There is little information to be teased out of the bland statement made this week by Qualcomm that its MediaFLO mobile TV technology will be used to launch services to Verizon customers.
The move was a complete surprise and it may have been a late change of mind. US analysts had predicted in October that the deal had in fact gone to Crown Castle and Verizon was supposed to be trying out the Crown Castle DVB-H technology.
Crown Castle has been operating a trial in Pittsburgh for almost a year as it waits for the right handset equipment to be ready once DVB-H chips become available in volume.
When we spoke to Crown Castle last month it would not confirm that it had signed with any operators yet, but talked about its plans to rollout a DVB-H network to the top 10 markets in the US by Summer 2006 and the top 30 markets by the holiday season in 2007. The cost of all this has been estimated by outsiders at some $500 million.
CFO Jay Brown would only confirm that Crown Castle had not yet announced any commercial partners, but that it expected to have them all lined up by the end of 2005 or early in 2006, including strategic investors in the business, likely to be content partners, equipment makers and hopefully a major US cellular operator. Crown Castle owns 11,000 US tower sites where cellular base stations are situated for each of the US cellular operators, and will use these sites to locate its DVB-H transmitters.
To a certain extent Verizon and Sprint, the 2 major CDMA carriers in the US, were the most likely Qualcomm targets since they already rely on Qualcomm technology for their basic cellular link, and are progressing down its data extension roadmap.
Cingular, and its nascent AT&T Wireless brand, the largest mobile carrier in the US, is just in the process of rolling out UMTS or W-CDMA, which Qualcomm has a lesser claim over in terms of intellectual property and market share, so it may be more comfortable working with DVB-H and Crown Castle, which partly originate with Nokia.
While neither of the US mobile TV broadcasting networks, MediaFLO or Crown Castle, need to remain exclusive to any one operator, it already requires the provision of an extra radio and antenna and more error correction logic, to be added to each of the handsets. To take both services would mean adding twice as much to the phones, and this would not only be too cumbersome for the phones, but increase the costs too substantially. Of course an operator could provide two separate types of phones and let the consumers decide which TV services they want, but they might lose out on economies of scale.
So it’s almost a certainty that each of the cellular operators will throw their lot in with a single mobile TV broadcaster, certainly until exiting analog TV spectrum is released in the 2008/9 time frame, when new players may emerge. In the same timeframe, RF chips that can scan much larger chunks of spectrum, defined purely by software (software defined radios) may also emerge, and then a subscription to multiple mobile TV services may be a simple download and play. But until then we expect it to be a case of one handset, one TV service.
Sprint Nextel has committed to building its own network with its existing holding of 2.5GHz spectrum and working with the US cable TV operators, although it may throw in its lot with either MediaFLO or Crown Castle in the short term. But eventually that leaves the US market looking like a 3 way draw, unless Crown Castle fails to land Cingular after all, and this too falls to Qualcomm. But that would mean the two operators would have to try to compete, offering the same or overlapping video content, so it’s unlikely.
One of the things that weighs heavily in the favor of Crown Castle is that it uses DVB-H technology, which has been chosen for around 20 trial installations around the world and which is far more heavily supported than either MediaFLO or the Korean alternative T-DMB (or its satellite cousin S-DMB). Because of this, handsets will enjoy greater economies of scale and are closer to being ready. Qualcomm will have to design all the chips for MediaFLO itself and the chips at least are likely to be more expensive, even if the handsets may not be (less profit for the handset maker).
However this win for Qualcomm means that a substantial proportion of Verizon top end handsets will have to have the Qualcomm chips in them or the MediaFLO designs will have to be licensed by chip companies like Texas Instruments or handset makers like Samsung, Nokia and Motorola, putting Qualcomm in the position of charging a hefty fee for either its intellectual property or for its chips.
Qualcomm says that in total it will have invested some $800 million in the MediaFLO technology and network by the time it is fully rolled out. It claims it is a superior technology to DVB-H and that it can deliver 20 parallel mobile broadcast TV channels offering full QVGA screens (320 x 240 pixels) in just a 6 MHz slice of 700 MHz spectrum.
Crown Castle bought spectrum up at 1670 MHz, so it will need more transmitters to create a network because radio propagation and penetration is not quite so good at that frequency. It will offer 16 TV channels and 30 channels of audio.
But Crown Castle is thought to be slightly ahead of Qualcomm in its ability to roll out, especially as some of the Qualcomm spectrum is still being used by analog TV in some of the US markets.
Verizon will have to rely on cellular delivered TV services, supplied by its Vcast service until the MediaFLO network is built out, and it is not likely to be the first to market and may lose between 6 and 9 months over any company that opts for Crown Castle as a partner. The Crown Castle service is ready to build out now and can hit the market in the first quarter of 2006, versus the last quarter for MediaFLO. Perhaps that’s why it was important to announce the MediaFLO deal first, to try to preserve some impression of Qualcomm being ahead.
Crown Castle also pointed out to us that, ‘We could launch the service without a major operator behind us. We could work with MVNO customers, or subsidize the handsets ourselves, but it won’t just be phones that tune into DVB-H services, PCs and PDAs are all likely to come DVB-H enabled and we know that both Intel and Dell have a big interest in this.’ So there are going to be plenty of mopping up operations, even after the big cellular partnership battles are over.
Texas Instruments, which makes the Hollywood chip for this market recently told Faultline that existing intelligent devices could add DVB-H functionality for just $10 per device, in volume, that being its estimated bill of materials for a full system as long as it is on a device that already has its own processor.
Qualcomm and Verizon will launch mobile TV services in around half the markets covered by Verizon EV-DO upgraded networks and gradually grow towards US wide coverage, the company said.