Description

Investment Summary

MaxLinear (“MXL” or the company) is an integrated circuit provider for the connected home, wired and wireless infrastructure, and industrial market. The company’s main products are sold to cable hardware manufacturers (Arris and Technicolor) and ultimately end up in homes and businesses within cable modems and other infrastructure products. Aside from its core markets, the company has grown substantially through $878mm of acquisitions over the last 4 years, and claims to have a heavy R&D operation, creating an aggregate $9bn future TAM opportunity versus $463mm of pro forma revenue in 2017. Management promotes this story of an innovative and R&D heavy company that is geared up for large future growth drivers, but our research as detailed below has given us cause to be skeptical.

We previously wrote up EXAR on VIC in November 2016, and while we believed at the time that EXAR was well-situated to be acquired, we were surprised when the buyer was MXL. We have since followed MXL and currently do not believe the MXL bull case based on our fundamental work and primary diligence. Our short thesis revolves around the following points:

• Intense and increasing competition in MXL’s core broadband segment
• Expected growth drivers also face intense competition and headwinds
• Our checks indicate little strategic rationale for the EXAR acquisition, and the acquisition is already underperforming expectations
• A host of accounting red flags including elevated DSOs and DSIs, “spring loading” of the EXAR acquisition (holding back revenue until the deal closed), obfuscating organic growth with numerous acquisitions, and other flags may be artificially inflating earnings

Forward estimates are currently for MXL to grow at increasing and accelerating rates each quarter starting in 2H 2018 and ramping through 2019. We see significant risks to estimates over that period especially, based on primary research and a detailed look at the accounting. One time boosts in 2017 create a scenario where the earnings base from which forward estimates are grown off of in 2017 is inflated, creating a higher need for growth than most investors realize to hit 2018/19 estimates.

We believe a pattern of earnings misses combined with lack of visibility into growth opportunities is likely to cause a scenario of multiple compression from MXL’s current 14.6x P/E and 4.1x EV/Sales multiple. We see MXL closer to 3-3.5x revenue or 11-12x P/E (where it has traded in the past) yielding a price target in the range of $17, or ~30% downside.

Company Overview

MaxLinear is a provider of “system on chip” solutions for broadband data/video and data transmission infrastructure markets. MXL started primarily as a company that made chips for cable set top boxes or gateways to convert input into high speed data and video and has since grown organically and through acquisitions. The business operates in three segments:

• Connected Home consisting of cable modems and set top box integrated circuits converting RF signals to data and video (moving data around the home). MXL says this is a “low growth” market with 5-10% growth with the main customers being Arris/Technicolor who manufacture modems and boxes for cable telcos
• Infrastructure consisting of broadband and wireless infrastructure integrated circuits (getting data to the home). MXL says this is a “high growth” market with 30-40% growth with the ultimate end customers being wireless providers
• Industrial and Multi-market consisting mostly of power management products for industrial applications. This segment is primarily from the EXAR acquisition. MXL says this is a “low growth” market with 5-10% growth, served through distributors

In 2017, 25% of MXL’s revenue was sold to Arris ($105mm), and another 8-10% to Technicolor - the two major global cable modem and set top box providers.

Over the years MXL has grown both organically, as well as through $878mm of acquisitions. It has made the following acquisitions since 2014:

• October 2014: Physpeed, a high speed data center interconnect business ($10mm)
• April 2015: Entropic, a chip company that specialized in data/video over coax cables ($289mm)
• July 2016: The wireless infrastructure backhaul business of Broadcom ($80mm)
• April 2016: The wireless infrastructure access business of Microsemi ($21mm)
• April 2017: G.hn, a last mile data access company over telephone and power lines ($21mm)
• May 2017: EXAR, an analog and mixed-signal fabless semiconductor company specializing in power management ($457mm)

The company’s acquisition strategy is to enter new markets and expand TAM which is supported by the fact that two thirds of its R&D spending the last few years has been related to markets where they currently have zero revenue. According to the company, the acquisitions and organic opportunities have a TAM of $9bn by 2020.

Note: All figures are $ in millions unless otherwise indicated. MXL’s fiscal year ends 12/31

Background and Bull Case

We believe looking at the short case for MXL starts with looking at the consensus/bull case. Over the last few years, management has spun a story of an R&D heavy company with entry into upcoming growth markets. The current consensus/bull case incorporates this view as a “second half 2018 story” where the company will start to ramp its new products, diversify its revenue, and eventually get a higher multiple as growth increases.

The bull case revolves along the following growth drivers:

• DOCSIS 3.1: Within its core Connected Home segment, there is an upgrade cycle from the DOCSIS 3.0 to DOCSIS 3.1 standard. This updated standard allows for faster data transfer rates into the home and is seen as essential to the “connected home” future. DOCSIS 3.1 enabled modems are currently being sold by some cable operators with full roll out expected to gain steam in late 2018 and onwards.
  
  
  
• Data Center Networking: MXL has been a first mover in the 400gb PAM-4 data center interconnect market, which is a new upgraded standard taking data transfer rates from 100gb to 400gb to better connect high speed data centers together. The company expects the TAM here to be approximately $1bn ($100 of content x 10mm servers) over the next 3-5 years.
  
  
  
• 5G Infrastructure: MXL has developed a 5G wireless access chip as well as wireless backhaul infrastructure, which is expected to be used in wireless modems as infrastructure is built out to upgrade 4G to 5G. The company expects the TAM here to be approximately $1bn depending on how many antennas go on each base station over the next 5+ years.
  
  
  
• Integration of EXAR and cross sell opportunities: MXL announced the purchased EXAR in March 2017, and at the time talked about cross sell opportunities by integrating power management onto its existing integrated circuits (such as in set top boxes), as well as integrating their RF chips onto EXAR’s products - namely the server power management space which has an $800mm TAM (10mm servers at $80 of content).
  

The company entered 2018 with $500mm+ revenue expectations, driven by the above growth drivers. At its 12/31 quarterly report, MXL talked about softer markets and delays which brought down full year estimates by ~$30mm or -7%. MXL is now a “back half 2018” story according to the 6 buy-rated analysts the cover the stock. Consequently, current the consensus estimates look like this:

And on a quarterly basis, things accelerate substantially at the end of 2018 and into 2019:

Short Thesis

We do not believe MXL’s “back half” story and think the financial model has further risks to the downside throughout 2018 based on increased competition, challenges within the acquired EXAR business, and meaningful accounting red flags that set up challenging compares. We believe the last point leads to a set-up whereby the earnings base upon which MXL will have to try to grow is lower than generally believed, as is the true underlying growth rate of the business as it stands today.

Intense and increasing competition in MXL’s core broadband segment

We believe MXL is facing intense and increasing competition in its core broadband markets. In this business, MXL partners with Intel and competes against Broadcom - with Broadcom taking ~2/3rds share within cable gateways. With the top tier technology, and multitudes of products, Broadcom is going to OEMs and selling bundles of the entire broadband/networking needs of a customer (data centers, data transmission to the home, data transmission in the home) while Intel+Maxlinear do not have that same portfolio of products.

At a recent investor conference, MXL’s CFO had this to say about Broadcom’s competitive tactics:

Source: Bloomberg transcript

Our conversations with industry professionals indicate that Broadcom’s bundling strategy has gotten more aggressive over the last 18-24 months (since the Broadcom/Avago deal closed). While Broadcom may sell thousands of dollars of content per unit to OEMs or vendors, MXL’s content in a standard cable gateway sits at just $5-10. This gives Broadcom greater power to grab share and squeeze down MXL as they are selling much larger dollar packages to customers. The company acknowledges this fact in its recent 10K with some newly inserted risk factor language:

Source: Factset 10-K Blackline

We think that this increased competition has been flying under the radar as it has been unclear in MXL’s results. Over the last 18-24 months, MXL has done multiple acquisitions as well as noted a $65mm run-off of some “legacy” business from their Entropic acquisition in 2015. Unfortunately, MXL does not formally disclose any organic growth measures to give investors better information on its core business.

Additionally, with the recent break-up of the Qualcomm deal, and AVGO’s stock price -6% YTD versus the Semiconductor Index +12% YTD, it’s possible that Broadcom sharpens their pencils into 2018 and further increases competition for MXL. While we don’t think MXL will meaningfully lose share (since each customer wants a second source supplier to Broadcom), we think it will be challenging for MXL to grow share or ASPs in the face of tough competition even as the industry goes through a DOCSIS 3.1 upgrade cycle. This compares with the company’s growth target of 5% to 10% in its Connected Home segment.

New and upgrade markets face intense competition and headwinds

Meanwhile, MXL has been attempting to diversify away from Broadcom-driven markets through its 5G Infrastructure and Data Center Interconnect products. It is anyone’s best guess what the ultimate market share of MXL will be in these new markets, but our checks indicate:

• The timing of these new markets is uncertain, and upgrades of this magnitude are often delayed
• MXL may be an early mover, but it faces intense competition from much larger competitors with better R&D scale in these markets

In the 5G Infrastructure market, products have not been fully designed yet. The current target is for a roll-out of testing products towards the end of 2018, with at least a 12 month testing process. So meaningful revenue will not start from this growth opportunity until at least 2020 with the potential for normal delays. In this segment, MXL is competing with Analog Devices ($6 billion in 2018E revenue versus MXL’s $476mm). ADI also recently became bigger with its purchase of Linear Technologies closed March 2017.

In the Data Center Interconnect market, MXL is a first mover with its 400gb PAM-4 products. However competition is upcoming from many companies including: Broadcom, Inphi, Macom, IDT, Qorvo, Microsemi, and Semtech. Given the TAM of $1bn or more, we would expect intense competition in this market. The market remains immature, with the main ramp happening in 2H 2019 according to management.

In both these growth markets, we came away from conversations with management with the sense that delays, or lack of traction in market share, are still very real possibilities in these early stage markets. This reinforced our view that the real bogey for meaningful growth in MXL’s revenue is at best late 2019 and beyond.

Separately, in the company’s core broadband segment we believe there is an underappreciated headwind to the DOCSIS 3.1 rollout. While the company has incremental dollar content on DOCSIS 3.1 modems currently being rolled out, in Comcast’s latest modems it also lost some legacy content. Our conversation with the company indicates this is a $1.5-$2 content headwind on ~25mm Comcast devices, over the rollout period. The company claims that this will be offset by other DOCSIS 3.1 rollouts - but our conversation with industry professional indicate that Cox, Comcast, Rodgers and Shaw all license similar gateways from Comcast. So it is possible that MXL has content headwinds aggregating more than 30mm subscribers. This converts to about at $10mm headwind per year assuming a 5 year rollout schedule of DOCSIS 3.1 gateways.

Our checks indicate little strategic rationale for the EXAR acquisition in 2017

With increasing competition, declining legacy products, and revenue from growth markets not showing up for years to come, we can see why MXL purchased EXAR. More specifically, towards the end of the Q4 2016 quarter, the CEO of MXL emailed EXAR unsolicited the day after EXAR closed the sale of its consumer-focused iML business and announced the acquisition by March 2017. We find this timing curious, as it coincides with the abrupt start of significant deterioration in MXL’s earnings quality (outlined below in the next section).

While an unsolicited acquisition is not an issue, our checks with industry professionals indicate little overlap and strategic rationale behind the EXAR acquisition which leads us to believe it was a rushed acquisition. The company has almost zero overlap in products, customers or channels. EXAR has hundreds of products and sells mostly 10 cents to a few dollar products through distributors. MXL sells far fewer products, at higher dollar content, directly to major OEMs. There is not an obvious cross sell strategy or sales force leverage to be had between the two channels.

Power management is also a commoditized and competitive space. While EXAR had some design wins to get it through 2017, we believe subsequent revenue visibility is limited. Our checks indicate that if there were cross sell opportunities with MXL, it would take at least 12-24 months of qualification/testing to get new design wins approved and therefore at least 24 months before meaningful cross sell revenue shows up.

At the time of the EXAR acquisition, MXL said the acquisition would be immediately accretive to Free Cash Flow. However, throughout 2017, the company reported the following LTM CFFO:

CFFO seems to have disappeared due to multiple factors: higher interest expense, tax payments, restructuring costs, acquisition costs, and working capital. While there was no one factor that drew cash, we think it’s clear that EXAR has not been “immediately” accretive to MXL’s free cash flow.

We also believe that the EXAR acquisition has been underperforming expectations. We back into its reported calendar year 2017 revenue based on MXL’s 10K and pro forma disclosures and compare it to consensus estimates for EXAR at the time:

We confirmed this general notion of an underperforming EXAR with the company. They noted three main issues that hit EXAR last year:

• Consumer win in its Force Touch segment faded as HTC sold part of its consumer handset business to Google and the product was not designed into any new handsets
• The long-awaited server power management win at HPE had delays and design issues
• Storage/encryption business wins did not materialize

In its Q4 2017 conference call, MXL also made confusing statements about the EXAR acquisition. They noted “softness across a range of power management and interface solutions” (possibly related to Force Touch) and then later noted how EXAR’s business was “difficult to forecast”. None of this gave us confidence that EXAR was a logical acquisition.

Accounting red flags throughout 2017 set up challenging comparable for 2018

Putting all of the above together, along with the following accounting red flags, we believe sets up ample evidence that MXL has more risk to missing 2018 numbers than is currently being priced into the stock. Based on careful review of the filings, we find the following red flags:

Increases in Days Sales Outstanding (DSO)

After doing our best to take out the acquired revenue, and acquired accounts receivable for EXAR and G.hn, we find that DSOs have been elevated over the last year, including 60-70% jumps in Q1-Q3 2017:

In its Q3 2017 conference call, MXL notes that increases in DSOs are driven by “revenue linearity” and “general lengthening of payment terms granted to some of our largest creditworthy direct customers”. In our view, elevated DSOs (especially in the face of declining organic revenue) is often indicative of channel stuffing, and/or pulling forward demand from customers. In this case customers are mainly distributors and OEMs like Arris or Technicolor. Granting terms, while it may be to customers who are credit worthy, essentially means that in some cases MXL could be putting product into the channel earlier than necessary. These tactics eventually need to work themselves out with future revenue reductions from customers as they realize inventories need to be right sized.

Increases in Days Sales Inventory (DSI)

We see a similar outlier change in DSIs (calculated based on revenue, not COGS) of 10-20% the last few quarters:

The company explains its inventory increase due to the EXAR acquisition - but adjusting for this we still see a meaningful increase. We view this as either excess fixed cost capitalization (providing a boost to reported income by expensing less costs) or an indication of elevated inventories due to slowing business activity. This is in addition to the aforementioned possibility that the company granted terms and potentially put excess product in the channel or with OEMs.

“Spring loading” of the EXAR acquisition

Our math indicates that it is possible that management held back EXAR’s sales to distributors until it closed the EXAR acquisition, and then recognized the revenue once EXAR was owned under the MXL umbrella. We approach this math in two ways. First, if we assume regular linearity of sales for EXAR throughout a quarter, it was owned for 48% of the 6/30/2017 quarter. However, looking at what was actually recognized relative to consensus estimates at the time, MXL recognized 58% of the quarters’ expected revenue. This equates to ~$2.7mm of “excess” revenue that was recognized from EXAR after the acquisition closed. Since MXL reported $104mm of revenue at Q2 2017 (versus $108mm guidance), this extra revenue was a meaningful ~2.7% bump to revenue in a quarter where MXL clearly missed guidance.

Another way we look at potential spring loading of EXAR is to calculate the implied full quarter revenue based on reported pro forma disclosures in quarter and adding back purchase accounting adjustments to revenue. We calculate that EXAR in the quarter would reported $24.8mm of total revenue, versus consensus estimates of $29.7mm. This equates to an almost $5mm difference in revenue, or -16% less than what should have been reported.

One explanation here may be that EXAR underperformed in its 6/30/2017 quarter - but at the same time it is possible that MXL encouraged them to underperform via delaying shipments to distribution so that revenue could be recognized under MXL’s ownership.

While the first analysis above could be explained by a normal non-linear pattern for revenues in the quarter, it is more powerful in combination with the second which suggests that the quarter was weak overall, but stronger late in the quarter. Either way, we find this flag troubling.

Odd trends in Deferred Profit account

According to its filings, MXL has some distributors that are based on “sell through” accounting. In this case, the company defers the revenue and profit from these sales to distributors until the distributor ships to its end customer, and the deferred profit is held as a liability on MXL’s balance sheet. In the last three quarters, we saw an odd buildup, then draw down, of this account:

An increase in this account coincides with the EXAR acquisition, but acquisition purchase accounting does not allow the company to acquire any deferred revenue (it is all written down within the purchase price). So the bump in this account is out of the ordinary. According to the company on its Q4 2017 conference call, this was due to “the increased distributor channel sell-through and a catch-up in a large customer's rebate payments that were accrued for but unpaid at the end of Q3 due to that customer's internal processing delays”.

While this may be the case, a drawdown on the deferred profit account indicates that an odd, and one-time, bump in gross profit was recognized through the income statement at Q4 2017 which will not be repeated next year.

Valuation

Putting it all together, we believe MXL inflated its revenue and profits throughout 2017, and obfuscated everything with acquisitions. Throughout 2018, however, MXL will have to lap these one-time bumps which sets up very challenging comparables from Q2 to Q4 2018E. The company will also fully lap the EXAR acquisition by Q3 2018E. With increased competition and lack of meaningful growth drivers throughout 2018, we believe this sets up considerable room for quarterly disappointments through the year.

Below we make some reasonable estimates for 2018E and 2019E revenue and compare them to consensus estimates. However we caveat our numbers below with the fact that if our analysis on DSOs and spring loading of EXAR is correct, organic growth could be much worse than what is shown below. In addition, given MXL’s high fixed cost base (and continued required R&D spending for new markets) and elevated DSIs, we believe operating margins have the potential to materially disappoint at any point during the year due to revenue misses, write-offs, or discounting. Finally, as mentioned above, given the gamesmanship in earnings in 2017, any one quarter has ample room for material disappointments in either earnings or guidance. Put simply, we “take the under” on the below model and the ~30% operating margins embedded in consensus estimates:

It is a challenge to set a firm price target for MXL shares given the accounting considerations. But we believe with earnings misses throughout the year coupled with possible push out of growth opportunities, MXL’s 14.6x NTM P/E multiple and 4.1x EV/Sales multiple have considerable room to fall. We would estimate a 3-3.5x revenue multiple and an 11-12x P/E multiple, yielding a price target in the range of $17, or ~30% downside from current levels.

Risks

• M&A (MXL sold) - With its current 4.1x EV/Sales valuation, a concentrated revenue base (25% to one customer) and lack of meaningful revenue contribution from R&D, we don’t believe anyone will be interested in a takeout of MXL at this time
• M&A (MXL buys) - it is possible that MXL buys another company. However, currently MXL is in debt pay down mode to shed the debt load from the EXAR acquisition. Throughout 2018 it will focus on debt pay down so the near-term risk of a large acquisition is muted.
• Any large announced or unannounced customer win which materializes into real revenue in 2018

Disclaimer

This report (this “Report”) on MaxLinear Inc (the “Company”) has been prepared for informational purposes only. As of the date of this Report, we (collectively, the “Authors”) hold short positions tied to the securities of the Company described herein and stand to benefit from a decline in the price of the common stock of the Company. Following publication of this Report, and without further notice, the Authors may increase or reduce their short exposure to the Company’s securities or establish long positions based on changes in market price, market conditions, or the Authors’ opinions with respect to Company prospects. This Report is not designed to be applicable to the specific circumstances of any particular reader. All readers are responsible for conducting their own due diligence and making their own investment decisions with respect to the Company’s securities. Information contained herein was obtained from public sources believed to be accurate and reliable but is presented “as is,” without any warranty as to accuracy or completeness. The opinions expressed herein may change and the Authors undertake no obligation to update this Report. This Report contains certain forward-looking statements and projections which are inherently speculative and uncertain.

Catalyst

Earnings misses and/or weak quarterly guidance throughout 2018