Qualcomm is a major player in the tech and telecom industry. From being the pioneer of CDMA to being one of the leading modem manufacturers on earth when it comes to LTE, Qualcomm is perhaps one of the best-known names in the processor and modem business. But the company has been having mixed fortunes of late – 2015 was a bad year thanks to the Snapdragon 810 fiasco, and while 2016 saw a revival of sorts, 2017 has been marked by anti-trust lawsuits in South Korea, Taiwan, and the US.
Along with these anti-trust lawsuits, Qualcomm is also battling Apple, which has stopped making royalty payments to Qualcomm. Apple feels Qualcomm’s royalty fees for patents and the conditions that tag along are too onerous and need to be revised, although Qualcomm (obviously) begs to differ. Recent news reports suggest that Apple may be looking to drop Qualcomm completely as its modem supplier from 2018 onwards.
Is Apple going the in-house way?
If Apple drops Qualcomm as its modem supplier, two things are likely to happen. Either Intel will replace Qualcomm as the modem supplier to Apple or Apple will develop its own in-house modem to be used in the iPhones, iPads and even the Apple Watch.
Apple’s silicon prowess has been on the rise ever since the company released the first 64 bit ARM processor, i.e., the A7. Apple has the best single core performance amongst all ARM processors and this year released its own in-house GPU, in the Apple A11. Two of the most important components of a SoC, i.e., the CPU and the GPU are now designed by Apple in-house.
It is very likely therefore that Apple would want to design its own modem and take control of all the important components of the SoC. The SoC plays an important role in determining the quality of Apple’s products, and Apple would want to develop as much of the SoC in-house as possible. It already designs its own CPU and GPU, and the modem is quite likely the next target.
Qualcomm depends on Apple…
Qualcomm has a 25-75 business model of sorts. Even though patent licensing makes up just 25 percent or so of Qualcomm’s revenues, it makes roughly 75 percent of its profits from patent licensing. Similarly, even though modems and processors make up close to 75 percent of Qualcomm’s revenues, they contribute just 25 percent of Qualcomm’s profits. It is important to note that the 25-75 split is a rough estimate, as the figures keep changing every quarter, but remain in the same ballpark.
Of all the smartphone manufacturers that Qualcomm deals with, almost all Android smartphone manufacturers purchase the Snapdragon line of chipsets which comes with Qualcomm designing all the components of the SoC, end to end, be it the CPU, GPU, modem or DSP. Apple however just purchases modems from Qualcomm and integrates the modem in their A-series of SoCs.
Despite purchasing only modems from Qualcomm, Apple is quite likely Qualcomm’s biggest customer. As I explained before, Qualcomm gets the vast majority of profits from its licensing division. Qualcomm’s licensing practice involves taking a percentage of the total device’s cost as licensing fee rather than a fixed fee. Apple is the second largest smartphone manufacturer on earth and commands the largest ASP amongst all smartphone manufacturers. By virtue of having the largest ASP, Apple is almost quite certainly Qualcomm’s largest and most profitable customer.
…and industry depends on Qualcomm
The Snapdragon 810 fiasco showed how reliant the smartphone industry is on Qualcomm. The Snapdragon 810 was supposed to be Qualcomm’s flagship SoC for 2015, and Qualcomm felt the urge to go 64 bit as Apple’s A7 was a 64 bit SoC. Evidently, this haste to get on board the 64-bit train led to Qualcomm making some poor design decisions as a result of which the Snapdragon 810 ended up being a very sub-standard SoC.
The 810 would heat up very easily under sustained periods of high workload and the heating would in turn kick in thermal throttling which meant the Snapdragon 810 would be able to maintain its peak performance for a very short period of time and heat up while being at it. The A7, by comparison, did not heat up like the Snapdragon 810 and was, therefore, able to maintain its peak performance for a much longer duration.
The Snapdragon 810’s poor performance meant that Samsung abandoned the Snapdragon SoC for its Galaxy S6 and went all in with its in-house Exynos SoC even as Apple with the A8 had one of the finest SoCs powering its smartphone. Pretty much everyone else in the smartphone market was stuck with Qualcomm which meant their smartphones were immediately at a disadvantage compared to the Galaxy S6 and iPhone 6/6Plus. And this affected the fortunes of a number of players, most notably HTC and LG, for whom 2015 was a crucial year, as both were being sidelined in the smartphone industry. Both smartphone manufacturers have not really been able to recover lost ground since.
The modem sole provider…
Qualcomm makes some of the best LTE modems in the market and has a near monopoly in the CDMA market. Intel has recently entered the modem market thanks to the support from Apple, but if Apple decides to go in-house for developing modems, then Intel’s foray into mobile-based modems would be a short affair considering no one apart from Apple sources modems from Intel.
There do exist companies like Mediatek c that integrate their modems along with their SoCs. However, the perceived quality of Mediatek modems is nowhere near what Qualcomm brings to the table. If Apple decides to go in-house with its modem and drag Qualcomm into expensive litigation for years, then that could have a huge effect on Qualcomm’s finances. A major reason why Qualcomm’s modems are considered best in class is that the company use profits generated from the licensing division to fund extensive R&D that goes into building world-class modems. If profits from the licensing division get affected because of litigation with Apple (which is almost certainly bound to happen), then the money spent on Qualcomm’s R&D is bound to go down. With reduced spend on R&D, Qualcomm would find it difficult to maintain the quality of its modems.
…which is why Apple pulling out could be bad news
Apple has the financial muscle and a business case to spend heavily on an in-house modem, considering the number of portable devices the company sells every year and the ASP every device commands. The same can also be said about Samsung and Huawei that develop their own SoCs and modems namely the Shannon modem in case of Samsung’s Exynos and Balong modem in case of Huawei’s Kirin.
So, if Qualcomm’s modem expertise were to come in jeopardy in future, the bigger players like Apple, Samsung and Huawei would be able to rely on their in-house units for sourcing modems. But other manufacturers that rely on Qualcomm might have to release products with inferior Qualcomm modems compared to the competition, putting them at a disadvantage. Qualcomm itself has witnessed a mild fall in quality because of profitability reasons. Earlier for its 32-bit SoCs, Qualcomm would develop fully custom cores that could any day compete with Apple’s offerings. However, in case of 64-bit, Qualcomm no longer develops a fully custom CPU core but a pseudo-custom CPU core that more or less retains ARM’s core architecture with a few modifications to suit Qualcomm’s needs. While this pseudo custom core is not bad, Apple’s core architecture that is truly custom made and tailored to Apple’s needs is far better.
That a company as big as Qualcomm can come under jeopardy because of Apple might seem a little far-fetched, but Apple’s stronghold of the smartphone segment is such that any company that does not fall in line with Apple’s demands almost always ends up dying. The case of Imagination is a perfect example of this, Imagination was the sole supplier for Apple’s GPU needs, and then Apple decided to make GPUs in-house which led to such a drastic fall in Imagination’s fortunes that the company ended up selling itself to private equity investors.
It could happen again. And that would be a hammer blow to many players in the market.