The Atlantic ran a story yesterday on the significance of Apple selecting AT&T for their iPad launch. The piece comments inspired me to pen this post and share a little secret... I love semiconductors - not the technology itself, but I get excited about lowest common denominator technology in general. If semiconductors are doing well and innovating rapidly, everyone else in the downstream ecosystem should do as well. I like the semi space for this reason; I also like building devices - another reason I'm drawn to them.
Okay Spinney..., but what the hell does this have to do with Apple staying with AT&T to launch the iPad, particularly when users complain about network performance, dropped calls, and 3G coverage? The answer, I think, can be found by examining lowest common denominator effects.
Everyone knows AT&T is one of two tier-I GSM carriers in the US along with T-Mobile - the smallest in the tier-I camp, and as such not a reasonable choice for Apple. Sprint and Verizon are CDMA carriers. Not everyone knows Qualcomm invented CDMA, owns most patents for it, and licenses these patents along with semiconductor sales for tons of dough to equipment vendors and handset manufacturers who supply Sprint and Verizon with their network gear and devices. If you're a mobile device supplier and want to launch a CDMA handset on a CDMA network, you're paying Qualcomm something - either for IP usage or for their chipsets. I can't imagine this sits well with Apple (it certainly doesn't with others like Nokia), but that's only perhaps one reason Apple launched iPad on AT&T instead of Verizon (Sprint, I suspect, was out of consideration since they too are CDMA plus connect the competitive Kindle e-reader).
The main reasons I suspect Apple continued a relationship of convenience with AT&T at this time was the A4 and Apple's vertical approach to product development, which is directly at competitive odds with Qualcomm's full system-on-a-chip Snapdragon offering currently being adopted by Android, Chrome OS, BREW, WinMo, and Linux. The situation may change when Verizon builds out a new 4G LTE network atop their 700 Mhz spectrum and when Apple will be free from Qualcomm IP licensing, but until then, AT&T looks like the long term winner of all wireless Apple products in the US because AT&T is a GSM network and at this time, that means freedom to control everything for Apple.
Back to the A4... Remember that little thing his Steveness mentioned last week?
The A4 is Apple's 1Ghz processor (they acquired the technology), similar to Qualcomm's Snapdragon 1Ghz processor. However, unlike Qualcomm's Snapdragon, the A4 doesn't include a 3G modem, GPS, or other full system-on-a-chip capabilities that most wireless semiconductor providers have built into 3G offerings. 3G support inside iPad is a separate module from the A4 processor. I suspect it's Infineon's offering since they rule the iPhone (on a side note, Infineon's GPS also supports the iPhone). Like Qualcomm did with their own BREW software, Jobs made it clear last week that Apple strives for super tight integration between their software and silicon to reach optimal performance, vertical scalability, and software development flexibility. Using Qualcomm-inside technology might have made it more difficult for Apple to achieve these goals while limiting their control. It's for this reason they chose to continue with AT&T!
Eventually, the same freedom-from-control that comes with vertical product development may be the reason Apple extends the A4 to build their own full system-on-a-chip that includes 3G and 4G modems, A-GPS, plus all the other features competing solutions have. I wouldn't be surprised if they did. Apple has proven controlling the whole stack and experience has competitive advantages while igniting explosive innovation with openness at the SDK and API level.