24 January 2008—Analyzing a U.S. wireless spectrum auction sounds about as exciting as watching a chess match. But as Bobby Fischer's recent passing reminds us, once in a while chess matches are pretty exciting.

Observers have called 700 megahertz the last great band of spectrum to come up for auction in this age of ubiquitous wireless devices. By virtue of having a lower frequency than today's third-generation services—which all operate at 1700 MHz and above—its signals can travel farther and better push their way through apartment and office walls.

All told, 214 companies have met the qualifications of the U.S. Federal Communications Commission (FCC) to participate in the complicated auction, which begins today and involves five different blocks of spectrum—A, B, C, D, and E—each with its own rules, advantages, and limitations. The spectrum, totaling 84 MHz, was freed up in the course of readying the United States' airwaves for digital TV. By February 2009, UHF channels 60 to 67 have to be off the air to make way for the auction winners.

The auction, with a total of 1099 possible licenses, could hardly be more complicated. All but a handful fall into the A, B, and E blocks, which together add up to just 30 MHz of spectrum and will be bid on in separate small geographical areas. The C block is broken up into 12 larger geographical areas, but you can go for nationwide coverage by making one of several permissible superbids.

The C block was already the most interesting chunk of spectrum to be auctioned off in years for three reasons: its large size, 22 MHz; its nationwide coverage; and its excellent propagation characteristics. Then, last summer, Google showed up to play.

Google offered to meet the FCC's minimum bid, which is now a hefty US $4.6 billion, so long as two conditions were met: the first was a rule that the winner had to allow onto its network any device or service, no matter who sold it, that complied technically with the communications standards and protocols the winner chose to use. That's in contrast to current spectrum licenses, in which the licensed carrier can make itself the exclusive gatekeeper of handsets and services, such as ringtones, sports scores, or Web access. Second, Google wanted the winner to be required to resell its network service on a wholesale basis. Another company could create a prepaid voice or data service, for example, that would operate in the same spectral band, and the licensed carrier would have to let it buy minutes or messages at a discount.

Though the FCC imposed only the first requirement, in November Google joined the roster of qualified bidders. And it's one of only a few companies with the big bucks needed to bid for a nationwide license.

Verizon Wireless also comes to the table with plenty of cash. The company objected to the Google-inspired rules, yet many observers expect it to be the most aggressive bidder. Tom Elliott, a vice president at Strategy Analytics, a marketing information company, says Verizon “is going to be pretty serious about it. They really want to build out their mobile data offerings, especially for delivering television and other video to cellphones.”

Besides Verizon and Google, the list of possible C-block bidders features many of the usual telco suspects, including AT&T and Alltel, as well as some other companies far from the telecommunications field, such as the oil and gas giant Chevron. But other major carriers will sit this one out, notably T-Mobile, which, less than a year ago, spent $3 billion on spectrum to beef up its third-generation data network, and Sprint, which is concentrating its resources on a next-generation network for which it already has enough spectrum. [See “Sprint's Broadband Gamble,” IEEE Spectrum, January 2008]

Another nonparticipant will be Frontline Wireless, a company that was created specifically to bid on D. That block gives the winner the right to use 10 MHz worth of nationwide commercial spectrum but comes with an unusual obligation to operate a parallel national emergency-services network. Two weeks ago, Frontline folded as a company, and now many observers wonder if anyone will come forward with the block's $1.3 billion minimum bid.

If you hope to follow the auction as it's happening, you may be in for a disappointment. For each round, the FCC will announce the amount of the highest bid for each license but not who offered it. It's a bit like hearing the moves of a chess tournament without knowing which player made them—and about as frustrating.

But anonymous bidding was needed to “reduce the potential for anticompetitive bidding behavior,” in the words of the FCC. Researchers looking into U.S. spectrum auctions going back to the 1990s—and as recent as the February 2007 auction in which T-Mobile spent $3 billion—have concluded that the major carriers were gaming the system. In April, economist Marshall Rose released a study in which he argued that “both empirical and theoretical evidence emerged that open auctions—auctions in which the identities and bids of all bidders were disclosed to the rest of the bidders—could produce anticompetitive, inefficient, and revenue nonmaximizing outcomes.”

Even if the action is obscured, surely industry observers have picked some favorites to win. Or maybe not. None of the analysts or other experts that IEEE Spectrum interviewed were ready to predict just who would bid on the C Block or how high the bidding might go. Nadine Manjaro, a senior analyst for wireless infrastructure at ABI Research, thought it might reach $6 billion, but perhaps not much more, “because you have to build [the network] as well. You don't want to put yourself in a hole, bidding so much for spectrum that there's no money left to provision it.” Carriers, she says, look at one figure in particular: the average revenue per user, or ARPU. “Even as wireless becomes a higher portion of revenue, actual month-to-month ARPU has declined, a sign that wireless service is becoming a commodity.”

Future networks, including ones based on the 700-MHz spectrum, will carry far more data traffic than voice. Both Verizon and AT&T have long-term plans to evolve their cellular services to a technology called, helpfully, Long Term Evolution. Noting that LTE requires a minimum of 20 MHz of channel bandwidth, Manjaro expects both companies to bid for the C Block to fill in their holdings. “Current networks might not have 20 MHz clear in every geographical area, and 700 MHz has better propagation than the [1700- to 2100-MHz] spectrum they already have.”

Action—or lack thereof—in the D Block's 10 MHz of nationwide spectrum is even harder to predict. The winner will have to build and maintain what the FCC calls a Public Safety/Private Partnership, using other spectrum in the 700-MHz range that has already been set aside for police, fire, ambulance, disaster relief, and other emergency and public services. One thing that will depress the bidding is an onerous build-out requirement. By 2019, the network has to cover 99.3 percent of the sprawling U.S. landmass. On the plus side, the operator can use the public safety part of the network commercially, so long as it doesn't interfere with public-safety traffic.

Frontline, which was formed by onetime FCC chair Reed Hundt and other high-profile industry insiders, seemed to have strong investor backing, including venture capitalists John Doerr and James Barksdale. And the nascent company was reportedly able to come up with the $128 million deposit needed to bid for the D Block. The $1.3 billion minimum bid, and a plan to make it back was, however, apparently a different matter. Mike Dano of RCR Wireless News predicts that no one will bid on the block. “Who in their right mind would want to share a network?” he asked in a 17 January 2008 column.

If no one does bid, the Public Safety/Private Partnership concept might fall apart. (Similarly, if the C Block's $4.6 billion minimum isn't met, the FCC can drop the “Google rules,” making the spectrum more attractive to traditional carriers like Verizon.) Mark Gaynor, a management and information technologies professor at Boston University, would like to see the FCC impose the Google-inspired openness rules on D Block and lower its price, or perhaps, in the public interest, give it away.

Even at the current minimum of $1.3 billion, Google just might fool everyone and bid on D instead of C, regardless of the formidable build-out costs. Running a public-service network with no clear business model might make sense to a company whose motto is still “Don't be evil.”