Last week, we took a look at the mobile semis market and raised some concerns we see for that market, especially for Qualcomm. That post encountered a fair amount of pushback in the comments and elsewhere online (but not from anyone actually at Qualcomm). Nonetheless, we wanted to try looking at Qualcomm from a different perspective. The company has a lot of smart people, and they know the mobile market better than anyone. So surely they must see what is happening and are taking rational actions in response. If you read Qualcomm’s recent investor materials and transcripts, two things stand out – Cars and IoT. I think we can all agree, that more and more things are going to get connected to the Internet, and so Qualcomm seems to be carving out a bigger space for it in those markets.
In their public comments, Qualcomm’s management spends a lot of time talking about autos, even though that segment contributed just a little over 5% of revenue over the past nine months. In some ways, this makes sense. As we pointed out last week, Qualcomm’s core handset market matured a long time ago, with slow growth, customer consolidation and clear pricing pressure. But autos are largely untouched. The global auto market is roughly 150 million units a year, with something like a billion cars on the market today. ABI estimates that about 30% of new vehicles sold last year had Internet (i.e. cellular connections). Qualcomm probably sells the majority of the modems that ‘connect’ those cars. Those chips are probably priced around $25 a piece (a mix of 4G and 5G, because auto product cycles are long). This squares roughly with Qualcomm’s Auto segments which is approaching a billion a year in revenue. Assuming Qualcomm can upsell their chips a bit (5G FTW!) and that connected penetration rate doubles, it is easy to see this becoming a $2 billion business for Qualcomm soon. This is great, but does not exactly replace the smartphone business.
So then we turn to the Internet of Things. IoT is a messy category, and means different things to different people. For Qualcomm, IoT consists of two things representing the present and the future. The present reality is that most of their IoT segment is the sale of Wi-Fi and Bluetooth chips to a range of consumer electronics, probably including connectivity chips sold into smartphones. The future of IoT is much broader. This is their slide on IoT below , it is lovely but just as hard to read as everyone else’s slide on IoT. There is a lot going on here, but it all really speaks to Qualcomm’s ambition to widen their addressable market to a whole new category of devices. As we said at the outset, it is safe to say that many more things will get connected to the Internet. Some of these things will be connected with low-priced Wi-Fi/Bluetooth, others will require higher margin cellular connectivity.
Reading Qualcomm’s slide is always a fun exercise in code-breaking. Note all the icons on the bottom highlighting all the things that Qualcomm can power in the IoT. The trick is determining which of these are discrete chips (more $ content) and which are just standard features in connectivity chips (less $ content). Voice UI, Security, Location, 2G/3G/4G/5G/6G, Bluetooth, Audio – those are just features. But computer vision, voice UI start to look more interesting, they imply that Qualcomm will not only sell basic connectivity but also some form of applications processor. And then there is CPU, GPU and NPU DS – those are intriguing as Qualcomm does not exactly sell those today (yes, Windows laptops, but…).
In both Autos and IoT, Qualcomm appears to be betting that not only can they sell connectivity but also extend into other core compute areas. This is important because those could be massive markets. Take auto as a simpler example. Everyone has a different estimate on this, but roughly speaking the average car has ~$600 in semi content today, mostly low margin/low intelligence devices like microcontrollers for rolling up the windows and connecting sensors. But everyone also expects that amount to grow substantially in coming years as consumers expect better electronics in their vehicles. And then we get to autonomous vehicles which will require A LOT more semis. It is way too soon to tell how much of this market Qualcomm can really claim, and if they even have a shot at competing in autonomy, but it definitely looks like they think they have reasonable prospects here. The same logic applies to IoT.
So is it possible that Qualcomm is betting its future on an auto market that is still many years away and an IoT market which is too fragmented for anyone to sanely analyze? We have a clue this week. After years of nursing their wounds from a near-death hostile bid from Broadcom and the never-ending NXP deal that finally ran aground on the rocky shoals of the US-China Trade War, Qualcomm has returned to the M&A market. Over the weekend, Qualcomm entered a topping offer to acquire autonomous vehicle technology provider Veoneer. For a decade Qualcomm has been largely dormant in the M&A market, always the lukewarm auction participant, and now they have leapfrogged into a competitive bidding process. The fact that it is deep in automotive and somewhat distant from the company’s core competencies, marks this as a big change. (NXP is a whole other post and definitely does not qualify here). Qualcomm is by no means finished in the mobile market (quite the opposite), but it is clear they have woken up to their need to diversify.
Photo Credits – The Simpsons