It is interesting to see some of the challenges facing once dominant semiconductor companies as the evolution of computing continues. Intel is and has been challenged to maintain the relevance of their architecture as personal computing shifted from desktops and notebooks to mobile devices. Similarly, Qualcomm is facing challenges in the next phase of mobile. Both Qualcomm and Intel have many similarities. Both focus on innovation to drive next generation chipset growth. Beyond advancing the capabilities of computing devices, the goal is to use this focus on innovation to keep the ASPs of their chipsets from falling. However, as hardware is commoditizing, maintaining ASPs is becoming more difficult.
With Intel, the writing on the wall happened with the commoditization and dropping ASPs of PCs. On top of this trend, the PC industry is not growing. Hence, most of Intel’s revenue strength has come from servers, not necessarily client PCs. Intel hit their mark of 40m tablet shipments in 2014, largely driven by the white box Chinese tablet market. But, outside of servers, Intel is facing headwinds with declining ASPs.
For Intel, growth markets like smart phones seem essential for the company to grow share. However, even that isn’t as black and white as it seems. While the smart phone market is huge, with our estimates above 1.5 billion smart phones sold in 2015, the market is seeing ASP declines faster than any segment. Intel has some insulation in the PC market but would have no protection from being forced to compete in commodity hardware in the smart phone sector.
Commodity smart phone hardware is what Qualcomm is facing as well right now. With a range of factors, from China regulatory and licensee disputes, to the potential loss of Samsung as a customer for Snapdragon, are all causing Qualcomm to issue lower guidance for next quarter.
Qualcomm is facing the reality that the era of smart phone growth where they could get a premium for their chipsets is about to end. Companies like MediaTek are able to compete more on price with Qualcomm at the low end of the smart phone market — which is where nearly all the device growth is going to come from in the next few years.
Both Intel and Qualcomm have to become competitive in the lower-price tiers, going against the grain of how these companies are oriented, if they want to capitalize on the upside volume. There is certainly a case to be made they can continue to serve the premium segment of the market with next generation innovations. However, the problem with that reasoning is it is becoming clear Apple is the only major player able to sustain in premium. Apple only needs a discreet modem from Qualcomm and does not use a Snapdragon processor. Similarly, should Intel have a shot at winning Apple it would only be with the discreet modem and not the entire SoC. Both companies get more money for the whole SoC over the just the discreet modem.
Looking at Qualcomm’s financials it is clear Apple’s impact on Samsung is having an effect on them as well. In fact, Apple’s role in this situation is hard to ignore. Their clear domination of the premium segment is locking out potential customers in the space where both Intel and Qualcomm could hope to sell more premium parts. This makes a strategy that Intel and Qualcomm are using to maintain ASPs or sell premium innovations, thus monetizing their RND spending, a little more difficult. If all you are left with are customers who want to sell cheap things, then this model is challenged.
What is interesting about this from a Qualcomm and, to a degree, an Intel viewpoint going forward is how ARM inherently created this situation. Since ARM is a licensable technology, from an architectural standpoint, the ARM ecosystem enables a wide range of competitors to compete in the ecosystem. Essentially, Qualcomm’s latest processor one year is the baseline ARM offering the next. I’m generalizing, but this is essentially the case.
This is what has enabled MediaTek, and really any other new upstart, to compete with a premium ARM licensee. And more to the point, when the market wants low cost chips, the competition in the ARM ecosystem gets incredibly intense. This headwind is what Qualcomm and Intel are up against in this next phase of mobile.
There are a couple of puzzling things:
– How did Apple get such a lead in ARMv8 ? Were Qualcomm that NIH-infused they refused to do a quick and dirty ARM-based respin à la A8, and tried to come out with a v8 Krait do-over straight away until their hand was forced ?
– Wearables are probably not a huge martket right now, but is it wise to let… Mediatek… lead with the first smartwatch-optimized SoC ?
– Oh , and also, have Intel backtracked on their desire to fab for others ? I’d assume Apple and Qualcomm to be the main prospects, yet the silence is deafening.
They didn’t see the 64-bit from Apple coming. All ARM designs on v8 were being looked at for server not client. So it caught everyone off guard and gave them a must-year advantage.
Since Apple is an architectural licensee of ARM, as is Qualcomm and others, they had access to ARM v8 to do as they please. There is a question as to how close ARM was to this situation and if they did work with Apple very early on to do this. From when ARMv8 was official, Apple’s turnaround was so quick it was either masterful, or ARM was working with them in advance.
To respond, Qualcomm is offering the 64-bit chip but it is not on the Krait architecture. Those chips are still coming.
And Intel is still doing some work manufacturing others as a part of their custom unit. Not much volume and no big name wins.
One thing that occurred to me today. If Qualcomm was not doing 64-bit when the iPhone 5s struck, then what were they doing? Were they doing something else to improve CPU performance, which didn’t really work out, or did they assume that smartphone CPUs were good enough as they already were?
Once upon a time, Intel was caught off guard in the shift to 64-bit on PCs, precisely because they were incapable of imagining that 64-bit would be used on anything else but high-end servers, hence Itanium. At least though, they were advancing clock speeds on the Pentium in the meantime. Qualcomm and ARM’s focus on servers seems to be a similar mistake, but I’m not sure what they were doing meanwhile for smartphone chips. Maybe they were focusing on quad-core designs… Maybe they weren’t doing much at all.
Not being able to offer market leading performance puts Qualcomm in a position similar to AMD. Yes, they can stay in the game, but being deprived of the best customers and unable to demand premium pricing for SOCs does make life much much harder.
Intel is lucky that they have a profitable legacy x86 business, but that is no longer the centre of gravity in computing and not a source of growth. Their risk is that even if they plough a fortune into mobile CPU development, they would have a hard time monetising any advantage because Apple owns the bulk of the profits in handsets.
One question which might help provide perspective.
Does or will Apple’s A8 chip outsell Qualcomm’s Snapdragon?
Given that high-end Android (but no longer Samsung) is mostly Qualcomm whereas mid-range is being eaten up by companies like MediaTek, I think that this could be possible.
I don’t think it matters since Qualcomm and Apple aren’t competitors.
Direct competition is not the only angle you should be looking at. A single perspective leads to a narrow understanding.
“blind men and an elephant”
http://en.m.wikipedia.org/wiki/Blind_men_and_an_elephant
The ability of OEMs like Samsung, HTC and LG to sell premium Android phones depends on Qualcomm. All of their best phones use Snapdragon processors (even Xiaomi uses Qualcomm on their flagship phones).
Typically, modular companies like Qualcomm can provide better performance products at lower prices because of the economies-of-scale. Vertically integrated companies are at a disadvantage in this regard because they can only sell internally. However, if Apple’s chip units becomes larger than Qualcomm’s, this no longer applies. Qualcomm will no longer have the economies-of-scale on its side. Apple will gain the advantages of both vertical integration and modularity, which means that they will have the best chips at the lowest prices exclusively for their own internal use.
The extreme case is if Samsung stops using Snapdragons and the premium Android market shrinks (ceding to iOS). In this case, Qualcomm will no longer be able to sell large volumes of high-end processors and they will inevitably shift R&D resources towards mid- to low-end ones. Then no-one will supply the high-end chips for Android.
I am predicting an extreme bifurcation of the smartphone market. I want more perspective into the possibility.
So I don’t think a comparison can be made generically because Snapdragon represents a family of chips that span price ranges. We can ask is Qualcomm’s high end chipset, the 810 will outsell the A7 and its a valid question but one we can’t really track so easily. Qualcomm only gives overall SoC shipments and does not break them out by model and I have never seen any model doing so from a third party.
Thanks.