Qualcomm & Semis – Barmy army.

Qualcomm answers the M-series challenge.

• Qualcomm has worked out how to challenge Apple’s current advantage in Arm-based CPUs but the fact that it will be 2023 before it hits the market underlines just how far ahead Apple is.
• Qualcomm held its 2021 analyst day in which it stated that it sees its serviceable market growing by 12% for the next 3 years and that it will have a proper challenge to Apple’s M-series family in 2023.
• This combined with the company’s highly credible expansion outside of smartphones has given the market confidence that there is life after smartphones and that Qualcomm will be able to mitigate the impact of losing Apple as a customer (as it might in 2025 when the current deal expires).
• This renewed confidence has triggered a spectacular rally in the share price with the shares touching $181 up some 50% since mid-October.
• Qualcomm’s CTO, Dr. Jim Thompson, announced at the meeting that the company is on track to produce an M-series competitor that will sample next year and should be in devices in 2023.
• I suspect that a large part of this new product line is a result of the acquisition of Nuvia for which Qualcomm paid $1.4bn despite Nuvia having no customers, product, or revenue.
• My thesis on this acquisition remains unchanged which is as follows.
• Qualcomm has had difficulty in designing a decent processor for personal computers although it has had a little traction in servers.
• The performance of the Surface X was reasonable with good battery life but it did not really wow the crowd making it far from a game-changer.
• Apple releases the new Macs with M1 and blows the doors off with its claims which are almost universally found to be accurate in independent reviews.
• Seeing this and comparing this to what it achieved with Surface Pro X makes Qualcomm realise that something big is missing in its Arm designs which need to be urgently fixed.
• In early 2020 when Apple got wind of what its ex-employees (Nuvia) were up to, it realised that the technical tricks that it had created to make the M1 as good as it is were either not patentable or it had not filed patents that cover these claims (unlikely).
• This could happen if the performance uplift is more about design, the 5nm TSMC process, and how one implements the Arm IP rather than anything radically new.
• As a result, Apple tries to interfere with the development of Nuvia by filing a lawsuit against one of the founders in order to keep its new implementation techniques in-house for as long as possible.
• Qualcomm realises that it can replicate the performance and power characteristics of the M1 much more quickly by purchasing Nuvia which is how this start-up could be worth so much to Qualcomm.
• It is the fruits of this acquisition that I suspect will give Qualcomm the ability to produce a viable competitor to the M-series which is bad news for Apple and Intel but good news for everyone else and for consumers.
• However, the time frame for its availability demonstrates just how far Apple has stretched its lead giving it another couple of years before there is real competition for its chips.
• This also gives it time to keep improving to stay ahead of Qualcomm but the fact that we have two superb chip designers competing like crazy against each other is excellent news for consumers.
• The days of having to make real compromises on battery life and performance may be coming to an end and thanks to Qualcomm this will be made available to spread through the Android and Windows ecosystems and not be limited just to Apple.
• This again raises the question of whether the x86 processor is now obsolete but because there is no x86 processor available on the equivalent node to TSMCs 5nm, this question remains largely unanswered.
• A good portion of Apple’s improvements with its new chips may be coming from TSMC’s 5nm process node and unpatentable design tricks rather than from its own IP and if so, they will not remain exclusive to Apple.
• Should this prove to be the case (which I am leaning towards), then Intel should be able to catch up and re-establish the performance superiority of the x86 processor.
• This is the multi-billion dollar question facing Intel investors at the moment and given Qualcomm’s rally, the valuation argument for Qualcomm over Intel is not as strong as it was.
• I have previously had a strong preference for Qualcomm over Intel (see here) but Qualcomm has since rallied 50% and Intel has gone nowhere.
• Hence, the valuation argument now favours Intel, but I think it can easily go lower and so I am still comfortably sitting on the sidelines waiting for a better entry point.