Connectivity: Bluetooth, WiFi (& GPS)

The current craze in the wireless industry is integration and the combination of many standards in a single chip, if not the baseband itself. So it's refreshing to see Apple reminding everyone that what actually matters outside the ultra-low-end is the quality of your product and how it affects the end-user experience, not how many cents you save on the final bill of materials.

To back that up, in a conference call with the analyst community in early May, a CSR executive made the following comment: "Bluecore6 is in mass production right now, we essentially reached mass production status right at the tail end of Q1. […] and we have produced several million units. […] I can't go into the details, but I can tell you that Bluecore6 will be used in some of the most hotly anticipated products in the mobile phone sector in 2008".

This is the wireless chip we referred to in the introduction. This comment was made at their Analyst Day while commenting the ‘Bluecore6: The best in the world’ slide. To fully understand why this Bluetooth design win nearly certainly refers to the 3G iPhone, it is important to look into the broader context of CSR’s forecast for Bluecore6. Mind you, this 'leak' is hardly CSR's fault and we hope they will not be punished for it by Apple; there's only so little you can say about a significant revenue source in your quarterly financial guidance.

In late February 2008, CSR pointed out that ‘BlueCore6 will start contributing to revenue in H2 2008 having started mass production following a design win at a top tier customer.’ – then, in on the 1st of May, they said: ‘BlueCore6 had an important design win at a top tier customer and started contributing to revenue in Q2 2008.’

We are therefore talking about a single design win generating, according to CSR at their analyst day, substantial revenue. So it’s a high volume product, which is also hotly anticipated and that nobody, not even some of the component suppliers, truly knew exactly when it would ship as recently as late February. Sounds familiar? Sure does to us. There are other indications too, but we’ll spare you all the redundant details. It could refer to another Apple product, of course, but we don’t really see Bluetooth being added to any of them in the short-term.

So what’s so magical about Bluecore6’s power that it‘s part of what warrants the first paragraph above? Not much, really: it's just downright better according to every single indicator. CSR claims power consumption to be 20% lower than their best competitor, and that seems to be accurate; Broadcom loves telling everyone about how great their process advantage is, but in reality that’s little more than marketing FUD. The radio part of the chip which takes most of the power doesn’t scale with Moore’s Law unless you go down the digital RF route (which Broadcom doesn’t), so it’s really all about design intelligence and not just brute force. While we’re at it, it’s worth pointing out that it seems Broadcom's solution was superior overall to CSR's Bluecore4/5, so it’s definitely not always so clear-cut.

Moving on, there’s also an interesting rumor involving Broadcom in the 3G iPhone: it has been claimed that they would be responsible for the GPS part of the chip. Initially, we were highly skeptical: Broadcom has little penetration in that market today, so why would Apple go to them instead of a larger player if the 3G iPhone does support GPS (which is still a big if, mind you)? But there’s actually a very good reason: their power consumption numbers for tracking mode are by far the best we could find (by about two times).

Given Apple’s clear emphasis on power consumption rather than simply cost, this would make good sense and is fully plausible. Other manufacturers which did not quote power numbers anywhere often seemed to focus more on cost in their press releases, which makes pretty clear what their focus is. So is it actually the case that the 3G iPhone has GPS and it is provided by Broadcom? We don’t know, but wouldn’t be surprised at all. At these power and board area levels, it’s probably hard to resist.

And then finally there’s WiFi. If we had to focus on minimising power consumption, we’d immediately think of Atheros' disruptive solution (see: en-genius’s excellent analysis). However, all indications (from their latest financial conference call) are that it won’t start contributing to revenue before 2H08, which is obviously too late for an 3G iPhone design win. There’s a slight possibility for the iPod Touch 2G, but even that seems unlikely to us.

We believe that Marvell, which had the design slot in the original iPhone, remains the most likely possibility; however, there is a chance that CSR or Broadcom might grab this one. There are compelling cases for both, but honestly the impact of such a choice to the end-user would be effectively imperceptible and it’s too speculative for any sane investor to care, so we won’t be examining those aspects here.

In conclusion, it is clear that shorter lead times are an advantage in more than just application processors: this additional flexibility seems to have made the 3G iPhone the first product on the market with what has been, according CSR, described by every phone manufacturer as the best Bluetooth solution available today (whether that’s completely true and every manufacturer does think is anyone's guess, but you’ll have to admit that it is a nice quote in this context).

Furthermore, Apple is proving that they aren’t going away from the low-integration route of the 2G iPhone just yet: they likely believe that being able to pick-and-choose components can deliver better performance and lower power consumption. And obviously they’re right; integration makes a lot of sense in quite a few cases, especially to reduce costs (assuming the process technologies aren’t too different) or lock-in customers, but it’s not a silver bullet and the craze about it does tend to get a bit boring after a while.