Qualcomm – Eye on Mobileye

Qualcomm is coming for Mobileye.

• Qualcomm has won its fight for Veoneer which will see it expand its offering in automated driving in an attempt to take on Mobileye in a market that is ripe for more competition.
• Qualcomm and SW Partners are acquiring Veoneer which will then be broken up and the autonomous driving piece sold to Qualcomm and the other sensor business will be found a new owner by SW partners.
• On the surface, this sounds like a trivial piece of news, but when one looks a bit deeper, one can see that this might be pretty important for Qualcomm in the long run.
• Autonomous driving is difficult to get right as the endless delays and now consolidation in the sector have demonstrated.
• This primarily springs from the nature of the driving environment which is a source of almost infinite permutations and combinations that are always changing.
• It is this characteristic that has made it extremely difficult for autonomous driving systems to perform to anything close to humans in difficult environments.
• The reason for this is almost entirely down to machine vision and not the logic that makes the driving decisions.
• Almost every crash or disengagement suffered by an autonomous system is because the system did not interpret its surroundings correctly not because it made a bad decision on good information it was given.
• This implies that the biggest bottleneck (there are others) to autonomous driving is the machine vision system by which the vehicle works out what is going on around it.
• All of the errors that one sees on Twitter or elsewhere of Teslas making horrible mistakes are due to visual interpretation errors and not incorrect decision making.
• This is why Mobileye is such an important company.
• Mobileye has an 80%+ market share of the machine vision market for vehicles because to date, its machine vision has been substantially better than anyone else’s.
• Furthermore, because it is primarily a camera-based system, it can be put into production vehicles at a tiny fraction of the cost that it would take to equip vehicles with lidar.
• Lidar is a great aid to machine vision, but because it is currently so expensive, no one is using it on production vehicles because consumers won’t pay for it.
• RFM’s research indicates that the machine vision market is ripe for competition and that a second supplier with a similar level of quality would be welcomed with open arms.
• Furthermore, the market also has good things to say about Qualcomm’s offering in this area and it is hoped that by the middle of next year that it will be able to compete on performance with Mobileye.
• This is what the acquisition of Veoneer is all about as it will round out Qualcomm’s offering in this area to include lidar, driver monitoring as well as software for machine vision.
• This combined with Veoneer’s relationships with the OEMs will help Qualcomm gain share against Mobileye over the next few years.
• The key to Qualcomm’s success will be how Mobileye reacts to this new competitive threat, but from a customer perspective, Qualcomm is pushing against an open door.
• Qualcomm has come back off its highs and is now trading at a relatively reasonable 14.5x 2022 PER.
• This is two points above Intel which is on 12.3x but Qualcomm’s short to medium growth outlook is far better and it does not carry the transformational risk that Intel does.
• Qualcomm’s growth forecasts are in line to better than many of the FAANG names meaning that it represents much more value than most of the rest of big US technology.
• However, we are in a big semiconductor upcycle driven by shortages and when that turns, Qualcomm will suffer with the rest of the sector.
• On balance, if I was in large-cap US tech (which I am not directly), this is one I would have a very close look at.