Apple (AAPL 1.93%) is well-known for designing its own applications processors for its iOS devices. The Apple A-series chips have evolved from being just another ARM implementation to being truly world-class designs in their own right.
However, despite the massive amount of chip expertise that the company has built up in-house over the years, it's still missing one piece of the puzzle that all of the merchant chip vendors Qualcomm (QCOM 2.19%), Intel, and many others have in-house. Heck, even Samsung has this particular capability in-house . The capability is the integration of a cellular baseband chip on the same piece of silicon as the processor.
Why integrate?
For today and for the foreseeable future, apps processors can't get enough transistors dedicated to graphics, processor performance, camera functionality, and so on. Smartphones have rapidly evolved into handheld, fully fledged computers. As a result, computing power is really the name of the game. Richer games, higher-definition movies, more sophisticated websites, facial recognition, and many more usage models that haven't even been dreamed up yet will all require more graphics and processing horsepower.
However, the name of the game is integration. Eventually it's going to make sense for just about every apps processor -- low-end or high-end -- to come packed with an integrated cellular baseband processor. As far as anybody outside of the company knows, Apple doesn't have its own cellular modem team in-house. Indeed, in its iPhones, the company uses discrete modems exclusively from Qualcomm. The real question now is what Apple plans to do going forward. Indeed, there are three options that make sense.
Option No. 1: Use discrete forevermore
Apple's iPhones are already sleek and sport excellent battery life despite packing a two chip solution. The obvious solution would be to simply use discrete cellular chips going forward, using Moore's law to pack in more CPU, more GPU, and more dedicated co-processors and functionality that are accretive to the user experience. For a vendor like Apple, one that plays almost exclusively at the high end, using a discrete solution makes more sense. It may drive up the bill of materials, but Apple can certainly afford that flexibility.
Option No. 2: Bring that expertise in-house
The next option would be to simply buy the expertise. Cellular modems are exceptionally difficult and are a patent minefield. Staffing up a design team from scratch by, say, poaching employees from other houses is probably not going to work. If Apple wants cellular expertise in-house, it'd have to outright buy it. Given that Apple has passed on two rather obvious takeover candidates in this space -- Renesas Mobile and the ST-Ericsson joint venture -- it's clear that Apple isn't quite ready to make a leap into this space. It's exceptionally difficult move.
Option No. 3: License the technology
The third and final option would be to simply license the relevant IP blocks from a vendor that's willing to license it out. Now, given how much "secret sauce" goes into a modem and how lucrative selling a discrete modem is, it's unclear just who would be interested in licensing the technology to Apple. Further, until CDMA network operators collectively make the switch to LTE-only networks -- which could take years -- Apple would need Qualcomm's modem IP in order to support much of its customer base. It's unlikely that Qualcomm would license its modem for pennies when it could sell the discrete modem for $20 or more.
Foolish bottom line
For now, it seems Apple will continue to use discrete modems in its iPhone product lines. These are premium phones, so Apple can quite clearly afford to use a separate chip. At some point, however, it will make sense to integrate a modem onto its apps processor. It will be very interesting to see what Apple has up its sleeve over the next several years here.
