Qualcomm is getting back into the network operator business it left years ago, this time investing $800 million in capital in a nationwide mobile multimedia network that the CDMA inventor hopes will jump-start the nascent video-streaming market.

Instead of optimizing existing CDMA technology, deploying over standard wireless spectrum and using time-tested business models, Qualcomm is rethinking the traditional notion of the wireless network. The company is building out a broadcast-only network called MediaFLO, capable of distributing dozens of one-way video and audio feeds over long distances (see figure). It will send those signals over the 700 MHz TV spectrum — channel 55 on the rabbit ears — that the company won last year at auction and will use a new multimedia transmission technology to deliver them.

Qualcomm is designing a new line of dual-mode, multiband chipsets to support those new frequencies and technologies. Most notably, it's handing the service derived from this network over to carriers entirely on a wholesale basis, giving them a bevy of common but crucial content options from a shared reservoir so carriers can reserve and develop their own data channels for higher-revenue content services.

“This network is in no way a substitute for cellular networks — the two networks will complement each other,” said Paul Jacobs, president of Qualcomm's wireless and Internet group, during an earnings call last week. “It points to something we believe will become increasingly commonplace — heterogeneous networks operating in conjunction with each other and offering a wide variety of services, all of which can be accessed by customers on their mobile devices.”

Some have questioned whether Qualcomm would have another hit on its hands along the scale of BREW or its initial launch of core CDMA technology, but few disputed that Qual-comm is once again living up to its reputation for unique thinking. While Qualcomm has been putting the elements of the proposed MediaFLO network together for more than a year — buying spectrum in 2003 and unveiling its Forward Link Only (FLO) multicast technology in March — the announcement came as a surprise to much of the industry.

“It's a very bold step,” said Michael Grossi, vice president of the wireless practice of Adventis. “It's obviously very risky and will take a lot of capital, but it's exciting to see Qualcomm develop an entirely new type of business model. A carrier certainly wouldn't have thought of it.”

FLO is designed to transmit from 50 to 100 separate national and local channels, including up to 15 live streaming channels, which Qualcomm hopes to populate with optimized content from major TV providers as well as unique mobile content. The platform delivers video in the quarter VGA format for small screen sizes at 30 frames per second and stereo audio, all at much lower power than a typical television broadcast, thus sparing precious mobile device battery life.

Because of the propagation characteristics of the 700 MHz band and the ability of broadcast multicasting architectures to reinforce signals, Qualcomm anticipates low tower density. Rob Chandhok, vice president of engineering at MediaFLO, said the average market will need between two and three transmitters, each situated on high towers — very similar to the layout of a broadcast TV network. This is far fewer than the dozens of cell towers it usually takes to blanket a typical cellular market.

“We're not taking an existing terrestrial technology and tweaking it,” Chandhok said. “This is not television. We're creating a low-power consumption system, optimized for mobility and in-building coverage. It's intended to take the load off of the cellular network.”

The service, however, is not meant to replace carriers' own streaming services, many of which are in their infancy, Chandhok said. Rather, it is intended to provide a foundation of common content most carriers wouldn't launch on their own — the equivalent of basic cable over the wireless network. Qualcomm is licensing out its MediaFLO technology to carriers so they can deploy their own interactive streaming content over their own W-CDMA or CDMA 1X EV-DO networks. Those streams will merge with Qualcomm's wholesale channels in a common client on the handset, so carriers can create integrated but differentiated content packages for their subscribers, Chandhok said.

CDMA Development Group executive director Perry LaForge said Qualcomm is hoping to emulate the success east Asian carriers have had with video streaming over EV-DO and UMTS networks. LaForge said that in South Korea, SK Telecom takes in its largest chunk of data revenue from video services and has used that content to drive 3G adoption to extremely high levels.

“Qualcomm is trying to do the same thing it did with BREW,” LaForge said. “They believed there was a market for downloading applications, so they decided to get the market started. Now they believe there is a market for streaming video. They want to jump-start that market, too.”

While Qualcomm expects media partners to be excited about the opportunity it is presenting, at least one established mobile video provider expressed some reservation. Phillip Alvelda, CEO of Idetic, which runs MobiTV over the Sprint and AT&T Wireless networks, said he was “cautiously optimistic” about Qualcomm's plans. While MobiTV is thrilled that a major industry player is throwing its weight behind mobile multimedia streaming, Alvelda said he's not convinced deploying an entirely separate network based on an entirely new transmission technology is the best way to encourage the industry.

“We want to avoid a new Iridium,” Alvelda said, referring to Motorola's failed $5 billion satellite phone venture.

Qualcomm's investment will be more modest. It estimates the capital costs for a nationwide deployment at $800 million over the next four to five years and expects to share costs with partners — presumably including carriers using the service. So far, the only carrier to publicly express interest is Sprint, and Qualcomm hasn't secured the commitment of any content providers yet, though Jacobs said some major content companies have expressed interest in investing.

There will be time for interest to develop. Qualcomm does not expect to have its first commercial market operation until 2006, and until that time, the vendor has to work out the regulatory mess surrounding its 700 MHz spectrum, which is still occupied by TV broadcasters in many markets.

Those broadcasters are supposed to vacate the airwaves by 2006, but regulatory caveats could alter that plan. Qualcomm said it would match its deployment plans to the conditions of the airwaves in each market, launching first in markets where no broadcaster occupies channel 55 or neighboring channels. In markets where channel 54 and 56 are occupied, the company is laboring to demonstrate to regulators and station owners that it can broadcast MediaFLO without interfering with neighboring signals. As for markets that have broadcasters sitting squarely in the middle of its spectrum, Qualcomm said it is working with the FCC and Congress to establish a hard date for those broadcasters to convert to digital or leave.

News Editor Dan O'Shea contributed to this story.