The focus in the baseband market is now very much on LTE. With major carriers rolling out LTE networks in the US, Northern Europe, Japan, Korea and a few other markets, handset makers are scrambling to find suppliers of 4G modems. In 2012, there was really only one choice, Qualcomm.
In a development that we had predicted, but are still shocked to see, in 2013 there will really only be one choice of LTE vendor, Qualcomm.
Almost all of the baseband vendors have announced LTE roadmaps. They have press releases, PowerPoint slides and talk of carrier certification trials ongoing. What they do not have is real volume shipping.
By 2014, this will change, but probably not in the way most people expect. Our estimate is the next baseband vendor to ship tens of millions of units will not be a US Tier 1 chip designer. From what we could glean at the show, the next big vendor of LTE will be Mediatek. This is important not only to the handset market, but also says a lot about the changing nature of the semiconductor industry. Mediatek did not really have a 3G product until two years ago, a full six years behind the rest of the industry. Now they look set to not only close the gap, but entirely surpass it.
Admittedly, they have some advantages. First, much of their initial volume will be of the TD LTE Chinese variant of the standard. We believe their product will support FD mode as well, but probably most of their volume will be the ultimately less common strand. They are also likely to ship first into markets with less grueling carrier standards. A big delay right now for everyone else is this lengthy certification process, US carriers are particularly exacting.
That being said, Mediatek’s progress is still very impressive. Not least because there is a good chance that they may actually ship into a US phone this year.
Among the others, it is hard to distinguish. Both Broadcom and Marvell seem to be equally positioned around LTE. They should both ship later this year, having earned a few design wins, they are now struggling with carrier approval. Nvidia is in a similar camp, albeit they still show signs of needing to better integrate their Icera acquisition into their production roadmap. Most shocking is the lack of progress at Intel’s Infineon unit. The company currently has a data-only capable LTE modem which means further work to integrate their 3G stack. The roadmap for their fully baked 4G chip is also very vague. At heart, Intel is struggling to bring together the ARM architecture of Infineon’s products (along with their Ceva and Tensilica DSP cores) into the x86 architecture of Intel and their Atom applications processor line. This is Intel’s fifth struggle to enter mobile, and from the show floor it looked like this attempt was fraying around the edges. Intel’s problems with mobile go back a long way, a topic for a different note. Most people in mobile feel that, at heart, Intel does not understand the mobile industry and tries to use the same approach they brought to the PC world. For the man with the x86-shaped hammer, every problem looks like a PC-shaped nail. Signs that the company will pick yet another non-mobile internal candidate to be CEO gives us little confidence that they can overcome the problem.
Finally, we had a chance to catch up with the other two Asian baseband vendors – RDA Micro and Spreadtrum. Although these companies are both fierce rivals with Mediatek, we detected clear signs of happiness that Mediatek was doing so well in 4G. They both see this as validation that industry is moving in their favor. As both have LTE roadmaps, they rightfully see Mediatek’s progress as helping to make the case for their products with a growing array of customers and end-markets.