Risky business.

Jeffrey KraussIn light of recent legislation allowing the FCC to auction off TV broadcast spectrum, it might be instructive to review some ill-fated radio spectrum plays.

Prior to the 1970s, the radio spectrum was largely used for either broadcasting or private communications, like taxicab dispatching and point-to-point microwave links. But starting in the 1970s, commercial use of the spectrum expanded, with satellite networks and telephone competitors like MCI using it.

And that was the start of a whole raft of competitive radio spectrum ventures, some more speculative than others. Many of them are dead and gone. Many found it was easier to convince the FCC than the marketplace. Some didn’t make it because the technology didn’t work or was more expensive than expected; some didn’t make it because the business plan was bad; and in many cases, it was a combination of factors.

The earliest one I can recall is XTEN, the Xerox Telecommunications Network, which was proposed in the late 1970s. The plan was to bypass local telephone company wires and offer document interchange service to businesses. This was before email, and the documents would be scanned in as faxes for transmission. Xerox asked the FCC to allocate spectrum, and the result was a service the FCC called Digital Electronic Message Service (DEMS), with some (but not much) spectrum allocated around 10 GHz. In the early 1980s, packet switching companies like Tymnet received FCC DEMS licenses, and a few transmitters were built, but they were expensive, and the market never developed.

Later, the FCC allocated more spectrum for DEMS at 18 and 24 GHz. In the 1990s, Teligent acquired licenses for DEMS spectrum at 24 GHz and built out local networks in about 25 markets. But Teligent was never able to compete with telephone company services and filed for bankruptcy in 2001. Around that same time, Winstar, which planned to deploy local networks based on microwave licenses at 39 GHz, also filed for bankruptcy. The planned point-to-multipoint microwave technology was not mature, and competing against the telephone company was a losing proposition.

Another late-1990s local network hopeful was Northpoint. Northpoint convinced the FCC to create the Multichannel Video Distribution and Data Service (MVDDS) in the 12.2-12.7 GHz band, the same band used by DirecTV and Dish Network for their broadcasting satellite downlinks. Northpoint argued that since home satellite dishes all point to the south (the satellites are located over the equator), that same spectrum could be reused by roof-mounted dishes that point to the north. But there were continuing questions about interference to home dish satellite reception. Some companies bid and won the auctions for MVDDS licenses, but none of them built successful networks.

The satellite field is also full of shaky ventures. While fixed-satellite networks like Intelsat and SES Americom have been around for years, proposals for mobile satellite networks (satellites that communicate with handheld mobile devices) and non-geostationary satellite systems have foundered.

Globalstar has been in and out of bankruptcy, and it recently had some licenses revoked by the FCC. Iridium, a network of 66 non-geostationary satellites developed by Motorola, was deployed in 1998 at a cost of $6 billion, but it quickly went into bankruptcy in 1999 because it could not compete against cellular phones. Private investors bought the assets for $25 million and successfully marketed the high-security capability to U.S. government customers. But it remains to be seen whether they can afford to replace the satellites when they reach their end of life.

Two other elaborate non-geostationary satellite systems were planned during this time period. Teledesic, funded by Bill Gates and Craig McCaw, originally planned to deploy more than 800 satellites, and then it cut back to 288 satellites, and then it scaled back even more. Teledesic’s plans were finally scrapped after it spent millions to achieve FCC spectrum allocations and international ITU-R recognition, but before it got into operation.

SkyBridge was another non-geostationary satellite system that was planned by the French company Alcatel. Like Teledesic, SkyBridge got FCC licenses and international ITU-R recognition in the 1990s before closing up shop.

Another forgettable venture was Sky Station, which got the FCC to allocate spectrum for communications between the ground and stationary platforms (blimps) that could hover for months at an altitude of 18 miles. Somebody spent a lot of money designing the technology. The FCC allocated spectrum, but no system was ever deployed.

And next we come to LightSquared. LightSquared has spent billions on a plan to deploy a land mobile, cell phone-like network using spectrum originally intended for satellite use. But the frequencies are adjacent to frequencies used by GPS receivers, and LightSquared transmissions would cause interference. The basic technical issue is that GPS receivers should have been built with better (more expensive) filters. But they weren’t. And so it became a political problem and a problem that LightSquared could not solve.

Most recently, Charlie Ergen got FCC approval to acquire two other bankrupt mobile satellite licensees: TerreStar and DBSD. Ergen hopes the FCC will convert these satellite licenses to terrestrial, since his frequencies are farther away from GPS than LightSquared’s are. Keep your fingers crossed, Charlie.

So maybe the lesson is: Spectrum plays are risky. Stick with wires and cables.