Description

NXP Semiconductors is a best-in-class semiconductor company trading at a meaningful discount to peers using conservative estimates of growth and margin improvement, driven by a broken merger, massive liquidity event, and show-me story that will play out over the next 1-2 years. High exposure to disruptive mega-trends such as autonomous vehicles,Internet-of-Things (“IoT”), computing at the Edge of the Cloud, and more should drive growth well into the future. In addition, large cash return and refocused management are near-term catalysts poised to re-rate the equity in-line with peers, if not at a premium. Under our base case scenario, we see NXPI re-rating to a price target of $105 (+39%) by the end of 2019, with a blue-sky scenario of $130 (+72%). Industry tailwinds and trough valuations support a downside of $67.00 (-11%), creating a 3.5x reward to risk with additional upside optionally.

One quick note. The initial model used for this report was created in Sept/early Oct. Since then, semiconductor reporting has been mixed, with certain players calling out no inventory/trade war issues (On Semi/TSMC/Infineon) while others have reported a slowdown (Texas Instruments). The below assumes no cyclical downturn or serious trade impact (we discuss these in detail below), but see model section for a second base case “short-term slowdown” implying a two quarter inventory flush and slight trade war related semi end-market slowdown. The price target even under this scenario is $89.70, implying NXPI is still almost 20% too cheap relative to comps factoring in results based on Texas Instruments expected slowdown and multiple change post-guidance of a semi downturn.

A variety of reasons explain why NXPI trades at a discount to peers, but the largest is the breaking of quite possibly the most crowded merger arbitrage trade in US history, leading to massive forced selling/de-risking. The natural buyer pool to offset this has been slow to react due to: (1) a lack of education on a company that has not been giving guidance, holding calls, updating their financial model, or generally engaging with the investment community in 2 years given the Qualcomm merger, (2) fear of an unengaged management (due to the same merger) and restrictions on business operation during the pendency of said merger, and (3) a lack of understanding of the NXPI that is emerging as a standalone company, given the evolution of the business over NXPI’s 8 year public life. NXPI IPO’d on 8/6/10 and was throughout much of its earlier years a much different company with lower margins, less disruptive end markets, and a shallower breadth of product offerings. The acquisition of Freescale (FSL) in 2015 changed the latter two, but led to a transformational year with margin erosion (integration) and higher debt. In mid-2016, NXPI set out to change this by announcing the divestiture of their lower margin Standard Products business (which targeted 30-34% gross margins and 20-24% operating margins vs their HPMS business they kept targeting 54-58% gross margins and 32-35% operating margins) in June of 2016 to Jianguang Asset Management and Wise Road Capital for $2.75bn. By the time Qualcomm entered the picture in late Sept 2016 (deal officially inked 10/27/16), NXPI was a much different company than it had ever been before. But, with the Standard Products divestiture not completed until 2017 and NXPI no longer holding earnings calls, marketing, or explaining the new value proposition to investors, most are still just now getting up to speed (NXPI held it’s first analyst day since 2016 on 9/11/18) on why this business deserves a much higher multiple than the NXPI of 2016.

NXPI operates in 4 main segments (the HPMS business), plus it’s Corporate & Other group. These are Automotive, Secure Connected Devices (“SCD”), Secure Interfaces & Infrastructure (“SI&I”), and Secure Identification Solutions (“SIS”). These four segments are exposed to the four major end-markets that NXPI targets: Automotive, Industrial & IoT, Communications Infrastructure, and Mobile. I break down, the four segments below by size, demonstrating the powerful revenue growth opportunities and leadership position of NXPI.

Auto (42.2% of 3Q18 TTM Revs [3Q18 est at midpoint of guide] at $3.98bn)

NXPI’s largest segment is it’s automotive segment, which focuses entirely on the automotive end-market as one would expect. NXPI holds the #1 global automotive semiconductor market position (~13% share, followed by Infineon at ~11%, Renesas at ~10%, and Texas instruments at ~8%) with key leadership positions in it’s core competencies (#1 market position in Connectivity, Radar [ADAS], Powertain & Vehicle Dynamics, Body & Comfort, Connected Infotainment, and Secure Network & Gateways).

Key products include in-vehicle networking (short range communication among vehicle systems and applications), entertainment (radio/audio such as amplifiers), telematics, keyless entry, radar sensors for ADAS functionality, and battery cell management for electrification of vehicles. NXPI is guiding for 7-10% revenue CAGR from 2018 – 2021, driven by positions in new applications (radar, network processing + Ethernet, ADAS, and EV) growing 20-28%, core applications (infotainment, powertrain, general purpose MCUs, secure car) growing 6-8%, and legacy products (8-bit MCU) falling -7 to -8%.

NXPI cites Strategy Analytics as having a 5-7% automotive semiconductor CAGR from 2018-2021, but other estimates are even higher. IC Insights for instance has a 12.5% CAGR from 2017-2021; Gartner has a CAGR of 7.1% from 2018-2021. Given NXPI’s #1 market share and broad portfolio, as well as high security-IP, NXPI should be able to outgrow the market by providing a complete platform/system solution to OEMs and Tier 1 auto suppliers.

Add on to that exposure to mega trends such as ADAS (there is too much to go into in this write-up, but to move from assisted driving ADAS levels 1-3 to Autonomous Levels 4-5 requires an increasing amount of sensors, especially Radar [as well as Lidar and Cameras] per vehicle, with estimates having the increase from Level 1/2 to Level 4/5 by as much as 15+ sensors per vehicle and 7+ radar sensors; NXPI has #1 position in Radar per their analyst day in Sept) and Electric Vehicles (NXPI focus on 25% CAGR power control in a market expected to penetrate 50% of all vehicles by 2030, up from <15% today).

This is a content story as global vehicle Total Addressable Market (“TAM”) increases at only 1-2%, but silicon increases at a much faster rate. Indeed, today the current vehicle has $380 worth of semi-content (implying ~$40 in content for NXPI).

• Medium-term, L-3 ADAS + Hybrid EV + Premium Infotainment can drive as much as $1,330 in semi-content per vehicle.

• Long-term estimates imply >$2,000 in semi-content, with NXPI being able to capture well over $500 in content. This can be seen in the BMW 7 series today, which has >$1,000 in semiconductor content and >$300 in NXPI semi-content.

Given these growth trends, NXPI’s market leading position, and the high-growth newer markets they are targeting, offset somewhat by tough competition (Infineon, Texas Instruments, Rensas in target markets; likely won’t compete in compute with NVDA and Mobileye/Intel), we forecast NXPI being able to grow at the mid-point of management guidance (we think conservatively), with an ability to accelerate in out years (but give no credit for this). Implies 2019e Auto Revs ~$4.38bn [8.5% growth], which is also slightly below their average annual growth rate implied by quarterly CAGR of 9.2% from 1Q16 – 3Q18, which has trended extremely steadily.

SCD (29.1% of 3Q18 TTM Revs [3Q18 est at midpoint of guide] at $2.75bn)

NXPI’s second largest segment is Secure Connected Devices, which is most exposed to the Industrial and IoT end-markets. However, despite the fact that most appear to extrapolate SCD numbers purely based off the Industrial & IoT end-market, NXPI disclosed at it’s analyst day that SCD is ~60% focused on Industrial & IoT, ~26-27% focused on Mobile, and ~13-14% focused on Auto. Based on the guided growth rates by end-market, this implies a potential SCD growth CAGR for NXPI of ~8% at the midpoint vs ~4.7% for the SCD end-markets:

The Industrial & IoT market is expected to grow at 3-5% CAGR from 2018-2021 [per Strategy Analytics as cited by NXPI; note that Gartner expects IoT semiconductor endpoints to increase at a 25.9% CAGR from 2016-2021, reflecting much higher expectations], while NXPI expects to grow at 8-11% over that same period (2.4x the market); the Mobile market is expected to grow at 4-7% CAGR over that period (per the same source), while NXPI expects to grow 4-6% (roughly in-line); and the Auto market is expected to grow 5-7% CAGR over that period (again per Strategy Analytics), while NXPI expects to grow 7-10% (~1.4x the market).

The SCD segment is comprised of ~2/3rds general purpose MCUs and ~1/3rd NFC/Mobile Transactions. The main drivers on the Industrial side here are the IoT end markets and broader MCU market in that Industrial and IoT section, with NXPI focusing on embedded microcontrollers, connectivity (short range wireless such as NFC, Bluetooh, etc), Security, and Sensors for driving smart industrials (factory automation), wearables, smart cities, and smart home, medical, and more. They will provide the technology driving communication and security between industrial robots, smart energy and appliances at home, an ever-cloud based and automated medical market, and more. They also provide audio and voice processing for Google Voice Assist, Alexa Voice Services, and Voice Solutions for China.

• Industrial/IoT & MCUs: In this 2/3rds of the market, NXPI is differentiated from it’s competitors based on a few main factors:

1. Scale as NXPI is the #2 MCU supplier (behind Renasas) and #3 ex-Auto MCU supplier (behind Microchip and Renesas, just ahead of STMicroelectronics) per Gartner [based on 2017 numbers]. The top suppliers have been GAINING market share in MCUs, driving a faster-than-average growth rate

2. Short Range Connectivity Expertise NXPI is a top Connectivity semiconductor supplier globally [#4 behind Broadcom, Qualcomm, and Mediatek], and has an especially strong position in NFC with >60% market share and positions in Bluetooth, Bluetooth Low Energy (BTLE), Zigbee, LE, and Thread. NXPI basically pioneered the standards for NFC, giving them a leg-up in the short-range connectivity space

3. Security As mentioned previously, NXPI is known for it’s security IP, and security is going to be key in the future connected wireless world as cyber-hacking can put businesses at risk of loss (think traditional cyber business loss and factory/grid shutdown) and people at safety risk (think connected hospitals, vehicles, and more)

4. Crossover Processors NXPI has invested in crossover processors, which is an expanding market opportunity growing from a negligible part of the market in 2018 to almost 10% of the embedded processing SAM by 2021 (almost $1.5bn per WSTS est). NXPI is currently the only player with crossover units between MCUs and APs per a sell side analyst

5. Investment in the Edge & Software Connected Solutions  NXPI’s focus and investment in edge computing as well as software based solutions to complement their hardware drive differentiation, higher margins, and a more sophisticated offering.

All of this should serve to drive higher-than market growth in MCUs. A PWC report [graphic per PWC below] has MPUs and MCUs growing at 6.8% between 2018 and 2019. Given management guidance for 8-11% in IoT end market CAGR, we conservatively estimate that on the 2/3rds of SCD business exposed to this that they grow at 8% between 2018 and 2019 [implying $1.95bn in ’19 revs for this portion of SCD)].

The other portion of SCD is Mobile Transactions (~1/3rd). This represents mobile payment apps and contactless payments via your cellphone at a range of vendors (ie Apple Pay). NXPI is the current leader in Mobile Payments with a #1 market share, which is also likely driven off of their short-range connectivity and NFC strength (uses the same technology). While a mature smartphone market expects 0-1% 3-yr CAGR growth, NXPI sees mobile growth of 4-6% built off of this leadership position and driven by low-end adoption.

NXPI currently sees NFC/SE attach rate for mobile wallets in 30% of phones, mostly restricted to Tier 1 OEMs (Apple, Samsung). As this expands to lower-tier OEMs (China OEMs such as Huawei) and wearables (Garmin, Fitbit), NXPI sees an attach rate of 50% by 2021. US adoption in general contactless payment has been disappointing (more on this in later segments), but mobile wallets are continuing to gain traction. Juniper Research said in June that the number of mobile contactless users will exceed 760mm by 2020, up from 440mm in 2018, a CAGR of 35%. Given NXPI’s NFC leadership, offset by lower pricing as it proliferates second tier phone adoption, we use a 12% growth rate for this segment, one third of the CAGR implied above, due to disappointing adoption rates in the past.

All of this drives a 9.3% blended growth rate in SCD, resulting in $2.96bn in 2019e revs. This is well below NXPI’s historical Q/Q growth rate annualized to imply a 19.6% CAGR going back to 2016 in the segment.

SI&I (18.7% of 3Q18 TTM Revs [3Q18 est at midpoint of guide] at $1.77bn)

NXPI’s third largest segment is SI&I, a segment most often thought of as Communications Infrastructure, but which also includes exposure to the Industrial & IoT and Mobile end-markets. This segment consists of: (1) Digital Networking Processors, (2) Secure Interface and System Management Products, (3) High-performance RF Power-Amplifiers (HPRF), and (4) Smart Antennae Solutions.

• Digital Netwokring: Digital Networking was a business inherited from Freescale that focused on communications and data center networking. This has been a non-core business in decline, and has in the past been mentioned as a divestiture candidate due to lack of synergies with NXPI’s portfolio (Forbes mentioned this in January this year and a WSJ article in Sept this year talked about how the NXPI/QCOM deal came about because NXPI CEO Rick Clemmer approached QCOM CEO Steve Mollenkopf about selling him the DN business). While NXPI doesn’t break out DN separately, it was a ~$1.0bn market in 2014 for Freescale. Digital Networking has been hit by competition (IBM), slowing China base stations, and capabilities being brought internally (PowerPC roll-up). NXPI has de-emphasized this market and is re-purposing much of the IP towards Auto and IoT. They also mention it is now a “small portion” of overall revenues, although they do not break it out (some estimates put it at ~$550mm - $600mm in revenues). NXPI mention that in 2Q18 this business declined -20% year-over-year in revenues. At a Deutsche Bank conference on 9/13/18, CEO Rick Clemmer said that the DN business is “…going to not be the same kind of impact it’s had before. It’s down ~5% of revenue now. It’s not going to have the material headwinds it had in the past.” At the same time, it doesn’t really stand to benefit. So we can assume this business will continue in decline, with our estimates baking in a 20% decline from 2018 to 2019, consistent with 2Q18 performance, despite the more recent update showing a slowing decline.

• HPRF: RF power is another key business here, and was also a 20% decline year-of-year in 2Q18. NXPI is the market leader in High-performance RF Power-Amplifiers (HPRF). However, NXPI at the DB conference in Sept pegged RF Power at ~5-7% of total company revs, implying ~$570mm in revs at the midpoint. NXPI says this segment is poised to benefit from 5G and adoption of mIMO, and NXPI are projecting 0-2% growth in this segment. Estimates for combined RFFE and Power Management revenues in 2019 could imply 6-7% growth in the space per Gartner. Despite this, we use 0% growth in 2019 for this portion of the business. NXPI also sees the Serviceable Addressable Market (“SAM”) in this business up ~7% CAGR from 2017-2019, increasing to a CAGR of 11.8% from 2019-2021.

• Interface Products: Interface products were the sole bright spot, up high-single digits in 2Q18 year-over-year. NXPI expects growth here from new high-speed interface standards such as USB-C, PD3.0, etc, which they see as having a SAM CAGR of 40%. NXPI sees the mobile business revenue increasing at 4-6% CAGR vs Market Research Future estimates expecting a Digital IC market CAGR of 7.6% from 2017-2023. We use a 5% growth rate on the remaining revenue in the segment, owing to new standard tailwinds, a historical growth position, and future Interface IC growth expectations.

All of this blends to create a -3% growth rate in SI&I from 2018-2019, implying a $1.7bn revenue for the segment, which we believe is more than conservative [extrapolating NXPI’s 3Q18 est revs for the segment + 1.5% growth in 4Q18 as that is traditionally the quarter-over-quarter growth in 4Q in the segment]. Also note that 3Q18 revenue in the segment is poised to accelerate for the first time since 4Q18. Moving forward, this segment has the ability to benefit from Wireless CapEx enhancement with 5G deployments, but we give no credit for that here. CapEx has been in decline for 3 years straight as 4G matured and spending was put on hold ahead of 5G. This is set to change for 2019+.

SIS (5.8% of 3Q18 TTM Revs [3Q18 est at midpoint of guide] at $550mm)

NXPI’s smallest segment, SIS, is entirely focused on the Communications Infrastructure end-market. This market was not really addressed at NXPI’s 2018 Analyst Day in September, likely owing to it’s smaller size. SIS represents NXPI’s secure smart card IC business, mainly focusing on chip-enabled cards for the following for markets: (1) Mobility and Retail [essentially Transit], (2) Banking [mainly EMV contactless bank cards], and (3) Secure ID [this is basically electronic, chip-enabled government documents like licenses and passports]. The devices here mainly consist of RF connectivity semis, microcontrollers providing security, and a software component for interacting with readers.

• Banking: This segment saw large declines and disappointment in 2016 as many were expecting US adoption of contactless payment cards, similar to China and Europe, which never really materialized. However, if you look at the market today, there are still many assuming that the EMV contactless payments will proliferate in the US. A Visa report has mentioned that in the US today, 38% of transactions originate from a contactless-enabled merchant, however many are using old MSD (magnetic stripe) technology. Visa is switching over on 4/13/19 to requiring all vendors support EMV contactless. Mastercard has similarly announced an initiative to force all terminals and cards to be EMV chip and contactless enabled by 2023 (cards by April 2019) in the Middle East, Africa, Europe, Latina America, and Asia Pacific. A.T. Kearny also recently released a study mentioning the amount of money banks could save due to the cost of cash by switching to contactless, and predicted a wave of contactless bank adoption increasing through 2019 and 2020. The study said <5% of cards in 2018 were contactless, but predicted a rise to almost 20% by 2019 and close to 40% by 2020. They based their report on data from Visa and other payment industry providers. Australia, the UK, Canada, South Korea, and China all have >50% of their cards equipped with contactless payment ability. Expectations are extremely low for this segment due to past disappointment, representing upside, however competition due to the commoditized nature of the product and Chinese vendors offsets that to some extent (and Infineon who has a lower-required profit hurdle for the segment). During NXPI’s Sept 2018 analyst day they said “We’re kind of through that period of time, where we’re seeing reduction. We’re not counting on a significant amount of growth, but we do have a leadership technology…we think that we’ve got a good position, but the bottom line is, we’re not expecting a lot of growth.”

• Mobility and Retail: Mobility and Retail refers to transit, where NXPI has the #1 position with >70% of mass-transit contactless market share, being deployed in >750 cities and >50 countries around the world. In 2016 during their analyst day, they presented a plan to have deployment in 10 cities in China. They now have >100 cities in China alone covering >500mm in Chinese citizens. NXPI expects an RMS of 5x the market in this segment. This is also the largest part of the segment at >50% per various broker estimates. The London Tube adopted the MIFARE standard contactless payment and now sees half of all of it’s rides using contactless payment (up from 25% just two years ago). NYC also announced a partnership with NXPI to begin using it’s MIFARE for subways beginning in mid-2019 and accelerating for 5 years thereafter. TSYS reports that Miami, Boston, Los Angeles, San Francisco, and Portland are also looking to go contactless. ABI Research has estimated that contactless ticketing will rise from 447mm to 550mm from 2018 to 2023, implying a CAGR of 4.2%. With NXPI’s 5x RMS, it could imply up to 20% growth for the largest piece of the segment.

• Secure ID: The final piece of the puzzle is eGovernment documents, with various broker estimates pegging at ~25% of the segment. This segment provides government identity cards with a digital format, allowing licenses, passports, etc to include digital information on the citizen/traveler. NXPI is the market leader providing solutions for chip-based cross-border passports, drivers-licenses, health cards and other government sponsored identification documents. A Research and Markets report has projected the Physical, Digital, and Mobile Government Identity Mobile ID growth CAGR of 24% from 2017-2022. It’s hard to really forecast the growth in this group, but SIS has been roughly flat since the beginning of 2017.

We assign 1% growth, the midpoint of the Communications Infrastructure end market, in 2019, representing increased adoption of transit and eGovernment solutions, offset somewhat by continued decline in EMV cards (further adoption more captured by low-margin competitors). There should be significant tailwinds due to contactless and digitization here, and other potential use cases include theme parks, interactive gaming, and supply chain tracking.  

So net, net we see NXPI growing it’s top-line at ~6% from 2018-2019, reflecting a notional revenue of $9.97bn [we forecast “Other” at 4% of revenues, roughly in-line with 2018, due to continued wafer and foundry supply agreements from Standard Products and FSL’s HPRF divestitures; this is a no gross margin business that they hope will overtime decrease].

Margins

The other major piece of the puzzle here is NXPI’s gross margins and operating margins. As mentioned above, these have historically been much lower than the high quality analog peers such as Texas Instruments and Analog Devices. And even still NXPI is targeting margins below those of their peers. Looking back to pre-FSL days, NXPI was running consistently <25% operating margins, which basically increased every single year (and have continued to do so through FY17). From the completion of the FSL merger onward, we saw gross margins of ~50-51% in 2016 as they suffered a drag from the lower margin Standard Products business, with operating margins in the ~26-27% range. After divesting that business, gross margins jumped to ~53% in FY17, with operating margins ~29%. We have also seen some weakness as we entered the early part of 2018, with 1H18 gross margins at ~53% vs 2H17 at ~54% and 1H18 operating margins at ~27% vs 2H17 at ~31%. On their earnings call for 2Q18, management said high wafer costs (80bps) and product investment have caused margins to be slightly below the long-term model, which underperformance they said they are not worried about changing their goals structurally.

They also called out ~80bps of margin hit from the QCOM merger in terms of gross margins, as well as product mix. Management plans to exit 2019 at a 55% gross margins. They have projected ASP Erosion of -160bps offset by mix improvement of +60bps and operational executive of +260bps. On operational execution they have said they’ll have pricing improvements of ~150bps, and the remainder will be mainly back-end execution as they have mostly finished off front-end fab improvements. So we’ve had this linear improvement since inception, which has kind of tailed-off in the recent two quarters if we look at gross margins.

Management has mentioned they will try to get to 55% gross margins earlier than exiting 2019, but won’t commit to it. To get to the higher end of their guidance, up towards 57% gross margins, they say they’d need to divest a non-core lower margin business (~100bps) or invest capex in positive return commitments (dedicating 10% of the Qualcomm termination fee to this). Giving no credit to either, it is however worth recognizing these are levers at management’s disposal.

So where does that leave us? If we grow margins linearly to hit their 55% target exiting 2019, average 2019 gross margins will be ~54.6%. They have increasing portion of the business coming from Auto and SCD, and those design wins typically come ahead of time. We also model a better mix than the company, simply due to worse expectations for SIS and SI&I. Thus, we think they will be able to largely drive their mix and volume given the above. To be conservative, we bake in the full ASP Erosion. And then we are left at operational execution at ~260bps. If we assume they can hit 50% of these targets (conservatively as management seems to have good view into this), we get gross margins of ~53.8% expanding linearly through exiting 2019 (really just around this level). Operational margins are tied to gross margins and management has said R&D trended higher when the Qualcomm merger was ongoing as they thought they’d be one piece of a bigger puzzle. So they should have some levers to pull there. If they pull R&D spend down to the middle of the target range of 14-16% and keep SG&A at the high-end of the range at 6-8%, we hit ~30.8% operating margins on this number.

EPS, EBITDA, and OCF

Some final items to round out before we get our ‘19e EPS, EBITDA, and OCF estimates. Management has guided to ~8% in non-GAAP cash taxes + $166mm in incidental taxes (past divestitures). They have also increased their stock-based compensation to $365mm for FY19. If we put this all together, we get EPS of $8.26, or $7.04 excluding stock-based compensation [can see share count info below in capital structure]. We get $3.4bn in EBITDA using a 3.4% D&A rate and $2.7bn in Operating Cash Flow using a 7% CapEx rate [management guidance for 2019 specifically on the 2019 Analyst Day]

As noted above, the impact of a semi slowdown could end up hitting NXPI’s numbers. Texas Instruments was the main semiconductor company to report a slowdown. Their 4Q18 Q/Q revs were -12% (vs normal cyclicality of closer to -7% to -8%) and they missed by -6%. An average of 2016 and 2017 4Q Q/Q growth for TXN is -4.6%, implying ~7.5% underperformance. For NXPI that number is +1.4%. In our new secondary base case scenario, we use a -6.8% Q/Q growth rate for 4Q18 (with 3Q18 as the midpoint of the guided range as management reiterated that guidance 82% of the way through the quarter), about in-line with TXN’s miss and Q/Q slowdown vs the past 2 years [with Auto -9% due to auto slowdown commentary, SCD -7% due to China slowdown commentary offset by strong IoT trends, and SIS and SI&I -4% due to cyclicality].

That results in revenues of $2.27bn in 4Q18, resulting in FY18 revs of $9.266bn. We assume no growth in 1H19 (despite the typically strong first half of the year and NXPI’s tailwinds; except SI&I which we conservatively keep at a -3% run-rate), and grow in 2H19 as the same run-rates as baked-in for FY19 in our base case model. Total revs become $9.5bn in 2019e (down from almost $10.0bn), with margins contracting to 54.5% GMs and 28.5% OPMs. See valuation impact below.

Capital Structure

If you take NXPI’s capital structure on 2Q18, they had ~$5.3bn in debt and ~$3.0bn in cash. They also had 347mm shares. They received a $1.5bn post-tax termination fee from Qualcomm in July, and proceeded to buy back 39.9mm shares for $3.76bn through 9/10/18. They are projected to generated ~$1.5bn in cash flow through 2H18 and buyback the remaining $1.24bn under their buyback capacity through year-end (~13.8mm shares at $90 per share, which seems conservative given the current price around $80). This will leave them with ~293mm shares. We add ~6mm shares for potential stock-based compensation, as they mention $365mm in stock-based comp for 2019 (~$4mm shares at $90, and include half this number for 2H18). This is conservative as most of these options will likely be OTM, but the impact is relatively small anyway. Total share count ~299mm. They also announced a $0.25 dividend, which should pay-out another $153mm in cash in 2018. Using the 293mm shares, it will reduce cash by an additional $329mm in 2019 [don’t increase to the long-term target of 20-25% of cash flow as management said with stock below $95 they’d be more inclined to buyback vs bump the div]. And using the CapEx from above, we get a cash flow of ~$2.7bn in 2019, bringing total debt to ~$2.1bn and Net Debt to EBITDA of just 0.6x by the end of 2019.

No 2019 Buyback

Management has also said they feel comfortable increasing Net Debt to EBITDA to 2.0x and will return cash to shareholders, which means in theory they could take on an additional ~$5.0bn in debt to buyback shares, an additional ~59mm shares, or 16% more outstanding, bringing the total to 31% of shares outstanding as of 6/30/18). In the below valuation, we do not give credit for this as we will explain that the market is currently somewhat worried about debt-levels in semiconductors. But, it is worth noting that they will have a large amount of dry powder for accretive tuck-ins or cash return to shareholders, additional levers to drive meaningful upside. And NXPI management is very seasoned in operating at leverage levels well above 2.0x [having done so for virtually the entire public life of the company], and has stated that the current <1.0x Net Debt to EBITDA levels are not where they want to be. While the market appears to be rewarding prudence around debt-levels at the moment, inefficient balance sheet utilization doesn’t seem like a good goal in a sector ripe with R&D opportunities and for NXPI specifically looking undervalued. This is a company doing close to $3.0bn in operating cash flow per year with $5.0bn in debt, even at a net debt to EBITDA of 2.0x they are in a very comfortable range.

2019 Buyback to 2.0x Net Debt

Valuation

For valuation, we look at NXPI’s closest product and end-market focused comps. These include analog-MCU peers with similar end market exposure [Auto and Industrial microcontrollers] and more broad-based peers with similar end market exposure [mainly auto application-specific standard products (“ASSP”) and wireless connectivity ASSPs]. This creates a comp set of Analog Devices, Texas Instruments, Infineon, Maxim Integrated Products, Microchip Technology, On Semiconductor, Power Integrations, Renesas, Silicon Laboratories, and STMicroelectronics. We exclude other oft-cited comps Intel and Broadcom as they are much more exposed to the data processing/compute markets [for Intel], wired connectivity [for both Broadcom and Intel], and handset/wireless capex trends [for Broadcom] than NXPI and do not have the same auto exposure as NXPI [with the exception of Intel’s Mobileye acquisition, which is focused on processing as opposed to edge compute and microcontrollers]. (NOTE: STM IM is actually 5.6x EV/EBITDA)

Looking at NXPI relative to this comp set, it trades at a fairly large discount on all metrics. The above peer set trades at an average of 14.2x ‘19e P/E (normalized to CY19 and excluding NXPI), with a range of 8.8x (On Semi & Microchip) to 20.6x (Power Integrations). NXPI trades at just 10.7x. Only Microchip and On Semi trade at lower P/E multiples than NXPI. The picture is similar looking at the EV/EBITDA, with peers trading at an average of 9.9x EV/’19e EBITDA, while NXPI trades at 7.2x (with only On Semi, Renesas, and STMicroelectronics trading lower). On an EV/FCF basis (using OCF as a FCF proxy for NXPI), peers trade at 15.3x ‘19e FCF, with NXPI trading at 9.1x, with only Renesas trading at a lower multiple.

‘19e P/E Valuation

‘19e EV/EBITDA Valuation

‘19e EV/FCF Valuation

One note on the above. we value NXPI using the forward capital structure, which could be a potential difference between our numbers and what you would see elsewhere. We think it’s only natural to look at NXPI’s future capital structure if we are talking about future EPS and EBITDA, especially as it is projected to change in a material way and impacts the EPS and EBITDA numbers. However, it is irrelevant to the picture, as NXPI screens extremely cheap even using the current capital structure. The question then becomes, is this discount justified?

The answer, in our opinion, is emphatically no. We can look at growth, profitability, shareholder returns, and debt levels. From a revenue growth standpoint, NXPI’s 2yr revenue growth profile between 2017 and 2019 is indeed on the lower-end of comps. But this is due to 2017, a year during which the pendency of the Qualcomm merger slowed supplier orders. Looking forward, our projected year-over-year growth rate for NXPI is actually above the average of comps at 4.7%, coming in at 5.96%. Furthermore, if we adjust for Microchip, who gained much of their growth via the recently completed Microsemi merger (increasing their revenue by roughly 50%), NXPI is even further above the adjusted average of 4.2%. And as mentioned, NXPI expects to grow at a CAGR of 5-7% from 2018-2021, which is higher than the average Gartner forecasts for all semiconductors over that period (4.8%) and higher in NXPI’s specific addressable markets of 3-5% as they mention. Analog Devices and Texas Instruments are actually projected to grow towards the lower end year-over-year from 2018 to 2019, but as we will see next their profitability helps explain why the command premium multiples (as well as exposure to attractive IoT/Industrial and Auto end markets which will drive long-term growth, but this is actually even MORE the case for NXPI). One quick note is that we are in the middle of earnings season. Texas Instruments and Power Integrations have swung their estimates from +~4% and +~10% growth in 2019 to -1.8% and +1% respectively. On Semi’s growth target has not moved much (beat and raise quarter), but they now look relatively more attractive on this metric than they did when I first wrote this up at the beginning of October. The lower growth partially explains their multiple, but also the consistency vs other players explains why it has actually seen a P/E expansion of 0.2x vs the space contracting 0.6x over the past month. If NXPI was to miss on 4Q18 revs by a similar amount as TXN and have no growth in 1H18, It is worth keeping in mind that these are currently moving targets (will get into the cycle and trade war below), but the long-term valuation discount of NXPI and growth drivers stand regardless of these  transitory headwinds.

Looking at profitability as measured by gross margins, operating margins, and EBITDA margins, the picture is very similar. NXPI’s gross margins of 53.8% are again the median of comps and right near the midpoint of 54.3% for all comps. Analog Devices and Texas Instruments have gross margins of 71.1% and 65.5% respectively, helping to explain why they have (and have historically) commanded premium valuations in comparison to NXPI. We also note that the lowest-margin comps are typically the European and Japanese competitors, Renesas (46.9% gross margins), STMicroelectronics (40.9% gross margins), and Infineon (38.0% gross margins). US semiconductors are known to have a different operating model and margin profile than these overseas competitors, which is a huge part of why Renesas (11.7x P/E; 5.0x EV/EBITDA; 8.5x EV/FCF) and STMicroelectronics (10.4x P/E; 5.6x EV/EBITDA; 13.1x EV/FCF) have lower valuations in general. The other notable comp with margins far below the average? On Semiconductor, which trades at some of the lowest multiples (8.8x P/E; 6.2x EV/EBITDA; 10.0x EV/FCF). Infineon likely receives a higher multiple despite this because it has the highest projected growth rates across our comp group, although it’s EBITDA multiple is pretty low. The picture is similar for operating margin, although NXPI is actually slightly above both the median and the average here, with an operating margin of 30.8% projected vs an average of 26.95%. Again, we see the lowest margin players are On Semi (17.6% OPM), Infineon (16.9% OPM), STMicroelectronics (13.% OPM), and Renesas at the bottom (11.8% OPM), while the highest margin companies are Texas Instruments (42.6%) and Analog Devices (42.6%). EBITDA margins also have the same picture, actually having the comps ranked in the same order with NXPI again just above the median and average of the range. Profitability is important because (excluding SLAB due to it’s high EV/EBITDA multiple), we can see a relatively-high R2 value when running a regression between valuation and margins (gross margins, operating margins, and EBITDA margins) of >0.6 (INCLUDING NXPI which skews the results). The regression implies that based on margins, NXPI should be trading at 10.0x EBITDA at a minimum.

‘19e Gross Margins

‘19e Operating Margins

‘19e EBITDA Margins

Regression of Operating Margin vs EV/EBITDA multiple (note: as of 10/19/18)

Moving to shareholder returns, if we look at the % of FCF returned to shareholders (as returned capital ex-debt / FCF), we also can see why certain semi stocks garner higher multiples. Those who return >50% of their FCF from our comp group our Power Integrations (returns 161% of FCF; trades at 20.6x P/E, 13.1x EV/EBITDA, and 13.5 EV/FCF), Texas Instruments (returns 108% of FCF; trades at 16.2x P/E, 12.3x EV/EBITDA, and 14.2x EV/FCF), Maxim Technologies (returns 104% of FCF; trades at 16.7x P/E, 12.0x EV/EBITDA, and 14.8x EV/FCF), and Analog Devices (returns 68% of FCF; trades at 15.7x P/E, 11.7x EV/EBITDA, and 13.9x EV/FCF). While NXPI has not been returning much cash to shareholders over the past two years, this year they will be returning >200% of their FCF (mostly due to the return of the Qualcomm termination fee via the $5.0bn buyback) and they have said they will payout close to 100% of their FCF to shareholders going forward as that is the “owners money” and they want to return it to shareholders. If they raise debt to get to 2.0x Net Debt to EBITDA in our upside scenario, they will again be returning well over 100% of FCF for a second year straight. As is clearly seen above, these companies all trade at above average valuations on all three metrics (after correcting EV/FCF for Infineon, which trades >2x the average).

Finally, when assessing the balance sheet health, the picture is similar. Many semi companies trade at negative net debt levels, and in general these garner some of the highest multiples (Silicon Labs, Power Integrations, Infineon, Maxim Integrated Products, Texas Instruments). The main exceptions to this are On Semi (1.3x Net Debt to ’19e EBITDA) and Analog Devices (2.0x Net Debt to ‘19e EBITDA) to a lesser extent, and Microchip Technologies (4.0x Net Debt to ‘19e EBITDA) and Renesas (jumps to 3.7x Net Debt to ‘19e EBITDA with the Integrated Devices acquisition they announced and will fund with ¥679 billion in debt) to a much greater extent. NXPI should end 2018 with a capital structure implying 1.3x Net Debt to ‘19e EBITDA, putting it in the same category as Analog Devices and On Semiconductors. While the market appears to slightly discount the valuations in response to debt, as seen by Analog Devices (which garners a slight premium valuation), this factor alone should not result in a significant discount. On Semiconductors is punished more for it’s margin profile and poor cash return in addition to it’s debt levels than for that alone. This is all before taking into account the fact that NXPI has experience managing debt at these levels and even if they decide to take Net Debt to 2.0x for their buyback should have no trouble meeting their financing commitments (not to mention how accretive those buybacks would be at these levels).

Add to these quantitative factors more qualitative aspects such as NXPI’s market leadership positions and exposure to attractive end markets like Industrial/IoT and Auto and there is absolutely no justification for this discount. We can clearly see the reasons for those companies that trade at significant discounts on one or various metrics, whether it is much poorer margin profiles (10%+ lower EBITDA margins than NXPI), much higher debt levels (~2.0x+ the HIGHEST level NXPI will conceivably have in Net Debt), lower shareholder return levels (<50% vs NXPI projected at 100%+), or revenue growth and market exposure (typically combined with one of the above as the regression on revenue alone doesn’t show much correlation). There is absolutely no justification for NXPI to trade at the multiples it does, and one might argue it could even deserve a premium multiple because it is at least average if not above average on every single metric, despite not having the margin profile of stronger comps like Texas Instruments and Analog Devices. But, as I explain below with our price target, this is absolutely not necessary to make this an extremely attractive investment and thus we conservatively use in-line multiples.

Price Target

As noted above, the broader peer group trades at 14.2x P/E, 9.9x EV/EBITDA, and 15.3x EV/OCF. On a P/E basis, our model shows NXPI generating $8.26 in EPS in 2019 excluding stock based compensation and $7.04 including it. There is disagreement in the market on how to value NXPI, and various comps report EPS with and without the inclusion of such stock based compensation. For the sake of being conservative, we will value NXPI including stock based compensation on their multiple of 14.2x, implying a price of $99.97 (as a note, if we used the EPS excluding stock based compensation, the price target would be $117.29). Looking on an EV/EBITDA basis, we get a price of $112.89. And on EV/OCF, we get a price target of $131.80 (as noted previously, NXPI trades at the cheapest on this metric relative to comps). So while a single valuation metric relative to a comp like Analog Devices may have you question the valuation level, it is clear that across all the valuation metrics NXPI is trading much too cheaply. A price target of $105 is a blend between slight discounts on EV/EBITDA (10.0x implying $107.00) and P/E (14.0x implying $99.00), while still remaining extremely cheap looking on an EV/OCF basis (12.3x, below all comps besides On Semiconductor and Renesas). Thus our base case is a 39% discount to fair value.

Downside protection comes from NXPI trading at essentially trough multiples already. Since 2014, NXPI traded this low on P/E basis only once, EV/EBITDA around the same period, and it is slightly above its trough EV/FCF level. If we assume that we see a temporary semiconductor cyclical downturn (more on this in the risks section below) and year-over-year growth in NXPI is actually -3%, we get total revenues of $9.5bn. We also contract gross margins by half the 130bps operational improvement (lose all pricing power) and also pull out the 60bps of mix improvement, resulting in gross margins of 52.6%. We also take operating margins to the lows post-Standard Products divestiture of 27.6%. Using these inputs, we see ‘19e EPS of $5.69 including stock based compensation, EBITDA of $2.9bn, and Operating Cash Flow of $2.2bn. Using trough multiples since 2014 of 9.0x, 8.5x, and 11.0x respectively on these numbers, we get a blended price of ~$67.00 on the downside, or -16.2% (note that P/E especially drags this down from a blend closer to ~$71.00).

So we have a 3.5x risk/reward before (before probability weighting, which would skew it higher), but we also have not looked at the bull case scenario yet. Without giving any additional revenue growth credit (I’ve mentioned above why we feel that our current estimates are conservative, but given the long-term outlook is 5-7% revenue CAGR, we think jumping immediately to the high-end would be too aggressive in valuation, even though our personal belief is management set a low hurdle). If we take margins to the 54.6% amount in 2019 that we would get by linearly growing to exit 2019 at management’s 55% estimate and grow operating margins to 31.6% in accordance, we greatly improve the numbers. We also adjust the capital structure to account for taking Net Debt to EBITDA to 2.0x and buying back an additional ~57mm shares. Using our new ’19e EPS of $8.99, EV/EBITDA of $3.5bn, and EV/OCF of $2.8bn, we get a blended price target of ~$130.00 in our blue sky scenario, or 72% upside. And as I note above, this is not, in our opinion, an implausible scenario as much as it is not the highest probability outcome.

In our short-term semiconductor/trade war slowdown, we have the following metrics for NXPI: $5.69 in EPS [without bumping the buyback], $2.4bn in OCF, and $3.1bn in EBITDA. We value NXPI relative to Texas Instruments, as their price is baking in the impact of the trade war.

• P/E: Pre-reporting, Texas Insturments traded at a 1.8x premium to the market, and despite their specific guidance miss, the space actually contracted more than they did, with Texas Instruments now trading at a 2.0x premium to the market. Thus, on a P/E, we conservatively value NXPI at the market rate (not adjusting for the 0.2x turns Texas Instruments gained). This implies a short-term price of $80.80 (although excluding stock-based comp the price is >$100 and factoring in a lower buyback (due to lower cash flow) of $3.4bn implies $6.80 in EPS at 2.0x turns and a $97.00 price at the comp multiple

• EV/EBITDA: Again, Texas Instruments actually increased it’s premium to the market on this metric, going from a 2.0x premium toi a 2.4x premium. Again, we conservatively value NXPI at the current market average multiple.This implies a short-term price of $91.85, with a buyback bumping the price to $92.27

• EV/FCF: Texas Instruments actually didn’t change on an EV/FCF basis, staying at a -1.1x discount to the market average. Because NXPI generates much more cash flow and always screens cheap on this metric, we actually use a 2.0x turn DISCOUNT to Texas Instruments (~15%), which is a 22% discount to the market multiple. This still implies a price of $88.16 with no buyback and $88.05 with the buyback.

• Taking a blend of all of the above gives us an $89.70 blended price target (50% buyback and 50% no buyback), implying NXPI still is almost 20% too cheap when factoring a cyclical downturn

Supply/Demand Imbalance

So where does the opportunity come from? At it’s simplest, it is a pure supply/demand imbalance driving a temporary pricing dislocation. As mentioned above, there is a fundamental misunderstanding of NXPI as a standalone company, a company that has a different margin profile and debt profile than it has in its entire history as a public company. Add to that the fact that it is coming off a failed merger after two years where they have been unable to directly engage with their long-term investor base to educate them on the development of the company, and into a sector with caution nonetheless, and you have an explanation for the lack of demand. As for the supply, NXPI was considered one of the most crowded event-driven/arbitrage trades of all-time. These shareholders are not long-term semi investors, and given their mandates or lack of understanding, they are either being forced to liquidate positions or selling to limit their own perceived risk due to the unknown.

Graph of Relative Performance of NXPI vs our Comp set

(Note: through 10/2/18. Since then NXPI is -11.2% vs Comps -14.2%, so some outperformance)

Risks

So what are the risks to the NXPI investment thesis, and what could cause the bearish price of ~$67.00 to come about? There are four main worries here (in order of importance): (1) Declining Semiconductor Cycle, (2) Auto/Industrial/IoT slowdown, (3) NXPI has lost ground due to the Qualcomm merger, and (4) NXPI has struggled to hit historical growth targets.

In regards to the declining semiconductor cycle, there have been a multitude of analysts calling for a decline in the semiconductor cycle, driven largely by checks of inventory build/double ordering, historical billing/semiconductor sales above the long-term trend line (a sign of potential double ordering and inventory building), declining lead times after over a year of quarter-over-quarter increases, and concerns over the macro picture driving decreased ordering [this will also tie in with the second risk]. While some of the above signs are cause for caution, and the primary reason for the recent retracement of stocks across the space, the reporting so far has been extremely mixed. We have commentary from market participants stating the opposite. NXPI themselves said on their 9/11/18 analyst day that if there was a change to their Q3 guidance, they would have updated the market. Taiwan Semiconductor recently reported 3Q18 earnings, and was specifically asked about inventory levels. While they had said inventory management could impact revenues, they said that even if there was some kind of inventory adjustment, inventory levels are lower than this time last year and they don’t think the impact on next year will be as severe as this year. Infineon recently commented “…July was strong, August was strong, still the September again was actually increasing the number of the accumulated orders on hand as well as the number of unconfirmed orders was still higher…”, on their 10/2/18 M&A Call. Lam Research, a semicap equipment company whose customers include the semiconductor companies, also recently reported strong Q3 earnings and a BEAT on 4Q18 revenue guide. On Semi reported on just 10/28/18 a beat and raise quarter, calling out strength in certain Auto and IoT areas (which NXPI should benefit from), although noting some weakness in China (which NXPI could suffer from). TXN has been the biggest semi to report bearish news thus far, missing their Q4 guide by 6%, calling out auto and industrial slowness, and mentioning a cyclical downturn. However, even in the event we have an inventory correction in this space, we should begin to see the recovery to growth by 2H19, and this ignores all the strong tailwinds that are currently coming to a head in the semiconductor space. Autonomous vehicles, 5G, IoT, Factory Automation, etc are not pausing and will create increasing demand, likely offsetting impacts from any elevated inventory levels.

The Auto/Industrial/IoT slowdown is really predicated on the fear of the trade war and what slowing Chinese demand could do to the semiconductor market. In Auto’s specifically, it also relates to the WPLT regulations and slowing of auto production. But again, we see real industry players directly refuting this point. Taiwan Semiconductor specifically said they have not seen any short-term impact on demand from China due to the trade war. NXPI on their analyst day specifically mentioned how they are Dutch which actually benefits them in the trade war and how they have been able to reduce the impact on customers to a de-minimis amount by the way they ship. Infineon said on that recent 10/2/18 M&A call “I don't see an impact on the order picture coming from China.” Even Texas Instruments, who guided below estimates, called out a slowdown, and had somewhat weaker Auto and IoT results, said they have seen minimal tariff/China impact. On Semi did call out some China weakness though, as have various Auto players (although there are already reports of China preparing to prop up the Auto market). NXPI also specifically mentioned how they do not see a material impact from WPLT near-term and negative profit guidance from a specific European Auto customer was a company specific issue. Offsetting both the semi cycle and China demand picture as well is the fact that on their 7/26/18 earnings call NXPI explicitly mention how SCD has been limited by capacity (MCUs) and that auto growth has slowed a bit due to a shortage of discrete capacity from other semiconductor companies. These supply issues mean that even if demand were to snap back to some extent, the supply has been a constraining factor to date and should offset any impact (current demand levels can’t even be hit for NXPI).

In regards to NXPI losing ground over the course of the Qualcomm merger, the main factors we can point to are that NXPI still has a leadership position across virtually all of it’s focus markets. They also mention how they have actually INCREASED their R&D investment over the pending Qualcomm merger. At their September Analyst day, commenting on why margins are not where they want them yet, NXPI Management said “And you definitely caught us out on R&D. So R&D, while we were into Qualcomm, we thought we can invest pretty much however we'd like and there's some great opportunities there for us to do things. So we spend more in R&D in the past 12 months than we would've done had we been an independent company.” NXPI has actually been funding MORE in R&D then they historically might have, which should be increasing their technological edge.

The last point is in regards to NXPI not hitting it’s historical growth targets. In some ways this is not true, as NXPI actually hit their growth targets, growing 10% top-line vs guidance of 5-7% (split between Auto and SCD outperforming and SI&I and SIS underperforming). However, the broader semi space grew 22% in 2017 and will grow 12% in 2018, while NXPI’s guidance contemplated them outgrowing their market. The only real point we can make on this is that it is backward looking, NXPI was in the midst of a deal, and the model revenue growth trends sanity check (again the drag being due to legacy Digital Networking and Bank-card market saturation).

Catalysts

Catalysts near-term to re-rate the stock include the turnover of the shareholder base from Event Driven/Arbs to long-term investors, management hitting their numbers, a resolution to the trade war with China (wouldn’t count on this, but represents upside), the resumption of the buyback after the earnings blackout period, and management re-engagement with the investment community.

Disclosure

I do not hold a position with the issuer such as employment, directorship, or consultancy.

I and/or others I advise hold a material investment in the issuer's securities.

Catalyst

Catalysts near-term to re-rate the stock include the turnover of the shareholder base from Event Driven/Arbs to long-term investors, management hitting their numbers, a resolution to the trade war with China (wouldn’t count on this, but represents upside), the resumption of the buyback after the earnings blackout period, and management re-engagement with the investment community.