Earlier this month, NVIDIA (NVDA -1.81%) announced that it would exit the cellular baseband market by either winding down or selling Icera, the company it purchased back in 2011 to enter the market in the first place.
In Seeking Alpha, Mark Hibben wrote that to continue to develop competitive smartphone processors, Apple (AAPL -0.08%) needs to build its own cellular modem solutions. And -- I'm sure you can guess where this is going -- he suggested Apple could benefit from taking Icera off NVIDIA's hands.
This sounds nice and all, but I believe it is highly unlikely. Here's why.
Better modem technology was for sale and Apple didn't bite
Apple has passed up many opportunities to pick up companies that develop cellular basebands. For example, Broadcom (NASDAQ: BRCM) in 2013 picked up Renesas Mobile for $164 million; Apple surely could have had this asset if it wanted, but it didn't bite.
Then, when Broadcom announced it was exiting the cellular baseband business, it, too, said it would either sell the division or wind it down. Apple, again, could have had this business, but instead Broadcom ended up winding it down.
So now we have NVIDIA, which was arguably in a weaker position vis-à-vis cellular basebands than Broadcom, also readying to exit its modem operations. If Apple didn't want Renesas Mobile or Broadcom's cellular baseband business, then it doesn't make sense to expect it to pick up Icera.
Apple has its own baseband efforts
It has been pretty clear for some time that Apple is building its own cellular basebands. For example, I just hopped onto Apple's job page and typed in "modem." The first listing that popped up was for a "Modem DSP Firmware engineer" as part of Apple's "Wireless Software Engineering Team."
Here's the list of responsibilities the position is expected to fulfill:
This is yet another piece of evidence that points to a cellular modem effort within Apple.
There is no immediate need for Apple to release a baseband
One reason I think Apple likely feels it can build a modem effort organically rather than via acquisition is that it has time on its side. There's no pressing need for Apple to integrate a baseband into its applications processors for the iPhone anytime soon.
Stand-alone basebands from third parties have worked very well for Apple thus far. Additionally, the road maps from Qualcomm (QCOM -0.55%) and Intel (INTC 0.58%) look compelling enough that Apple can likely use stand-alone basebands to great effect for several generations of iPhone to come.
The only place where an integrated part might be really nice to have sooner rather than later is the Apple Watch. The Apple Watch S1 system-in-package is highly compact and densely packed with chips, as you can see in this image from ABI Research:
If Apple can integrate the baseband onto the same die as the applications processor, it could dramatically reduce the board space required to implement cellular functionality.
However, I think that when or if Apple places its own baseband inside one of its devices, it will be the product of an organic baseband development effort.