent projections.

“The performance tuning of open-channel SSD for Alibaba and another PATA customer using these new controllers is progressing smoothly and we remain on track to begin volume commercial production of our open-channel SSD around the middle of this year.”

Based on its progress, Silicon Motion also shared it has been able to demonstrate its controllers' capabilities with other potential data center customers.

“We are in parallel processes for both enterprise open-channel solutions in China and enterprise controller engagement for the U.S. market.”

Considerations

I've been covering Silicon Motion for five years. It has consistently offered cautious guidance and often beat on the high side – the epitome of under-promise and over-deliver. In a nutshell, management has proven its trustworthiness.

Thus, I have faith in Silicon Motion's caution for 2019.

“Without a clear sales forecast, we believe that guiding directionally for the full year is more prudent than guiding with false precision to a specific sales range.”

I also have faith in their sentiment about the marketplace in the near-term.

“This is a very unique situation right now.”

Until the U.S. and China trade negotiations are firmly established, consumer spending forecasts will be somewhat questionable and foggy.

And, yet, I have faith in the company's long-term potential.

The company has nearly $290 million in cash and short-term investments and no debt. It pays an annual dividend rate of $1.20 per ADS. At a stock price in the $40 range, it is trading at a P/E ratio less than 15 based on 2018 GAAP earnings and less than 12 based on non-GAAP earnings.

There may be fog on its horizon but it seems clear, for the long-term, Silicon Motion is a safe haven investment.

Disclosure: I am/we are long SIMO. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: I belong to an investment club that owns shares of SIMO.