Qualcomm (QCOM 1.70%) is the world's leading smartphone-processor and cellular-modem vendor. Over the last several years, the company has dramatically increased its research and development spending, particularly to bolster its system-on-chip development capabilities. These investments have paid off in spades. Just about every high-end smartphone is either powered by Snapdragon or -- in the case of Cupertino's prized phone -- Gobi modem. A roadmap detailing Qualcomm's future chip plans just leaked, but the big question is whether it's real or not.
A questionable roadmap
According to MyDrivers.com, this is Qualcomm's latest and greatest roadmap:
(Source: MyDrivers.com)
According to the roadmap, we see things mostly as expected. The revised version of Snapdragon 800 launches late in the fourth quarter or early in the first quarter. The Snapdragon 805 hits the shelves in the late second quarter or early third quarter, along with revised modems. Then a next-generation 20-nanometer, 64-bit Snapdragon hits the market in late Q4 2014/early Q1 2015. This seems legitimate, but there is something here that actually doesn't make sense and seriously damages this roadmap's predictive validity. What is it?
Take a look at the "9635M" discrete cellular baseband platform that is scheduled for a mid/late Q3 2014 launch. First off, Qualcomm recently announced the 9x35 modem and gave the following information:
- Support for Category 6 (300Mbps speed)
- Built on Taiwan Semiconductor's (TSM 4.22%) 20-nanometer process
- Sampling in the first half of 2014, shipping in the second half of 2014
This roadmap suggests a few things that contradict Qualcomm's official announcements. First, it claims that this new modem will be built on TSMC's 28-nanometer process, which is clearly not the case. Second, it looks as though this roadmap is claiming Category 7 support. Now, the interesting thing about Category 7 is that it's still the same near 300 Mbps downlink speed, but the uplink speeds are doubled from 51.0 Mbps to 102.0 Mbps. This means that while data can be downloaded at the same rate, actually uploading data becomes much faster.
However, is this true? Given that Qualcomm would not have a whole lot of time to change the silicon mid-flight like this -- especially given that one of the more difficult parts of developing a modem is getting it certified on a carrier network -- a move from 20-nanometer to the older 28-nanometer process combined with a performance/feature enhancement seems implausible.
Still very competitive
There's absolutely no doubt about it -- Qualcomm's apps processor, cellular baseband, and integrated part roadmaps are absolutely compelling. The fiercest competitor here, Intel (INTC 4.77%) through its Infineon assets, is rolling out a Category-6 LTE-Advanced modem in the first half of 2014, built on a foundry 28-nanometer process. This will be the stiffest competition that Qualcomm has faced here in a while. But on a 20-nanometer node, Qualcomm will still likely have a die-size, performance, and power-consumption edge on the 28-nanometer Intel part.
The only risk to Qualcomm here isn't performance or features from a competitor, but the potential of a very aggressive Intel winning share at all costs -- i.e., a price war. It is looking increasingly likely that Intel won't be able to win the next-generation iPhone, even as a second-source supplier, so Qualcomm's biggest and most profitable discrete modem customer is likely safe. However, Qualcomm does have customers other than Apple, and those could be at risk. Also, with a credible LTE platform, Intel can sell its apps processor + modem to compete with Snapdragon platforms elsewhere.
Foolish bottom line
While the roadmap above likely has kernels of truth, there's a large-enough mistake on it to suggest that it may not be entirely accurate. Nevertheless, one doesn't need a leaked roadmap to prove that Qualcomm will continue to be exceptionally competitive -- i.e., the leader -- in mobile applications processors and modems for a long time to come.
