Is Qualcomm a good Buy?

Published on 03/31/21 | Saurav Sen | 4,995 Words

The BuyGist:

  • This is our thesis as posted on March 31st, 2021 - now in magazine format.
  • The Thesis Summary (somewhat redacted) is available to all readers. 
  • Thesis details - including 1) BUY or HOLD or IGNORE conclusions 2) whether we still hold the stock, and 3) Valuations - are for members only, accessible upon login. 
  • Please see our subscriptions plans for full access, if you haven't already.

Thesis Summary

Qualcomm chips are at the core of the global 4G to 5G transition. To buy its stock, we need to believe that Qualcomm will 1) maintain its dominant market share in 5G chipsets and 2) will make more inroads into ancillary growth markets like AVs and IoT. In the sections below, we dig into the positives and negatives of the story to see if we can believe it.

Positives

Competitive Advantage & Moat: Most focused company in 5G Chips. Full stack offering.

It’s hard for us non-engineers to pinpoint exactly why Qualcomm’s chip design & architecture are better than their rivals’. So, let’s start with the assumption that there isn’t some inherent genius in their chips. Let’s try to stick to facts. Here are some important ones:

  1. For decades – from 1G to 5G – Qualcomm’s core competency has been the Cellular Modem.
  2. Today, it remains the most focused company in the cellphone chip industry.
  3. This has translated into a high market share of about 30%.
  4. Qualcomm dominates the high-end segment cellphone chip market.
  5. Qualcomm offers a “full stack” set mobile chipsets – including Sub 6, mmWave and RF Frontend.

Does Qualcomm make the best chips? If we go by their market share in high-end phones, then we’re inclined to believe that. Even Samsung – one of the largest cellphone manufacturers in the world used Qualcomm chips in many of their phones. This is worth mentioning because Samsung also manufactures its own chips. That says something. Let’s just agree that they make “one of the best cellphone chips”.

They key question now is: what prevents competitors from driving Qualcomm out of business? In other words, what is Qualcomm’s Moat? We can distill it down to 2 factors:

  1. Qualcomm’s end-to-end presence in cellular chipsets.
  2. Huawei and Hi-Silicon’s retrenchment from the cellphone chip world.

It’s the combination of these two factors that’s the real economic moat. Qualcomm’s involvement with 5G technology (and previous generations) is end-to-end in the sense that they get involved right from the start – from Standards all the way to production. Standards are a set of broadly accepted technological rules that all participants of the 5G ecosystem – chip makes, handset makers, backhaul equipment manufacturers etc. – agree upon. They need to have common set of rules to ensure that all their products talk to each other. This multi-year process of setting Standards occurs before every generational leap, going back to the first generation of mobile phones. Qualcomm has been at the forefront of all standard-setting efforts since the beginning. Here’s an estimate of 5G Standards contributions by company – Qualcomm ranks #2. But their management argued (during their November 2019 Analyst Day Presentation) that in terms of quality (however that’s measured) they rank #1. In terms of quantity, Huawei ranks #1.

Speakin’ of Huawei – they’ve had a horrible 2020 thanks to the Trump Administration embargo. This has seriously shaken up the 5G world, and it may work in favor of Qualcomm. Huawei is (was) the most prolific 5G company in the world. They are (were) truly end-to-end. They’re involved in Standards. They’re the biggest 5G backhaul equipment manufacturer. The make cellphones AND they make (made) the chips that go in these cellphones. In terms of being the true end-to-end 5G company, they’ve beaten Qualcomm fair and square. Until now…

The Trump Administration embargo was a gut punch to Huawei. One of the biggest fallouts of the embargo was the restrictions imposed on TSMC (one of our largest holdings) and other chip foundries to manufacture HiSilicon chips. HiSilicon is Huawei’s semiconductor arm. In the 5G era, they have been one of Qualcomm’s main rival. And suddenly, HiSilicon could not get its flagship Kirin chips fabricated anywhere. Based on this data, that’s about 12% of the 5G chip market is now suddenly evaporating. But keep in mind that HiSilicon made chips only for Huawei phones.

So, what will Huawei do? Will it abandon its flagship smartphones? It appears that, at temporarily, Huawei is willing to use Qualcomm Snapdragon chips in its flagship phones. And, if this report is to be believed, Qualcomm chips will now being used in Honor phones. Honor used to be Huawei’s “lower-end” phone brand, which was divested last year into a standalone brand.

It’s all a bit political. But Qualcomm only stands to gain from a change. Otherwise, it’s back to normal times, in which Qualcomm was doing quite well anyway. As 5G becomes more mainstream, Qualcomm is the most prolific chipset maker after Huawei. Does that make Qualcomm better?

In the 2019 Analyst Day Presentation, Qualcomm’s management spent a lot of time bragging about their end-to-end capabilities. We were interested in one particular question about competitive moat. When asked what Qualcomm’s competitive moat was, CEO-elect Cristiano Amon responded by saying (we’re paraphrasing): the end-to-end comprehensive chipset offering makes life a lot easier for the cellphone manufacturer because it makes testing that much easier. It’s easier to test different components of Qualcomm rather than a smorgasbord of components from various manufacturers. In 5G, the smorgasbord approach (presumably to save a few $$$) may be dangerous because it’s a big technological shift. We believe that the transition from 4G to 5G is a much bigger leap than was the case in previous generations, which would bolster Cristiano Amon’s argument. However, while the argument is believable, we can’t be sure.

There is one other peculiarity of 5G that may play to Qualcomm’s advantage – the 2 main bands of 5G airspace. They are Sub6 and mmWave. Sub6 basically refers to the bandwidth between 3.5GHz and 6GHz. This frequency band was not accessible by 4G technology. But there is another band on top of Sub6 that can go all the way up to 40GHz. This is an ultra-high frequency band that only accommodates a certain kind of (very short) frequency called mmWave. Short frequency waves don’t travel too far. Here’s a helpful stack of wavelengths depicted by Ericsson (also one of our holdings).

Anyway, the point of explaining this is that 5G phones will need to be very adaptive phones – different types of data and workloads will require different frequencies. And a 5G chip inside a phone will need to be able to handle it all. Very few companies will be able to master this – especially mmWave. Qualcomm is on top of a very shortlist. At the moment, Qualcomm dominates the 5G chip space.

So, Qualcomm was the next-best thing in terms of end-to-end offering, after Huawei. Right now, it seems to be the most prolific chipset maker – after Huawei and HiSilicon have relented. This is the time for Qualcomm to go for the jugular and impose its dominance. Take a look at this diagram below from Counterpoint Research, which compares Qualcomm’s footprint inside a smartphone compared to others – this paints a picture of Qualcomm’s supposed competitive advantage.

For this competitive advantage to be durable, Qualcomm would need to outdo its other rivals in terms of processing power and speed. So far, judging by its market share in high-end phones, it seems to be the case. But it’s still mostly a 4G world. In 5G, the bull-case is:

  1. Qualcomm’s end-to-end chipset is a big competitive advantage.
  2. In this end-to-end comprehensiveness, there is no serious rival after Huawei’s retrenchment.

Will Huawei/HiSilicon chips come back to the market soon? It’s possible. But Huawei, ironically, may be “too comprehensive” for US and EU comfort. In a 5G world, would the western world want a Chinese company with close ties to the Communist Party have end-to-end jurisdiction of all that sensitive data? Huawei will need to weigh cellphone sales on one hand vs. insisting on an end-to-end presence on the other. This might be Qualcomm’s big chance.

We realize that the prospect of Qualcomm striking while the iron is hot is somewhat speculative. We view it as positive optionality. In the sections below, you’ll see that we’ve left enough room for skepticism in our revenue estimates. We don’t expect Qualcomm to rule the 5G world. But the stars have aligned for them to make a serious revenue push.


Mngt. Strategy & Investments: Now focused on adjacent markets – IoT, AVs, Cloud, PCs. 

A bulk of the 5G investments have already been made. Qualcomm has spent about $17 billion in the last 3-4 years to be ready for the 5G era. The investments have paid off judging by their market share in 5G chips. Cellular modems will always be Qualcomm’s core competency. But 5G isn’t like previous generations when it was just about cellphones. 5G has a much broader reach.

In Qualcomm’s 2019 Analyst Day, the management team made it quite clear that they will be pursuing “adjacent markets” with gusto. The most obvious adjacent market is Automobiles – as cars get more connected and “softwarized”, they become more like mobile datacenters. They need to keep talking to the Cloud or to other cars, especially as they get more Autonomous. Management bragged about the hundreds of design wins in ADAS (Advanced Driver Assist Systems). This is a believable growth vector. But it’s also Qualcomm’s smallest business segment, and probably the most crowded one. This segment’s growth trajectory is the hardest to quantify.

There are 2 other growth vectors that look even more credible to us:

  1. IoT
  2. PCs

We’re bullish on Industry 4.0 – because we believe that there is strong business case for manufacturing firms to digitize their factories as much as possible. This requires the kind of chips that Qualcomm makes. Qualcomm’s management believes that the SAM (Serviceable Available Market) will be worth $21 billion by 2022, growing at a CAGR of 7%. We suspect the growth rate will be higher. In fact, in order to buy the stock today (March 26th, 2021), one of the assumptions we’d need to believe is a 10% CAGR in their IoT segment. More on that later.

If you ask us, “what is the market underpricing about Qualcomm”, we’d say it’s the PC business. Now, this is a speculative bet, but we think that as computing becomes more heterogenous (partly in the Cloud, partly in an Edge device), Intel’s x86 dominance will wane. ARM-based RISC chips – such as the one Qualcomm makes – will be more in demand because they’re less power-hungry. A few months ago, Apple took a giant leap by furnishing their new Macbooks with their own custom-made, ARM-based chips. For Qualcomm, Apple has opened the floodgates. In fact, Qualcomm already has its designs in PCs – Microsoft’s Surface Pro X comes furnished with a Qualcomm-designed SQ1 chip. Why is this not the future?

Incidentally, we used to hold Intel. But we got rid of our position because we think that the ARM-based chips like the SQ1 will be the norm in PCs in 3-5 years. If that happens, Qualcomm is in a good position to be a major player in this new market of power-efficient CPUs. We believe that the lines between Mobile and PCs will blur. At some point, all PCs will be connected to the cellular network like a Mobile Phone. Qualcomm, more than any other company we know, stands to gain from that. But it’s hard to quantify this growth vector, which is why haven’t factored it in our estimate of Sustainable Revenue. It’s all gravy on top if it works out, a.k.a Positive Optionality.

The other growth vector Qualcomm’s management team is excited about is Cloud Datacenters. Here, we’re much more skeptical. Our first reaction is that Qualcomm is a day late and a dollar short in this space. They’re going to compete with entrenched players like Intel, AMD, Nvidia, Xilinx etc., apart from the constant threat of the Big 3 Cloud companies – Amazon, Microsoft and Google – designing their own chips. So, we haven’t factored in any growth from this space into our revenue estimate either.

So, what have we factored in?


Growth Drivers: Content increase in 5G chips. Adjacent markets – IoT and AVs.

Handset Volume Growth: Right off the bat, we can tell you that we have not factored in any growth in the total volume of handsets. It’s safe to assume that almost anyone who can afford a smartphone in the world has one. According to equipment maker Ericsson, there will be an increase of about 800 million handset units by 2026 – mostly attributable to 5G phones. But we won’t factor this growth into our estimates of Sustainable Revenue.

However, we will factor in some volume increment based on our assumption that Qualcomm will capture some of HiSilicon’s market share. We discussed this in the previous sections. More on this below.

Core Chipset Pricing: We don’t have definitive data on pricing – they are (we suspect) negotiated in custom contracts between Qualcomm and each of its customers. However, we have context to believe that some price increase relative to 4G chips is warranted. In the November 2019 Analyst Day, Qualcomm’s Management claimed that they see a 50% increment in “content per chip” in 5G compared to 4G. We think this is a reasonable assumption. Will that translate to a 50% increase in pricing per chip? Maybe not. However, we believe that, in general, 5G chips will be pricier. We’ll enumerate our assumption on this below. But first, here’s some context for that 50% increase in content per chip or, put simply, processing power per chip that’s needed in 5G phones.

The qualitative case for all this 5G hype is hinged upon what a 5G phone will be able to do. But we’ve long believed that the real upside in 5G is in the business world, not so much in the consumer world. The real promise of 5G is in heterogenous computing – the idea that computing will be optimized between the Cloud and the Edge. This is not guaranteed because engineers still need to find solutions to the problem of mmWaves – that it can’t travel long distances or through any obstructions. But we tend to bet on engineers finding solutions when there is a multi-billion-dollar market at stake. Let’s assume that mmWave is feasible. And that 5G chips that support mmWave will be pricier than 4G chips. We’ll put some numbers around this below.

Adjacent Markets: Qualcomm’s Management Team wants to take a bite out of Heterogenous Computing. They recently acquired Nuvia – a company that designs CPUs for datacenters. On paper, this acquisition makes sense. Qualcomm is dominant in mobile chips. Nuvia claims to be “reimagining” CPU designs for datacenters. Together, they could make Cloud Datacenter CPUs that are optimized for Heterogenous Computing. To us, Qualcomm’s foray into Cloud Datacenters is unquantifiable – but it is massive positive optionality. We wish we could, but we haven’t factored in any revenue growth from this adjacent market.

The most believable adjacent market is IoT. Qualcomm already has a big business in IoT. In Fiscal Year 2020 (ending September 30th), IoT accounted for more than 18% of Qualcomm’s chipset revenue. We expect this number to grow. IoT is a natural extension of Qualcomm’s cellular modem legacy and competency. We’re generally bullish on Industrial IoT as a growth vector over the next decade, so we’re willing to factor in some growth from this segment.

As for the other adjacent market – Autonomous Vehicles – this is a nascent industry. While we’re bullish on this “near-inevitability”, there is no way to predict a growth rate for this market, but we have a work-around. See below.

These are Qualcomm’s business segments (the first 4 are grouped under “QCT” or Qualcomm Chipset Technology):

  1. Handsets
  2. RF Front end
  3. Automotive
  4. IoT
  5. Qualcomm Technology Licensing (QTL)

We will assign our estimates of Annual Growth for the next 3 years according to these business segments. But we won’t just do some pie-in-the-sky thinking. We’ll go about this methodically. Here’s how we’ll do it:

  1. The starting point is twofold: 
    1. Qualcomm’s revenue (and revenue growth) over the last 12 months ending December 31st, 2020.
    2. We know from our Watch List that we need to believe that Qualcomm’s revenue would grow by about 40% cumulatively for us to buy the stock. Is this believable?
  2. Then we take into account Management Guidance – CFO Palkhiwala gave us SAM (Serviceable Available Market) CAGR expectations over the next 3 years (shown below). This is a good base to start with.
  3. We’ll make some of our own assumptions about Qualcomm’s core (and more easily quantifiable) business.
  4. We’ll back-solve into revenue growth for the segment that’s least quantifiable – Automotive. Then we’ll ask ourselves if that’s believable.

Recall that we’re practitioners of Expectations Investing – meaning that we’d rather back-solve into “revenue growth we’d need to believe to buy the stock” than make obscure revenue growth projections based on the past data. In other words, we try and estimate what the market has underpriced or overpriced in the stock. In Qualcomm’s case, we’ll use a combination based on Management Guidance, our qualitative view (as delineated so far) and some back-solving.

So, here’s the starting point – how Qualcomm’s revenue has grown over time:

Then we consider Management Guidance – their estimates of SAM growth over the next 3 years:

Based on Management Guidance on SAM, we’ll tack on some incremental growth rates on top of SAM growth based on our qualitative assessment of Qualcomm’s core business as delineated in the sections above. Here are the assumptions and rationales:

We find this scenario very believable. The above set of assumptions is even more believable considering that we have factored in ZERO growth from:

  1. Qualcomm chips for PCs – like the SQ1.
  2. Qualcomm chips for Datacenters – we have no idea how they will compete in this sector. If they do compete, it’s gravy for us.

The main line of questioning toward Management in Qualcomm’s Investor Day in November 2019 was this: “why aren’t you more aggressive with your growth rate assumptions?” Management made the usual disclaimers about conservatism and such, which we think is the right way to approach this. Management did not reveal Revenue CAGR projections by segment. We suspect they’re similar to ours, since all they were willing to disclose was that “our revenue CAGR will likely outpace our SAM CAGR”. We agree, and we’ve done our best to quantify it.

The next 3 years will probably be much better for Qualcomm than the last 3 years. 5G’s moment is NOW. Qualcomm is most likely to dominate 5G chips. But it’s not a certainty because Qualcomm’s not the only sheriff in town.  

Negatives

Competition: Custom chips by large customers. And Mediatek.

We’ve discussed at length the Huawei/HiSilicon saga. So, we’ll take a look the other threats to Qualcomm’s growth trajectory. The biggest one right now is MediaTek – a Taiwanese mobile chipset company that’s popular among lower-end phones. As far as we can see, they’re the biggest threat.

Companies like Qorvo are a threat but they don’t have the end-to-end capabilities of Qualcomm. Again, this diagram is very believable. Does it translate to an economic Moat? We can’t be sure but judging by Qualcomm’s high market share in 5G phones, we’re inclined to say Yes.

At this point, we should address the elephant in the room – IP license issues. Qualcomm had a very public with Apple, soon after Apple started making its own Bionic chips for iPhones. Qualcomm claimed that Apple used Qualcomm IP. Apple fought back. But the whole thing was settled with Apple agreeing to pay Qualcomm huge license fees. We don’t know the exact amount – we haven’t factored in any in our revenue growth estimates. It is also interesting to note that Intel exited the 5G chip market after this Apple-Qualcomm settlement.

Overall, we don’t think competition will be any worse than it is today.  If anything, the Huawei/HiSilicon episode gives Qualcomm some time to go for the jugular and increase its lead in 5G phones.


Key Risk: China Exposure. Geopolitical tensions can hurt revenue.

Huawei had to relent, and it now open to using Qualcomm chips in its phones because of the US embargo on HiSilicon chips. Huawei has no choice. What can it do? Stop selling its phones? That’s its bread and butter. It’s unrealistic for us to not expect some sort of retaliation. This is a bit scary – take a look at Qualcomm’s China revenue:

Most 5G handsets in the world will probably be sold by Chinese companies like Huawei, Xiaomi, Oppo and Vivo. If China suddenly says that no American chips can be used, and only Mediatek chips will be used in Chinese phones, this would be a problem for Qualcomm. Now, the US can retaliate by saying that no US company can sell any components to Chinese phone manufacturers. This would create problems for China in, say, semiconductor equipment or software. What if Google is prohibited from selling Android OS to Chinese OEMs? This escalation can go on and on, and ultimately, it’s bad for everyone.

We were considering allocating 5% of our portfolio to Qualcomm. But considering this unpredictable key risk, we curb our enthusiasm and stick with 3%.


Strategy Risk: Scatter-brain - possibly too late to succeed in Cloud and AV markets.

The doomsday scenario in terms of strategy is investing too much time and money into ancillary markets (which it may not be able to succeed in) at the cost of its core business – 5G chipsets. In the November 2019 Analyst Day, CFO Palkhiwalla said that Qualcomm needs just an incremental $200 million R&D expenditure (on a base of about $6 billion annually) to gain market share in ancillary markets like IoT, AVs and Cloud Datacenters. If that’s true, it’s not a big risk.

We do believe that it will be tough to crack the Cloud Datacenter market. There are too many entrenched players who dominate what they do. Qualcomm + Nuvia may have something to offer. But we don’t know what that is. As such, we haven’t factored any growth from this market in our revenue estimates.

Our assumption about growth in AVs is our most aggressive one. If you recall, this is the variable we back-solved. A growth rate of 19% hinges on the assumption that the next 3 years in “softwarizing and connecting” cars will far outpace the trend we saw over the last 3 years. Given that Qualcomm’s AV business grew by 44% in 2020, we believe that 19% CAGR over the next 3 years (about 70% cumulative) is not an unreasonable assumption. But we are aware that the market has some heavy hitters like NXP (which we used to hold before). Incidentally, Qualcomm tried to acquire NXP a few years ago. The deal fell through and NXP has since made great strides in AVs. But NXP’s products are different – they sell Microcontrollers, which are not the same as Mobile CPUs. Qualcomm’s product is much more complex.

The main competitors in CPUs would be Intel and Nvidia. Intel acquired an Israeli company called MobilEye, which is focused purely on Autonomous Driving. Nvidia – because of its proficiency in GPUs used for AI applications – is also an ADAS favorite (and our most profitable holding ever). Qualcomm’s focus, however, may be more in Telematics – that would be a natural extension of their core competency.

Overall, if new CEO Cristiano Amon (a veteran telecom engineer) doesn’t overstretch and take on losing battles at the cutting edge of AI, we think Qualcomm will bolster its place atop the 5G chipset world. We would like to see Amon and team focused on their core competency.

Final Thesis Summary

Qualcomm chips are at the core of the global 4G to 5G transition. To buy its stock, we need to believe that Qualcomm will 1) maintain its dominant market share in 5G chipsets and 2) will make more inroads into ancillary growth markets like AVs and IoT. Both scenarios are believable. We would allocate 5% of our portfolio to Qualcomm, but considering the China Exposure risk, we'll settle for 3%.


We use cookies on this site to ensure the best service possible.