Description

SIMO is a merger arbitrage opportunity with a very large 62% spread, expected to close by 7 August 2023 for a 108% ungeared IRR. If the deal does not get approval from the Chinese competition authorities (note this a Taiwanese company, not Chinese), shareholders will be left with arguably even better value: a semiconductor company run by its founder that has doubled revenue over the last three years, with high stable margins (50% GP margins, 25% operating margins) and pre-tax returns on tangible capital (40%+) trading at an ex. Cash LTM P/E of 7.9x. Historically over 10 years the company has traded at a P/E of 15x – 20x. SIMO is a dominant player in the NAND flash controller space for non-enterprise customers, with an estimated 40% market share. SIMO is very likely to benefit from the long-term secular tailwinds propelling the flash memory market (growing data usage, increasing memory complexity, the shift from HDD to SSD, IoT and electric vehicles) without the fluctuating margins and high capex (SIMO is a fabless designer of controllers that uses TSM to manufacture its products) that are associated with other memory makers.

I think the market is confusing uncertainty with risk and view an investment in SIMO at these levels as a win-win. There is a chance that the coin lands on its edge and China invades Taiwan in the next 9 months, but I view this risk as very small (although it looms large as these types of headline risks often do) and investors are well rewarded for taking it.

There have been several SIMO write-ups over the last few years, the most recent by driftwood explaining the current merger arbitrage opportunity that exists. So why another SIMO write-up so soon? Although these write-ups identify the opportunity that SIMO offers (the most important part of any investment thesis is certainly finding the prince after kissing many frogs), I believe that sketching the full picture in more detail will assist VIC members in seeing the opportunity more clearly. There may be a little repetition since I’d like the write-up to give readers a clear picture on its own, but I’ll try to keep this to a minimum. In addition to more discussion about the MaxLinear (MXL) deal’s regulatory approval by SAMR in China, I explain why I think that even if the deal is declined by the Chinese regulator, this is an even more attractive longer-term opportunity for SIMO shareholders. Building on the write-ups by specialk992, I’ll provide detail around how they managed to grow to a $1bn revenue company so quickly, go into more of the competitive positioning for SIMO in the market, and explain why I see SIMO as an outstanding opportunity to take advantage of the longer-term secular growth trends for memory without the typical margin fluctuations that accompany other memory makers. Lastly I’ll give my two cents on the China-Taiwan tensions which I see as the main (albeit very small) risk to the investment thesis.

This write-up will be divided into three scenarios that each have different payoffs. In the scenarios I will consider the MaxLinear deal and its likelihood, then the SIMO business and its longer-term prospects if the deal falls over, and lastly a China-Taiwan invasion.

Scenario 1: MaxLinear’s proposed deal to purchase Silicon Motion

The proposed deal by MXL is $93.54 in cash and 0.388 MXL shares for every SIMO share. At the current MXL price of $35,35 this gives SIMO shareholders $107.26 in value for a 62% return at the current SIMO price. The deal is approved by SIMO shareholders and fully funded, with MXL having debt committed by Wells Fargo and expecting leverage of about 4x EBITDA by transaction close.

The main stated reason for MXL buying SIMO is to increase its significance with TSM as a supplier – SIMO has about 3x the volume with TSM that MXL does. In addition, MXL who have made several acquisitions in the past, see some cost synergies and the ability to sell SIMO controllers to their enterprise customers where SIMO doesn’t currently have much of a presence. Importantly there is virtually no product overlap between MXL and SIMO currently.

When the transaction was first announced on 5 May 2022, SIMO shares jumped from $81 to $95. At that time MXL was trading at $53 making the deal value about $114 per SIMO share. Rumours of a potential deal had been surfacing during April when SIMO was trading in the $70s. Strangely, despite the deal having been announced, SIMO is now trading below its pre-deal price and is in the $60s. This is despite continuing to deliver decent results in a tough near-term memory market. Clearly the market attaches zero probability to the Chinese regulator approving the deal from a competition perspective (I disagree). And obviously the recent tensions between Taiwan and China in which the US has been involved haven’t helped.

SAMR approval

China’s State Administration for Market Regulation (“SAMR”) needs to approve this deal because a portion of SIMO and MXL business is with Chinese companies. It is SAMR’s job to ensure that any deal does not negatively impact Chinese companies, or China more generally, from a competition or strategic perspective. 

For a detailed overview of the most recent deals approved by SAMR in the semiconductor industry, I refer you to the document in the link below. The deals discussed in that document were all done before the recent US chips ban on China, but I think they are still instructive with regards to SAMR’s past behaviour. The discussion largely addresses issues relating to approval timing, competition issues identified and remedies proposed by SAMR:

https://www.competitionpolicyinternational.com/lessons-from-samrs-recent-merger-clearances/

Consider the following statement, quoted from the document:

“The most significant determinant of a lengthy review in the SAMR process turns obviously enough on whether SAMR identifies competition concerns. A further important determinant is whether SAMR has a sectoral or industrial policy interest in the transaction. When these two factors combine – as they did for four out of the six conditional clearances issued by SAMR between January 2021 and January 2022 – a long review is inevitable.”

SIMO and MXL really have no product overlap to speak of. SIMO makes NAND flash controllers and MXL makes communications systems-on-chip solutions for use in broadband, mobile, data center and industrial applications. So from a competition standpoint, there really is minimal concern here. However, since the deal is in the semiconductor industry, SAMR will pay close attention to it simply because of China’s intense focus on this industry and their worries about being over-reliant on the US for semiconductor technology.

Initially, MXL tried to speed up the approval process by applying to SAMR under a simplified filing. From the perspective of product overlap this made sense, but considering this deal is in the semiconductor industry MXL’s approach was perhaps a tad optimistic. SAMR then asked MXL to refile under the normal procedure. A timeline of the events that transpired is as follows:

•  27 June 2022: the waiting period under US anti-trust HSR Act expire

•  6 July 2022: MXL make a simplified filing with SAMR

•  31 August 2022: SIMO shareholders approve the deal

•  31 August 2022: SAMR ask MXL to refile the deal under the normal procedure in China. MXL “can’t predict the length of the review under the normal procedure, but expect a final determination by SAMR in the second or third quarter of 2023”

• 1 November 2022: MXL’s application for SIMO filing is said to be accepted by SAMR under the normal procedure (meaning they’re starting to look at it)

So if we put this deal in context with other semiconductor deals that needed SAMR approval, it looks as follows:

• The SIMO deal is the smallest of those in the table above. This arguably could allow it to fly under the radar somewhat.

• As mentioned, the SIMO deal has no real competition concerns whereas the first three in the table all seem to have some issues. This means that from a timing perspective I wouldn’t expect the SIMO deal to take as long as the others. The DuPont / Rogers deal was recently terminated due to SAMR taking too long and going beyond the drop-dead date. Admittedly the Rogers business did apparently deteriorate after the announcement.

• The AMD / Xilinx deal didn’t seem to have a direct product overlap but I suspect the sheer size of that deal and Xilinx’s share of the Chinese FPGA market prompted SAMR to apply some remedies.

• SAMR does like to impose behavioural remedies, especially ones that keep the commercial terms of business transactions with Chinese companies unchanged (“FRAND” terms) after the deals. So if the SIMO deal is of any concern (or even if its not) SAMR may impose these types of restrictions on SIMO which sells a sizeable amount of controllers into the mobile market to Chinese companies and also does a little business with the Chinese hyperscalers like of Alibaba and Baidu. To quote the document again: “As a rule, SAMR uses behavior remedies to address vertical or conglomerate concerns.”

• The Intel / Tower Semiconductor deal could provide us with an early-warning of SAMR’s approach to more recent transactions.

Overall I’d say that the SIMO deal compares favourably with previous deals, purely from a competition standpoint. Their expectation for approval timing and likelihood thereof also seems reasonable. 

US chips ban on China

The obvious fly in the ointment here is the recent chips ban by the US on China and whether this will have a bearing on SAMR’s decisions. It would be naive to think that decisions by SAMR are made in a vacuum without input from the broader Chinese government. The price that the market currently places on SIMO certainly suggests that the US chips ban makes the likelihood of a deal with MXL almost zero. I don’t think this is the case and I’ll explain why.

China is likely to act in its own self-interest. By stopping all deals – either by denying them or just dragging the approval out indefinitely until the deal collapses – it sends the message to US companies that doing business with Chinese companies, especially in the semiconductor space, may make your own company unsellable. From my perspective this is the opposite of what China wants. They are trying to close the technology gap and reduce reliance on US semiconductor firms. Cutting off access to the products of those firms, if they decide dealing with China just isn’t worth it, makes closing the gap much more difficult. In addition, by not approving the deal, SAMR would be robbed of the opportunity to impose some restrictions on the current SIMO and MXL business that would help Chinese customers of these firms.

One could argue that China will act out of spite / retaliation and block all semiconductor deals. Although I can’t rule this out, this would strike me as quite short-sighted and arguably would exacerbate the problems caused by the chips ban. The SIMO deal is also not a large one that SAMR / China could make a statement with. Such expected behaviour by SAMR also doesn’t explain why the spread for the Intel / Tower Semiconductor deal is only 17% at present. 

Of course there could be another reason for China to block this transaction and I certainly can’t pretend to understand all of the CCP’s actions. But if there is another reason, I can’t see it.

Scenario 2: Silicon Motion’s long-term prospects if the deal collapses

If the deal is denied approval or approval is not forthcoming by (the latest extension of) the Outside Date of 7 August 2023, then SIMO will be paid a termination fee of $160m by MXL (assuming they don’t renegotiate the deal on better terms):

At this point, investors who were hoping for the gains from a merger arbitrage situation will become longer-term shareholders so it is instructive to understand exactly what the prospects for SIMO will be then.

I’ll start with a brief background into the competitive positioning of SIMO, provide some thoughts on their recent financial history and then break down the valuation. 

Background – SIMO now dominate their niche and should benefit from growth tailwinds

Since SIMO is a supplier of NAND flash controllers, it is useful to start with what this actually means. NAND flash memory (which is different from RAM since it provides long-term storage as opposed to short-term processing) stores data and is present in PCs, mobile devices, servers, and increasingly automobiles, industrial devices and appliances. NAND flash requires a controller which does what it says – it controls the use of the memory by telling it how and where to read / write data, corrects errors, helps to prolong memory life, provides in hardware-level encryption and performs many other functions. Critically, the controller is a small part of the total flash memory cost but is a critical component. As memory becomes more sophisticated (and cheaper), its need for more advanced controllers increases.

SIMO’s business can be broken down into two main product types and three customer types that serve multiple end-markets:

• Although a lower-end product, the eMMC market is growing strongly because of the multiple uses in automotive, industrial and IoT applications. SIMO are dominant in eMMC, especially since Samsung exited this market in 2021.

• UFS memory is starting to make its way into more mainstream smartphones, driving growth in the category.

• SSDs (solid state drives) are slowly replacing HDDs (hard disk drives) as this superior memory technology becomes cheaper. So despite PCs and laptops not really being a growing market, the shift to SSDs results in growth for SIMO in this category.

• SSD solutions are complete memory products (memory + controller) including enterprise products sold to Chinese hyperscalers Alibaba and Baidu through their Shannon brand. This segment has some fluctuation in margins due to the underlying price changes of the memory. Marvell are the dominant players in the enterprise controller space.

• Many of the flash manufacturers have their own internal controller teams focusing on their newest memory but are increasingly outsourcing controllers for the bulk of their memory, mainly to SIMO. Historically the main concern for SIMO is that they would be cut out of the process by the flash manufacturers. But there is a recognition that a) making controllers is difficult and requires focus and b) a company like SIMO can aggregate and spread its R&D efforts across many different customers, reducing the eventual cost.

Module makers have historically been a large part of SIMO’s business, but this customer category is less sticky. They tend to be more volatile in their demand for controllers compared to OEMs that have long product runways. Once won, an OEM contract is 3 – 4 years long making for much stickier business. This is where SIMO have seen the most growth in the last few years.

SIMO have benefitted from outperforming their competitors who have slowly exited the merchant controller market. Marvel exited in 2018 and focusses on enterprise controllers. When Samsung’s fab in Austin (where it made its controllers) shut down in February 2021 for a month due to an ice storm, they decided to exit the eMMC controller market. Phison, another Taiwanese company, are the only other major competitor but they struggle to compete – Kingston previously just used Phison controllers and was a major shareholder, but recent have ramped up use of SIMO controllers. SIMO therefore dominate the merchant controller market and this has significant economies of scale advantages, allowing more spend on R&D and helping to ensure they maintain product leadership.

Financials – strong growth, high margins & returns on capital, strong cash flow

The following table provides a six-year snapshot of the SIMO business from a financial perspective. Take note of the stability in the GP margins, the operating leverage that has developed from the revenue growth, the low capital expenditure and the return of capital to shareholders in the form of dividends and buybacks:

Some further notes on SIMO’s financial history:

• The revenue dip in 2019 was due to the loss of SK Hynix as a UFS controller customer. They were quickly replaced by Micron. In 2021 SIMO added a second UFS controller customer and five more in 2022.

• The substantial revenue growth in 2021 was partly due to the factors mentioned above, namely multiple OEM design wins and the exiting of Samsung from the eMMC controller market. This year also benefitted from stay-at-home demand, but interestingly SIMO struggled to get all of the wafers it needed from TSM due to the chips shortage. As testimony of their importance in the supply chain, some of SIMO’s customers actually lobbied TSM to increase SIMO’s chips allocation during this period. Critically, SIMO have held onto their market share gains in 2021 in the last twelve months, despite the recent downturn in the memory market.

• The impairments were a write-down of the Shannon enterprise business where SIMO has struggled. They serve some of the Chinese hyperscalers like Alibaba and Baidu. If this segment does eventually succeed it would just be further upside from here.

• The large capex in 2018 was mostly due to $58.9m spent on a property in Taiwan which SIMO have plans to develop into a head office. The capex is slightly above depreciation, averaging $20m per year after adjusting for the property versus $15m for depreciation. But the difference is small making EBIT a fair valuation metric.

• The free cash flow in recent years has been lower (although still healthy) than one would expect after the recent growth because of the investment into working capital necessitated by this growth. If topline growth did slow down, free cash flow should improve substantially.

• The inventory buildup in the last twelve months is substantial, well above historical levels as indicated by the net working capital days. Without management transcripts recently (they ceased when the MXL deal was announced) it is difficult to know exactly why the inventory has built up so much. Management did say in their Q3 earnings release that recent slowdowns in the PC and smartphone markets have impacted the module makers significantly but that (stickier) OEM sales continue to accelerate. I don’t expect much in the form of write-downs of this inventory as SIMO has almost no history of this. But the most recent quarter did see GP margins of 47.5%, below the longer-term history, which has partly been put down to mix and partly to discounts. So there are some signs of market pressure.

• Buybacks have really ramped up in recent years and dividends have grown consistently. Note that both have ceased with the announcement of the deal with MXL, so cash is building up on the balance sheet. I expect both to return and accelerate if the deal does not materialize.

Overall this is a very high quality business as shown by the cash flows and returns on capital. Management is also shareholder friendly, as indicated by its dividends and buybacks. The Chairman and CEO each own more than $25m of stock which compares favourably from an alignment perspective with the total executive compensation of $4.23m in FY21. They have no debt and run conservatively with cash on the balance sheet. The business has substantial longer-term secular growth tailwinds that should drive continued topline growth since they are the dominant merchant controller supplier to the mobile and PC markets. Their substantial R&D spend (more than $750m in the last six years) over many years has put SIMO in a dominant industry position to continue winning business and potentially even break through into the enterprise market.

Management projections from the merger filing

It is interesting just to note management’s projections for the business, which they included as part of the merger filing. This should be taken with a pinch of salt since management teams are often too optimistic about their own businesses. But for me it has some credibility considering that a projection which is too aggressive may have caused shareholders to ask some questions about the price they accepted from MXL.

Note the continued topline growth, steady margins and increase in free cash flow as the topline growth moderates.

Valuation for SIMO if the deal is not approved – cash is building on the balance sheet

As mentioned above, the dividends and buybacks have stopped so cash is building up on the balance sheet. This cash will be boosted by the termination fee that MXL will have to pay if the deal is not approved. My estimate of the cash we can expect in September 2023 if the current memory market struggles continue and the topline does not grow in the interim is as follows:

Note that for the operating cash flow I’ve assumed that the latest quarter’s GP margins of 47.5% persist for another year on flat revenue and the opex grows by 10%. I’ve made no allowance for a release of the balance sheet inventory to more normal levels which could add a further $150m to this figure. 

After including this cash in the valuation and assuming no earnings growth in the next year the business just appears to be unreasonably cheap, especially when factoring in its quality and growth potential:

Besides the extremely healthy pre-tax yield of 16% compared with a risk-free 10-year government bond yield (by then) of ±5%, the EV / EBIT multiple should also be seen in the context of SIMO’s own 10-year history where the business has frequently traded between 10x and 20x EV / EBIT and averaged 12.2x EV / EBIT over this period (as mentioned, EBIT is a fair proxy for the pre-tax cash yield since depreciation and capex are fairly similar):

All things considered, SIMO is probably at least 50% undervalued if the deal does fall over and a price of over $120 would not be a stretch at all considering its future prospects.

Scenario 3: China invades Taiwan

Now I won’t pretend to have any special knowledge about this major global issue (I tried to arrange tea with Xi Jinping – he asked me to just call him “X” in our chats  – but our schedules just wouldn’t line up). But I wanted to just explain my thinking on the matter as to why I think a near-term conflict is extremely unlikely. I think this is important because it’s the only real reason I can see for the current disconnect in the SIMO stock price. Others may disagree with my views here, seeing them as too simplistic, and view a near-term conflict as likely. This is a judgement call.

Obviously an invasion of Taiwan is going to be disastrous for SIMO’s business (and the world). Although their head office is currently in Hong Kong, TSM does their fabrication and most of SIMO’s operations are based in Taiwan.

My reasoning for an invasion being extremely unlikely in the near-term is as follows:

• China’s handling of the Hong Kong situation was done in such a way to not create too much conflict. They bided their time and when the world was busy dealing with COVID, they made their move. China strikes me as patient and as many have said, they think long-term. 

• The US have come out regularly in support of Taiwan. It would be an optimistic view for China to take that an invasion of Taiwan would not be met by retaliation from the US and the West more broadly. The recent treatment of Russia over the Ukraine invasion shows what steps the West could take in such a situation. The US (and the world) stand to lose massively, much more than from a Ukraine invasion, if Taiwan is invaded considering their importance to global semiconductors. Can China afford a large retaliation by the West right now?

• China are coming out of a difficult period from an economic perspective. They’ve had issues in their property market and COVID zero has not been kind to their economy. Even if there is an invasion eventually, this taking place in the next year strikes me as extremely unlikely considering their recent economic weakness. 

• Contrary to some media headlines it does not appear like China’s stance on desiring peaceful reunification with Taiwan has changed. I found the following article quite useful in this regard where they compare the changes in Chinese rhetoric with respect to Taiwan during Xi Jinping’s recent re-election. (For what its worth, this site does not appear to be heavily pro-China. It also has a podcast on the abuse of the Uyghur people in west China.) Of course the recent reaction to Pelosi’s visit of Taiwan sends a different message, other articles can be quoted that strongly support the likelihood for an invasion, and China does not rule out military action if deemed necessary.

https://warontherocks.com/2022/11/listen-to-xi-jingping-about-taiwan/

• A recent strong win by the KMT party over the DPP party in Taiwanese mayoral and county elections should also give China a reason to hope for peaceful reunification eventually and to bide their time. KMT traditionally favours close ties with China whereas DPP does not. Although this recent win does not allow the KMT members to have a direct say on China policy, it may point to the outcome for the 2024 presidential elections.

An investment in SIMO could be reassessed in September 2023 and I think the risk of an invasion in the meantime is tiny. A January 2025 Put option on TSM (SIMO doesn’t seem to have longer-dated options so TSM can be a proxy) with a strike at the current price will cost you about 18%, which would reduce your merger-arbitrage returns by about 30% (assuming the Put expires worthless eventually … probably overly conservative since you may be able to close it out sooner) and last for 2 years giving SIMO time to return to fair value if a deal collapses. This should protect you against a catastrophic invasion scenario.

Buffett’s recent investment in TSM is also noteworthy here. It was sizeable (making me think it was Buffett who made the investment and not Ted Weschler or Todd Combs) at $4bn, although this is still a very small portion of Berkshire overall. This does suggest that Buffett does think the probability of a Taiwan invasion is very unlikely.

Conclusion

SIMO is a high quality business that appears deeply undervalued. A deal with MXL is more likely than the market seems to think and would provide a very nice return of 62% in just 9 months. No deal means you get a high cash flow generating asset in a growing industry with shareholder-friendly management and a dominant market position at least 50% below fair value. Your major risk is a Taiwan invasion of China, but this seems very unlikely in the next 9 months. Seems like a lucrative win-win situation.

Catalyst

Deal approval by SAMR before 7 August 2022

If no deal approval, reinstatement of dividends and buybacks