Qualcomm Snapdragon Summit – The Gauntlet

Qualcomm throws down the gauntlet to both Apple and Intel.

  • Qualcomm is mounting a bold challenge to both Apple and Intel in PCs with the launch of the Snapdragon X Elite but no matter how good the chip is, everything will depend on how well it runs the Windows ecosystem.
  • Qualcomm held its Snapdragon Summit last week and given the importance of what it announced, the tech press surprisingly paid it very little attention.
  • This Summit was important for three reasons.
    • First, in-house processor cores: which are important because this is how Apple achieves the best-in-class performance and power consumption in its M-series of chips.
    • Qualcomm is returning to designing its own cores for the first time in many years and if Qualcomm’s claims are correct, then its first chip using its new core design is 13% faster than the best that Apple currently has (M2 Max) on a single thread.
    • Furthermore, at the same level of performance, Qualcomm is claiming that the Oryon consumes 30% less power than the M2 Max.
    • This is crucial because it means that great performance and very long battery life may finally make it to the Windows ecosystem after several years of being handed its hat by the Mac.
    • However, Apple will probably launch the M3 this week where it is almost certain to proclaim its return to supremacy but either way, Qualcomm is promising to close a gap that has plagued the Windows ecosystem for several years.
    • Qualcomm is also claiming that Oryon can offer pretty much the same performance as Intel’s fastest available processor (i9-13980HX) and do so at 70% less power consumption.
    • In the Windows ecosystem, this is where the real competition is and every PC maker in the market will be taking an interest in this new chip.
    • However, there is a big caveat which is software.
    • The Windows ecosystem runs on the x86 processor meaning that all of the software old and new will need to be ported to run on the Arm processor.
    • Anything that is not ported (which will be a big piece of the legacy) will need to be emulated which creates a performance and power overhead which can destroy the user experience.
    • Apple managed this transition brilliantly as hardly anybody noticed when software was being emulated and so this was never a concern.
    • By contrast, Microsoft has tried this before and it made such a mess of it that hardly anyone wanted to buy the devices.
    • Unless this transition is as seamless as Apple’s and users don’t even notice that they are not running on an x86 processor, the whole proposition will fall to pieces.
    • This will be down to Qualcomm and Microsoft working together which immediately makes it more difficult because Apple owns and controls everything immediately making the transition easier.
    • If Qualcomm and Microsoft pull this off, then Intel will be facing the biggest challenge in its history making Intel’s return to being the best chip maker in the world all the more crucial.
    • Second, new market: Qualcomm has had a rough year with smartphones, and the Elite X promises to open up a new market segment for the company.
    • This is not a small segment as although far fewer laptops are sold every year compared to smartphones, the processors sell at much higher prices.
    • Furthermore, as Apple only makes chips for Macs, there will be no Arm-based competitor in the Windows ecosystem for quite some time.
    • Hence, if the Elite X and its descendants fulfil their promise then there is a good chance that Qualcomm will begin eating away at Intel and AMD in the market for laptop processors.
    • This could be a meaningful revenue stream over the medium term and so far, none of this potential is being reflected in the estimates or valuation of Qualcomm’s shares.
    • Third, Smartphones: where Qualcomm launched the Snapdragon 8 Gen 3 but more interestingly, promised that the Oryon core would be coming to smartphones in 2024.
    • The Snapdragon 8 Gen 3 offers improvements of 30% CPU, 25% GPU and 98% NPU which is where Qualcomm focused most of its efforts this year.
    • This is because the best performing and most private models will be those that run on the device rather than in the cloud and here there have been real improvements.
    • Snapdragon 8 Gen 3 can now run a model of 10bn parameters at 20 tokens per second which is enough to run a decent voice assistant.
    • The 1bn model that Qualcomm demonstrated at Mobile World Congress took around 10 seconds to generate an image on Gen 2 but on Gen 3 this has improved to 0.6 seconds.
    • I suspect that handset makers will mostly use this performance for computational photography in commercial products as that is where most of the competition is at the moment but there is also potential for other types of generative AI.
  • Qualcomm stopped making its own cores when Apple surprised the industry by jumping to 64bit computing in 2013 and it needed something in a hurry which Arm had already developed.
  • Since then, it has purchased its processor core designs from Arm and this marks a return to doing its own cores after 10 years which is why this Summit is the most consequential for some time.
  • The intention is clearly to quickly roll this right the way through its entire product line up with smartphones coming in addition to PCs in 2024.
  • This is the source of the current dispute between Arm and Qualcomm, and it looks to me like this boils down to a dispute over how much money Qualcomm will pay Arm now that it will no longer buy its cores from Arm.
  • Hence, I suspect that this will settle somewhere in the middle of what each party wants, and I was surprised by how little the market cared about this risk when it came to Arm’s recent IPO.
  • The two winners here are Qualcomm and Arm (even though they are fighting) as I see potential (software dependent) for Qualcomm to make a large dent in the laptop market and it will bring Arm with it.
  • Despite this potential, the shares continue to be weak and trade at less than 13x 2023 PER as no one seems to care or does not want to hear about a proper challenge to the M-series processor dominance.
  • I have a small position in Qualcomm and am considering buying some more as the growth story is becoming clearer but the shares keep on going down.

RICHARD WINDSOR

Richard is founder, owner of research company, Radio Free Mobile. He has 16 years of experience working in sell side equity research. During his 11 year tenure at Nomura Securities, he focused on the equity coverage of the Global Technology sector.