A lot of pieces go into a smartphone to make it work. From the display to the battery, there is an entire ecosystem of suppliers in China based in Shenzhen that is responsible for the Lego-like pieces that when put together in the factories of the likes of Foxconn and Pegatron result in the smartphones that we all know. During the early days of the smartphone market, there were a number of component manufacturers competing with each other. But as time passed by and the smartphone market kept becoming commoditized, the number of smartphone makers and subsequently the number of component manufacturers has been dwindling.
The absence of alternatives
Take the Qualcomm Snapdragon 835 for example. It is supposed to be the flagship chipset this year for Android smartphone manufacturers. However, very few smartphone manufacturers have been able to get their hands on it. The LG G6 which is supposed to be LG’s flagship handset for 2017 has been launched with the Snapdragon 821 which is what the OnePlus 3T was launched with last year. Now, I understand that having the latest processor is not the only thing that makes a smartphone great, but when someone is paying the top dollar, they do expect their smartphone to have the latest of everything.
Why would someone want to pay USD 600-700 for a smartphone that has yesteryear’s processor while something like the Galaxy S8 will probably retail for a similar price and come with the latest processor? Similarly, if one can content oneself with a Snapdragon 821, then why wouldn’t he or she get a OnePlus 3T at half the price?
But why did LG use a Snapdragon 821 instead of Snapdragon 835, and not something like MediaTek’s latest or best? This is because of Qualcomm’s monopoly over the North American and CDMA market in general. It would be in Qualcomm’s interest to supply all its customers with the latest possible SoC as that would definitely fetch Qualcomm a better price than an older SoC. But again, Qualcomm’s hands are tied just like LG’s. Just like the SoC is at the heart of a smartphone, foundries are at the heart of SoCs. SoC manufacturers need foundries to manufacture. Most of these foundries are expensive to build and the lower the process node, the more expensive it becomes.
Since Qualcomm is a fabless chip designer, it needs to outsource the manufacturing part. Qualcomm has a monopoly over the CDMA market which forces smartphone manufacturers to use Qualcomm alone if they need to serve the North American market and that market is unavoidable. But just like Qualcomm’s monopoly over CDMA, there are certain semiconductor manufacturers that have a monopoly over lower process node fabs. In the current scenario, Samsung and TSMC are pretty much the only options for fabless semiconductor designers that want to fabricate their SoCs on a lower process node. Considering the sheer volume of Apple’s Ax series of processors, most of the capacity in these lower process node fabs is taken up by them. This leaves others like Qualcomm and Huawei with little capacity to share from.
Power to a few players?
Qualcomm wants to fabricate its Snapdragon 835 on Samsung’s 10nm process but the output of Samsung’s fabs seems to be low enough, so that Samsung has apparently made it clear that unless and until Samsung’s appetite of Snapdragon 835 is fulfilled for the Galaxy S8, no one else gets to touch it (in meaningful numbers, at least). This is apparently why LG could not launch the G6 with the Snapdragon 835 and rather had to settle with the Snapdragon 821.
The problem over here is multi-layer. Not only is Qualcomm the only option for many smartphone manufacturers, but even the foundries that Qualcomm, Apple, and others use are very limited. Also, Snapdragon 835 is not the first time that such a problem has arisen, it has happened in the past as well with the Snapdragon 810. The Snapdragon 810 was supposed to be Qualcomm’s flagship chipset for 2015 but heating and other problems made it one of the most disastrous chipsets of its time.
Barring Samsung, Huawei and Apple, almost every smartphone manufacturer was at a disadvantage during 2015. They either had to put in an underclocked Snapdragon 810 or settle with a Snapdragon 808. On the other hand, Samsung was able to make use of its Exynos line throughout all the SKUs, Apple had its Ax series and Huawei made use of its Kirin line up.
Despite its consistently falling position in the smartphone market, the Chinese manufacturer, Xiaomi, plans to hedge its bets against Qualcomm by investing in custom silicon. Xiaomi’s Mi Mix device proves that the company isn’t afraid to venture into unchartered territories and can deliver as well. Considering the amount of time and money it takes to develop competitive SoCs, Xiaomi’s Pine Core based Surge S1 might not be an instant hit but would provide the smartphone marker the same kind of leverage Huawei enjoys with its Kirin line up.
The latest battle has been between Apple and Qualcomm. Even though Apple makes its own line of Ax series of SoCs, it still has to rely on Qualcomm for the modem. Apple in conjunction with anti-trust regulators worldwide has sued Qualcomm claiming that its business practices are illegal and harm competition. What comes out of the lawsuit remains to be seen but for the sake of maintaining competition, Apple might have decided to compromise on quality a little bit.
The iPhone 7 and the iPhone 7 Plus were among the first iPhones from Apple to sport a modem from Intel as well. Ever since Apple had stopped using modems supplied by Infineon in iPhones, Qualcomm pretty much had a monopoly over the modems used. But considering that the relationship between Qualcomm and Apple has been strained over the years, Apple decided to start using Intel modems in the iPhone 7 and iPhone 7 Plus as a hedge against Qualcomm.
But Cellular Insights soon released a report detailing how the Intel modem fared relatively poorly as compared to the Qualcomm modem in areas of low signal strength. Now anyone who had bought an iPhone with an Intel modem was obviously not going to have the same performance as someone with a Qualcomm modem would have. But Apple had no choice but to start using Intel modems in at least some iPhone models as the sheer volume (tens of millions every year) meant that if one component manufacturer has a monopoly, then that could affect Apple’s bargaining power and profits tremendously.
In order to level the performance, Apple issued a software update to smartphones running on the Qualcomm modem to make their cellular performance match the level of the iPhones running Intel modems. It is safe to say that for a company like Apple which places so much importance on hardware perfection, this is definitely not the kind of experience it would want its customers to have.
Not having control over the silicon has been a long-term problem even for companies like Google. The entire stock Android/Nexus community was angry when Galaxy Nexus did not receive the updates it had been promised since the SoC that was being used in it, i.e., the OMAP processor was no longer supported by its manufacturer Texas Instruments.
The need for more players to “chip” in
Be it the Snapdragon 810, the Snapdragon 835, the Galaxy Nexus or Apple’s Qualcomm vs. Intel tug of war, one thing that is eminently becoming clear is that having control over the silicon that goes into the smartphones is now more important than ever. If one does not control the silicon, then one is at the mercy of a few players like Qualcomm, whose own performance has not been consistent all and/or who can then demand onerous terms for their products like modems where they have a monopoly.