Apple (AAPL 2.15%) is the world's largest, most successful consumer electronics company. The company single-handedly revolutionized the cell-phone industry, refined and redefined the "tablet" category, builds (arguably) the world's best PCs, and has one heck of a silicon development team responsible for its "A-series" of chips. The company develops world-class SoCs with leadership IP, much of it custom (and increasingly so).
However, there's one piece of IP that Apple doesn't appear to have today: a cellular modem/baseband. This is a curious omission for a company for whom 52% of sales comes from the sale of cell phones, and an even more curious omission given the strength of the rest of its silicon teams.
How about modems?
What's interesting is that Apple doesn't design its own modems -- in fact, it buys off-the-shelf discrete modems while the rest of the merchant chip industry (as well as the players that build their own chips, such as Samsung) continues to integrate cellular basebands into their mobile system-on-chip products. In short, Apple seems to be the only company that designs apps processors that does not also have modem capability.
Now, given that Apple has demonstrated that it can build world class SoCs and absolutely monstrous CPU cores, it only seems logical that Apple also has internal efforts dedicated to the development of its own modem IP. Why? It's simple: As long as Apple continues to use discrete modems, it will be at a material disadvantage against competitors like Samsung, HTC, and others from both a BoM and a power/board space perspective. It needs to integrate its own modems at some point down the line or risk simply needing to exit the chip-development business for its phones entirely.
Why Apple won't stop designing its own chips
Apple likely has two options: Get some in-house modem action or simply exit the chip business. It is unlikely at Apple intends to exit the chip business considering how aggressively it hires chip designers:
(Source: Apple.com)
As you can see, Apple has 60 job openings pertaining to CPU design alone -- let alone the rest of the SoC! So, clearly Apple is very committed and focused on developing its own applications processor for a while longer. This suggests to me, then, that Apple has no choice but to do its own modem.
Why is Apple hiring so many RF IC architects?
A look at the job board over at Apple shows that Apple is hiring plenty of RF IC engineers, too:
(Source: Apple.com)
Apple is apparently hiring folks to build both an RF front end and a RF transceiver. It wouldn't make sense for Apple to do RF front end/transceiver and to not go all of the way to build its own cellular baseband for integration onto the A-series SoCs.
Conclusion
It looks like Qualcomm's (QCOM 1.01%) days could be numbered inside of the iPhone and this makes it increasingly unlikely that Intel's wireless division (or Broadcom's, or anybody else's) has any chance of gaining any content share within the iPhone (if Apple does indeed develop and integrate a cellular baseband). Now, before you go ahead and say that Apple "can't" do it, understand the following:
- Apple has one of the best silicon design teams in the world
- Apple has a lot of money
- Every other mobile SoC vendor, even ones far weaker than Apple, have their own cellular baseband and RF effort
- Apple ships an absolutely enormous number of smartphones, more than justifying the need to stop handing over margins to Qualcomm and others
Apple ships about 150 million iPhones per year with Qualcomm's baseband content estimated at somewhere between $15-$20 per phone-implying a roughly $2.25-$3 billion business that Apple could cease giving to Qualcomm (and could be realized as a significant gross margin savings).
