Joining the hunt for the 3G iPhone

There has already been a vast amount of speculation as to the timing of the 3G version of the iPhone. This is for good reason. A late year launch would mean drastic price cuts would be needed to keep sales going over the summer. The handset market moves much to fast to keep a year old device selling for $400 relevant to the majority of consumers. Conversely, a late spring launch would bring a jolt to iPhone sales that I think would go well beyond most analyst expectations. The customer satisfaction numbers for the iPhone are off the charts, but the fact that it is well known that a 3G version will launch in '08 has kept many consumers on the sidelines. Nobody wants to buy the current iPhone only to see a more advanced version released a month later.

So the question is, how do we narrow down the launch window for the iPhone? I think the best clues come from the suppliers. The baseband and applications processor for the iPhone are, by necessity, going to require the latest and greatest chipsets.

The Baseband

Infineon won the 2G iPhone slot with their ARM9 enabled S-GOLD2 EDGE baseband. This likely gives them the inside track for the 3G iPhone, though it is by no means a lock. Infineon has stated that their S-GOLD3 HSDPA processor will be in volume production in the 2nd quarter. They have been close mouthed about the identity of any customers for this chip.

Qualcomm is also a possibility though I'd place the odds considerably lower than for Infineon. One advantage to Apple for going with Qualcomm is that they have a much better application processor combined with their baseband. The MSM7200a is a 65nm chipset with a 528MHz ARM11 enabled apps processor. In the past, Qualcomm has indicated that this chip would be capable of running the iPhone. I have my doubts though that it could do so without another dedicated apps chip. However, coupling the MSM7200a with an apps processor would provide more than enough horsepower for a fully featured iPhone (5 megapixel camera, GPS, 30fps video capture). The MSM7200a sampled in the 1st quarter of 2007 and I believe that volume shipments would likely begin in the spring.

Broadcom is the last of the likely candidates. At one time, I had thought that the iPhone could be their break through win in the mobile space but I think that this has become unlikely. The press has focused their attention on the single-chip HSUPA chip that they announced in October, but this chip wont show up in handsets until 2009. Their ARM11 enabled HSDPA chip that was announced last February (BCM2153) is a possibility, but Broadcom's public comments on wireless have downplayed the prospects for wireless in 2008. It seems unlikely that this chip is going to get much traction.

Apps Processor

Many choices, but porting the iPhone OS to a new applications processor seems unlikely unless it involves the battery/cost savings of a converged chipset. That means that Samsung's 2G design win is going to give them a considerable advantage in the 3G slot. Their 2G chip was the S3C6400, which was a 667MHz ARM11 enabled processor on a 90nm process. The 65nm version of this chip (S3C6410) is due to begin sampling in May with volume production in the 3rd quarter. As they did with the S3C6400, it is entirely possible that Apple will get early access to a custom version of this chip.

My best guess is that Apple will go with a repeat of their 2G iPhone partners. Infineon and Samsung both look to be in position to have volume shipments of the required chipsets by the summer. While the exact timing would be nice to know, I dont think it is necessary to profit on Apple. The key is that they should be able to avoid the worst case scenarios of a November launch. That would have been disastrous for the summer/fall sales and would have left them launching directly opposite Nokia's expected S60 touchscreen device. A late spring or summer launch avoids this and sets Apple up to do very well in the back half of the year.