George Gilder
Forbes ASAP
October 1, 1994
Print ArticleNew low earth orbit satellites mark as decisive a break in the history of space-based communications as the PC represented in the history of computing. Pay attention to much-maligned Teledesic. Backed by Craig McCaw and Bill Gates, it is the only LEO fully focused on serving computers

-- George Gilder

"They'll Be Crowding The Skies."

THUS STEVEN DORFMAN, president of telecommunications and space operations for GM Hughes_the colossus of the satellite industry_warned the world of a new peril in the skies. Planning to launch 840 satellites in low earth orbits, at an altitude of some 435 miles, were a gang of cellular phone jocks and computer hackers from Seattle going under the name of Teledesic. Led by Craig McCaw and Bill Gates, they were barging onto his turf and threatening to ruin the neighborhood.

You get the image of the heavens darkening and a new Ice Age looming as more and more of this low-orbit junk_including a total of some 1,200 satellites from Motorola's Iridium, Loral-Qualcomm's Globalstar and Teledesic, among other LEO projects_accumulates in the skies. Ultimately, from this point of view, you might imagine the clutter of LEOs eclipsing the geostationary orbit itself, the so-called Clarke belt, some 21,000 miles farther out. Named after science-fiction guru Arthur C. Clarke, the geostationary orbit is the girdle and firmament of the Hughes empire.

In an article in Wireless magazine in 1945, Clarke first predicted that satellites in orbit 22,282 miles (35,860 kilometers) above the equator, where the period of revolution is 24 hours, could maintain a constant elevation and angle from any point on Earth. In such a fixed orbit, a device could remain for decades, receiving signals from a transmitter on the earth and radiating them back across continents.

The Clarke orbit also posed a problem, however_the reverse square law for signal power. Signals in space attenuate in proportion to the square of the distance they travel. This means that communications with satellites 22,000 miles away typically require large antenna dishes (as much as 10 meters wide) or megawatts of focused beam power.

Now, however, a new satellite industry is emerging, based on gains in computer and microchip technology. These advances allow the use of compact handsets with small smart antennas that can track low earth orbit satellites sweeping across the skies at a speed of 25,000 kilometers an hour at a variety of altitudes between 500 and 1,400 kilometers above the earth. Roughly 60 times nearer than geostationary satellites, LEOs find the inverse square law working in their favor, allowing them to offer far more capacity, cheaper and smaller antennas, or some combination of both. Breaking out of the Clarke orbit, these systems vastly expand the total available room for space-based communications gear.

It is indeed possible to "crowd" the Clarke belt_a relatively narrow swath at a single altitude directly above the equator. But even this swath does not become physically congested; collisions are no problem. The Clarke belt becomes crowded because the ability of antennas on the ground to discriminate among satellites is limited by the size of the antenna. Spaceway and Teledesic both plan to use the Ka band of frequencies, between 17 gigahertz and 30 gigahertz, or billions of cycles per second. In this band, reasonably sized antennas 66 centimeters wide can distinguish between geostationary satellites two degrees apart. That's some 800 miles in the Clarke belt. Thus no physical crowding. But it means that there are only a total of 180 Clarke slots for Ka band devices, including undesirable space over oceans.

LEOs, however, can be launched anywhere between the earth's atmosphere and a layer of intense radiation called the Van Allen Belt. The very concept of crowding becomes absurd in this 900 kilometer span of elevations for moving orbits that can be 500 meters apart or less. Thus the 21 proposed orbital planes of Teledesic occupy a total of 10 kilometers of altitude. At this rate, 70 or more Teledesic systems, comprising some 65,000 satellites, could comfortably fit in low earth orbits.

Nonetheless, it was clear that the LEOs, one way or another, were crowding Hughes. Hughes commands satellite systems or projects that compete with every one of the LEOs. Hughes responded to the threat of Teledesic by announcing the expansion of its Spaceway satellite system, then planned for North America alone, to cover the entire globe. Then, invoking the absolute priority currently granted geostationary systems, Hughes asked the Federal Communications Commission to block Teledesic entirely by assigning Spaceway the full five gigahertz of spectrum internationally available in the Ka band.

On May 27, Dorfman summoned the upstarts, Craig McCaw and Teledesic President Russell Daggatt, to Hughes headquarters in Los Angeles for a talk. Busy with Microsoft_the Redmond, Wash., company that in 1993 temporarily surpassed the market value of General Motors_Teledesic partner Bill Gates did not make the trip. But as the epitome of the personal computer industry, his presence haunted the scene.

Together with Spaceway chief Kevin McGrath, Dorfman set out to convince the Seattle venturers to give up their foolhardy scheme and instead join with Hughes in the nine satellites of Spaceway. Not only could Spaceway's nine satellites cover the entire globe with the same services that Teledesic's 840 satellites would provide, Spaceway could be expanded incrementally as demand emerged. Just loft another Hughes satellite. Indeed, Spaceway's ultimate system envisaged 17 satellites. With "every component proprietary to Hughes," as Dorfman said, the satellites only cost some $ 150 million apiece. By contrast, most of the $ 9 billion Teledesic system would have to be launched before global services could begin.

Nonetheless, the new LEOs marked as decisive a break in the history of space-based communications as the PC represented in the history of computing. Moreover, Teledesic would be the only LEO fully focused on serving computers_the first truly "global Internet," as McCaw's vice president Tom Alberg depicted it. It brings space communications at last into the age of ubiquitous microchip intelligence, and it brings the law of the microcosm into space communications.

If you enjoyed the New World of Wireless on the ground with its fierce battles between communications standards, technical geniuses, giant companies, impetuous entrepreneurs and industrial politicians on three continents_you will relish the reprise hundreds and even thousands of miles up. Launching Teledesic, McCaw and Gates were extending bandwidth abundance from earth into space. Observers, however, often did not like what they heard.

Bad Press For Two Billionaires

Every so often, the media is taken by the notion of technology as a morality tale. In place of a gripping saga of unjustly obscure geniuses enriching the world by their heroic creativity in the teeth of uncomprehending bureaucrats and politicians, the media treat technology ventures as a school for scandal. We have mock exposes of computer hype, monopoly, vaporware, viruses, infoscams, netporn, securities "fraud" and deviously undocumented software calls. Pundits gabble endlessly about the gap yawning between the information rich and the information poor, thus consigning themselves undeniably, amid many yawns, to the latter category. While American market share climbs near 70% in computers, networks, software and leading-edge semiconductors, analysts furrow the brows of the Atlantic Monthly with tales of farseeing foreign teams, spearheaded by visionary government officials, capturing the markets of American cowboy capitalists. They spiel implausible yams of tough-minded trade warriors prying open the jaws of Japan for Toys "R" Us, closing down vicious Korean vendors of low-priced dynamic RAMs, or blasting through barriers to U.S. telecom gear in the Tokyo-Osaka corridor, saving the day for Motorola's soon-to-be cobwebbed factories for analog cellular phones.

One of these sagas began early this year with two Seattle billionaires, McCaw and Gates, allegedly boarding McCaw's sleek yacht and going on an ego trip. With McCaw pitching in an early nickel, and the boat, and Gates hoisting his name as a sail, the two tycoons seemed to sweep away from the shores of rationality, as the media told it, into a sea of microwaves and arsenic. Spinning out Teledesic to build an information superhighway in the sky, they proposed to strew the heavens with 840 satellites, plus 84 spares. All would whirl around the world at a height of 700 kilometers (435 miles), using what they told the FCC would be some 500 million gallium arsenide microchips to issue frequencies between 20 and 60 gigahertz from some 180,000 phased-array antennas. The entire project seemed suffused with gigahertz and gigabucks. "We're bandwidth bulls," says Teledesic President Daggatt.

In case the hype of the sponsors failed to keep the system radiant and aloft, fueling it also would be a total of 12,000 batteries fed by thin film solar collectors stretching out behind the satellite "birds" in some 130 square kilometers of gossamer wings. Working at 4% efficiency, these cells would collectively generate 10 megawatts of power, enough to light a small city, but, so the critics said, insufficient to reach Seattle at microwave frequencies in the rain. (The Teledesic frequencies are readily absorbed by water in the air). To manage the elaborate mesh of fast-packet communications among the satellites and ground terminals, the constellation would bear some 282,000 Mips, or millions of instructions per second, of radiation-hard microprocessors and a trillion bytes or so of rad-hard RAM. In effect, Teledesic would be launching into space one of the world's largest and most expensive massively parallel computer systems.

At a mere $ 9 billion, to be put up by interested investors, Teledesic's lawyers told the FCC, the price would be a bargain for the U.S. and the world. (By contrast, current plans call for $ 15 billion just to lay fiber for interactive TV in California). But former Motorola, now Kodak, chief George Fisher fresh from pondering numbers for the apparently similar Iridium projects_suggested that $ 40 billion for Teledesic would be more like it. (Teledesic had the improbable result of making Iridium's 66-satellite plan, greeted in 1990 with much of the scorn now lavished on Teledesic, seem modest). Just rocketing the 840 satellites into orbit was said to entail a successful launch every week for a year and a half at a time when hoisting satellites is still a precarious and sometime thing.

Even if Teledesic succeeded in getting the things up, so other scientists suggested, the satellites would then be impaled on some 7,000 pieces of space debris in the chosen orbits. In any case, so it was widely reported, 10% would fail every year, some tumbling out of orbit, others joining the whirl of litter, where they would fly ready to impale the remainder of the satellites and the remnants of the two billionaires' reputations.

Surely these sages know that by the year 2001, when the systems would be up and running, the world will be swimming in the bandwidth of "information superhighways." Why support this lavish launch of technology for a communications system that would be dwarfed by capabilities already demonstrated on the ground?

Summing up a near-consensus of critics, John Pike, director of the Federation of American Scientists' Space Policy Project, declared to the Wall Street Journal, "God save us. It's the stupidest thing I've ever heard of!" Provoking Pike may have been the origins of the multisatellite architecture in the Star Wars "brilliant pebbles" program. Teledesic's most amazing achievement to date has been to displace the Strategic Defense Initiative as Pike's peak example of stupidity.

While McCaw and Gates could be dismissed as tyros in the satellite field, Hughes is world champion. Since 1963, the company has put 107 communications satellites into orbit. With 19 in 1994, this year should be its biggest ever. In 1993, well before the Teledesic announcement, Dorfman announced the first version of Spaceway_a $ 660 million, two-satellite system offering voice, data and video services_as a contribution to "information superhighways."

In the midst of all the terrestrial uproar surrounding superhighwaymen Al Gore, John Malone of TCI, Raymond Smith of Bell Atlantic and scores of other telco and cable magnates, however, no one paid much attention to Hughes.

Then came Gates and McCaw with Teledesic and claims of 20 million potential subscribers, two million simultaneous connections, billion-bit-per-second "gigalinks," bandwidth on demand and an array of other features, all advertised at a cost for Spaceway-type services nearly three times lower per bit per second. Everyone noticed Teledesic.

At the end of July, though, Hughes raised the stakes. With successful launches under way in China, Brazil and French Guiana to provide exclamation points, Hughes made a new submission to the FCC, extending Spaceway into a nine-satellite global system costing $ 3.2 billion. McGrath plausibly claimed it could be in place long before Teledesic and offer nearly all its functionality at a third of the price.

Already planned to be in place by 1998, however, were several other LEO projects, led by Motorola's Iridium and Loral-Qualcomm's Globalstar. As mobile phone projects, these systems could not readily offer service at T-1 data rates. But their sponsors promised availability for simple E-mail, faxes and paging.

By mid-1994, Motorola seemed to command the financial momentum. The company succeeded in raising some $ 800 million in equity investments from companies around the globe, including Lockheed and Raytheon (which would build the satellites), Great Wall of China and Khrunichev Enterprises of Russia (which together would launch a third of them), the Mawarid Group of Saudi Arabia (which pitched in $ 120 million) and Kyocera, Mitsui and DDI, which together put up another $ 120 million (Kyocera will build the dual mode handsets for Japan and DDI will sell and service them). On August 10, an Indian consortium purchased a 5% stake and a seat on the board for $ 38 million. Motorola claimed its share of the equity was dropping to 28.5%, well on the way to the company's final target of 15%. Motorola estimates that much of the additional $ 2 billion in the plan could come from debt securities and loans.

Iridium's attractions are impressive. It provides ubiquitous global phone service at a premium price with little or no dependence on local terrestrial facilities. In times of disaster or political crisis, or in places with sparse or unreliable local service, the system can route calls among the 66 satellites in space bypassing all infrastructure on the ground. For an elite of government officials and corporate figures operating in remote areas, the availability of Iridium should be worth the money. A bold and visionary concept when it emerged in 1987 from a team in the company's satellite systems engineering group, it endows many regions of the earth with voice and limited data communications for the first time. For example, it actually focuses on polar domains, such as parts of Siberia, poorly served by other satellite systems. Kazuo Inamori, the venerable chairman of Kyocera, also believes that Iridium will be popular in the 60% of territorial Japan not currently covered by cellular.

"Give Us Spectrum, Let Others Fight"

None-the-less, beyond the bold and ingenious concept (Daggatt calls Iridium "the real pioneer of LEOs"), the system suffers from technical flaws. Were it not for Globalstar, perhaps these flaws would not have become evident until alter the 66 birds were aloft. A far simpler and cheaper solution, Globalstar uses 48 satellites with no links between them. Each functions as a "bent pipe" transponder, receiving signals from a phone on the ground and passing them back to any gateway within the satellite's 1,500-mile-wide footprint, linked to locally available telephone networks. Because Globalstar uses local phone systems rather than bypassing them, the system has been able to raise a total of some $ 300 million in support from Alcatel, France Telecom, Vodafone (serving the United Kingdom, Australia and Hong Kong), Airtouch-U S West, Hyundai and DACOM in Korea, Deutsche Aerospace and Alenia.

This amount may seem small beside the billion raised by Iridium. But Globalstar has capital costs (at $ 1.8 billion) one-half Iridium's, circuit costs one-third Iridium's, and terminal costs (at $ 750 each) one-fourth Iridium's. With no intelligence in space, Globalstar relies entirely on the advance of intelligent phones and portable computer devices on the ground; it is the Ethernet of satellite architectures. Costing one-half as much as Iridium, it will handle nearly 20 times more calls.

The advantages of Globalstar stem only partly from its avoidance of complex intersatellite connections and use of infrastructure already in place on the ground. More important is its avoidance of exclusive spectrum assignments. Originating several years before spread-spectrum technology was thoroughly tested for cellular phones, Iridium employs time division multiple access, an obsolescent system that requires exclusive command of spectrum but offers far less capacity than code division multiple access.

Like conventional cellular or radio transmissions that differentiate signals by time slot or frequency, TDMA sharply restricts the reuse of spectrum in nearby cells. By contrast, CDMA is a form of spread-spectrum communications that differentiates signals by a spreading code and allows the use of the same frequencies all the time, everywhere. Just as you can reduplicate wireline spectrum merely by laying another fiber, you can now manufacture new spectrum in the air merely by breaking large cells into smaller ones.

Among some six companies seeking low earth orbit satellite approval from the FCC in 1993, only Iridium used TDMA, requiring national and international bodies to pick it as a winner from the outset and assign it exclusive spectrum. By contrast, in a majority report issued to the FCC on April 6, 1993, CDMA companies in the U.S., including TRW, Loral-Qualcomm, Celsat and American Mobile Satellite, could all agree to share spectrum and let the market choose winners. A Motorola lawyer explained to Space News, "Give us the spectrum and let the others fight for whatever's left." In the face of alternatives with no need for exclusive spectrum allocations, Iridium could fly only if it offered radically superior performance or capacity. But TDMA dooms it to generally inferior performance and capacity.

Unlike TDMA systems, which can "see" only one satellite signal at a time, CDMA handsets have "path" diversity, using "rake receivers" that can combine a number of weak signals into an intelligible stream. Iridium and other TDMA systems compensate by using more power. But no practical amount of power can propel a satellite signal through a tin roof. And excess power means larger handsets or heavier satellites. Iridium satellites together use 80% more power than Globalstar's, yet employ antennas nearly twice as large and offer 18.2 times less capacity per unit area.

Teledesic also suffers from the use of TDMA. But Teledesic's T-1 capabilities would compensate with 100,000 times more bandwidth and with a bit error rate that can accommodate the new fiber standards such as SONET-ATM (synchronous optical network/asynchronous transfer mode), which send packets without retransmission. The issue is whether these features can justify the political, financial, and performance costs of using a modulation scheme_TDMA_that severely limits spectrum sharing and path diversity.

So what is this, another saga of hubris on the information super-highway_to go with the Raymond Smith-John Malone follies? Perhaps good new ideas are harder to come by as company revenues grow into the billions, and Gates and McCaw disinvest and diversify as fast as they can from their increasingly cumbrous vessels of wealth. Having recently passed the billion-dollar mark in his systematic process of disinvestment from Microsoft_he retains $ 8 billion or so_Gates at times seemed embarrassed by his link to this gigantic project. He told us it was too early to write about Teledesic.

No, the story is in fact more interesting. Impelled by the onrushing rise in the cost-effectiveness of individual chips compared to multichip systems, the Law of the Microcosm dictates decentralization of all information architectures. During the 1980s, this centrifuge struck the mainframe computer establishment of IBM. During the 1990s, the personal teleputer, summoning and shaping films and files of images from around the world, will collide with the centralized establishments of TV broadcasting. At the end of the century, Teledesic and the other LEOs will usher in the age of decentralization in space.

From this point of view, Gates' participation becomes more readily intelligible. Gates seems always to follow the microcosm wherever it leads. A vision of software for decentralized systems of personal computers informs everything Microsoft does.

In 1994, for example, Microsoft made an investment in Metricom, a wireless terrestrial system that supplies links of up to 56 kilobits per second to portable computers or personal digital assistants. Within cells, the devices can communicate directly with one another; outside the cell, Metricom routes its calls through an expandable mesh of nodes each the size of a shoebox and costing less than $ 1,000. Based on spread-spectrum technology, the system operates at power levels low enough to avoid the need for FCC licenses. Yet it can be expanded to metropolitan-area dimensions.

In many respects, Teledesic is Metricom in the sky. It is focused on computer communications. It routes packets by the most convenient path through a mesh of nodes. It is based on microprocessor technology. (Both Teledesic and Metricom plan to employ devices from Motorola's 68000 family). As Gates explains the system: "Some functions are most efficiently performed by large numbers of small processors working together, rather than a few large ones." The entire new generation of low earth orbit satellite systems relies on this centrifugal force of the microcosm.

It was not supposed to happen this way. Just as Grosch's Law of the computer industry implied that computer power rose by the square of the cost, there was a similar law of the satellite industry that held satellite efficiency to be proportional to see. In a popular text, "Communications Satellite Systems," published in 1978, James Martin cited an AT&T study showing that just six satellites could carry all the long-distance traffic from the American continent; no fiber optics would be necessary. "The next major thrust in the space segment should capitalize on the economies of scale which today's technology offers," wrote Martin, urging creation of "massive hardware" as heavy as several tons and "immensely powerful satellites with large antennas beaming as much information as we are capable of using to our rooftops." Many satellite advocates, led by Arthur C. Clarke, viewed with impatient scorn the expensive terrestrial systems that somehow forestalled the manifest destiny of big birds to rule the world of communications.

Bringing The Microcosm To Space

In 1994, the big bird dream still flourishes in Spaceway, the international consortium Inmarsat, and the new launch this summer of direct broadcast satellite technology by Hughes's DirecTV, Hubbard's USSB, TCI's Primestar, and Rupert Murdoch's imperial systems in Europe and Asia. Using centralized satellites in geosynchronous orbits, DBS is the ultimate broadcast medium, reaching billions of potential customers at the cost of reaching hundreds of thousands through cable-TV systems. But these geostationary satellite systems suffer from the same flaws as mainframes: sclerosis by centralization. At a time when customers want the choice, control, convenience and interactivity of computers, the big birds offer one-size-fits-all programming at specified times, with little ability to control the flow or interact with it.

The real showstopper in the long run, though, is a nagging half-second time delay for Clarke orbit signals. Bad enough for voice, a half-second is near eternity for computer communications; for the living-room and desktop supercomputers of 2001, a half-second delay would mean gigabytes of information to be stored in buffers. While companies across the country, from Intel to Digital Equipment, are rushing to market with cable modems to allow computer connections to CATV coax, geosatellites remain mostly computer-hostile. Even with the new digital cosmetics of DBS, geosynchronous satellites are a last vestige of centralization in a centrifugal world.

By contrast, Teledesic brings the microcosm to space. Rather than gaming economies of scale from using a few huge satellites, Teledesic gains economies of scale by launching as many small birds as possible. Based on Peter Huber's concept of a geodesic network_a mesh of peers equally spaced apart like the nodes in a geodesic dome_Teledesic is not a hierarchy but a heterarchy. Distributing the system responsibilities among 840 autonomous satellites diminishes the requirements, such as message throughput and power usage, for each one. Building redundancy into the entire constellation, rather than within each satellite, yields higher overall reliability, while reducing the complexity and price of each unit.

As Craig McCaw explains, "At a certain point, redundant systems create more complexity and weight than they are worth. Rather than having each satellite a 747 in the sky with triply redundant systems, we have hundreds of satellites that offer self-redundancy." Eschewing the Hughes philosophy of "every component proprietary to Hughes," Teledesic will manufacture and launch a large number of satellite peers, using off-the-shelf parts whenever possible. This approach also provides economies of scale that, according to a study by brilliant pebbles contractor Martin Marietta, could lower unit costs by a factor of one hundred or more.

Just as microcosmic technology uses infinitesimal low-powered transistors and puts them so close together that they work faster than large high-powered transistors, Teledesic satellites follow the rules of low and slow. Rather than one big powerful bird spraying signals across continents, Teledesic offers 840, programmably targetable at small localities. Just 435 miles out, the delay is measured in milliseconds rather than half-seconds.

The total computing power and wattage of the constellation seems large, as is needed to sustain a volume of some two million connections at a time, four times Spaceway's capacity. But with other link features equal, between 1,226 and 3,545 times more power is needed to communicate with a geostationary satellite than with a LEO.

Perhaps most important, unlike Iridium, TRW's Odyssey, and Globalstar, Teledesic from the outset has targeted the fastest- growing market of the future: communications for the world's 125 million PCs, now growing some 20% a year. And Teledesic has correctly chosen the technology needed to extend computer networks globally_broadband low earth orbit satellites. The real issue is not the future of Teledesic but the future of Iridium.

In the short run Iridium's voice services cannot compete with Globalstar's cheaper and more robust CDMA system. But in the long run Iridium could be trumped by Teledesic. Although Teledesic has no such plans, the incremental cost of incorporating an "L" band transceiver in Teledesic, to perform the Iridium functions for voice, would be just 10% of Teledesic's total outlays, or less than $ 1 billion (compared with the $3.4 billion initial capital costs of Iridium). But 840 linked satellites could offer far more cost-effective service than Iridium's 66.

Iridium's dilemma is that the complexities and costs of its ingenious mesh of intersatellite links and switches can be justified only by offering broadband computer services. Yet Iridium is a doggedly narrowband system focused on voice.

Iridium eventually will have to adopt Teledesic's broadband logic and architecture. To protect its global lead in wireless communications and equipment, Motorola should join with Teledesic now, rather than later. Working with Lockheed, Motorola is making impressive gains in satellite-manufacturing technology. Supplying both handsets and space gear for computer networks, Motorola could turn its huge investment of time, money and prestige in Iridium into a dramatic global coup in wireless computer services. As part of a broadband system, Iridium could still become a superb brand name for Motorola. But persisting in a narrowband strategy in the name of avoiding Teledesic's larger initial costs, Motorola's executives will end up inflicting serious strategic costs on the company.

Most of the famous objections to Teledesic are based on ignorance or misinformation. Launch anxieties spring chiefly from the GEO experience. LEOs are 60 times nearer and between a tenth and a third the weight. Teledesic satellites are designed to be hoisted in groups of eight or more. From Great Wall in China to Khrunichev in Russia, companies around the world will soon be competing to supply low-cost launching facilities for the system. Orbital Sciences, an entrepreneurial dervish near Washington's Dulles Airport with some $ 190 million in revenues, has developed a low-cost method for lofting groups of LEOs from an adapted Lockheed 1011 Tristar.

Other fears are similarly fallacious. Teledesic will work fine in the rain because the high minimum vertical angle (40 degrees) of its satellite links from the ground reduces the portion of the path exposed to water to a manageable level. By contrast, geostationary satellites must operate at eight degrees, passing the signal through a long span of atmosphere. Made of tough new composite materials, Teledesic satellites will endure the kind of debris found in space mostly unscathed. The solar arrays can accept holes without significantly damaging overall performance. All in all, Teledesic's designers expect the birds to remain in orbit for an average of ten years. With most of its key technologies plummeting in price along with the rest of electronic components, the system may well cost even less and perform better than its business plan promises or George Fisher speculates.

Indeed, widely charged with reckless technological presumption, the designers of Teledesic in fact seem recklessly cautious in their assumptions about the rate of microchip progress. For example, their dismissal of CDMA assumes that the high speed of the spreading code functions_requiring digital signal processors that race at least 100 times the data rate_pushes cheap T-1 performance far into the future. Yet in early 1995, Texas Instruments will ship its multimedia video processor, a marvel that combines four 64-bit DSPs, a 32-bit RISC CPU, 50 kilobytes of on-chip memory, a floating-point unit and a 64bit direct memory access controller all on one chip. This device now performs two billion operations per second and, with an upgrade from 35 megahertz to 50 megahertz clock rate, soon will perform three billion. The estimated cost in 1995 is around $ 400, or a stunning $ 133 per bop (current Pentiums charge three times as much for 100 mips). Five years from now, when Teledesic gets serious, that kind of one-chip computing power can implement CDMA for broadband data without any cost penalty. Future generations of CDMA systems may be able to offer, at a dramatically lower price, the same broadband services in mobile applications that Teledesic now promises for fixed services only.

Assuming that Teledesic meets the CDMA challenge, the other fear is that terrestrial systems will capture enough of the market to render Teledesic unprofitable. This fear, however, can come true only if governments delay this supremely beneficial system well into the next century.

Unlike the competition, satellite systems can provide global coverage at once. Whether for $ 9 billion or $ 90 billion, no terrestrial system will cover the entire world, or even the entire U.S., within decades of Teledesic. As soon as it is deployed, it will profoundly change the geography and topography of the globe. Suddenly the most remote rural redoubt, beach, or mountain will command computer communications comparable to urban corporations today. The system can make teleconferencing, telecommuting, telemedicine, and teleschooling possible anywhere. Gone will be the differences among regions in access to cultural and information resources. People will be able to live and work where they want rather than where corporations locate them.

This change transforms the dimensions of the world as decisively as trains, planes, automobiles, phones and TVs changed them in previous eras. It will extend "universal service" more dramatically than any new law can.

Moreover, Teledesic can eliminate the need to cross-subsidize rural customers. Determining the cost of wire-line services are the parameters of population density and distance from the central office. Rural customers now cost between 10 and 30 times as much to serve with wires as urban customers do. Teledesic will bring near-broadband capabilities to everyone in the world at the same price.

Most important, this expansion of the communications frontier will foster the very economic development that will fuel the demand for the service. Today, it does not pay to bring telecommunications to poor countries that might benefit most. Teledesic and other satellite services break the bottleneck of development. Simultaneously opening the entire world, it enriches every nation with new capital exceeding the fruits of all the foreign aid programs of the era.

Teledesic is a venture worthy of McCaw and Gates. In its impact on the world, it may even rival the Herculean contributions of its sponsors in cellular and software. The issue is not the technology or the commitment of the principals. The issue is the readiness of the U.S. government to accommodate this venture. Before Teledesic can be approved internationally, it will have to attain a license from the FCC in the U.S. It has taken four years to approve Iridium. It took 30 years to approve cellular. How long will it take to approve Teledesic?

Currently Teledesic, Iridium and Globalstar face several political obstacles. The International Telecommunications Union's Radio Regulation 2613 gives GEOs absolute priority over LEOs. For Spaceway, Hughes is now demanding an exclusive license for the full five gigahertz available in the Ka-band worldwide, leaving no room for Teledesic or any other Ka-band LEO. Under current law, Hughes or other GEO systems could usurp any LEO that was launched.

LEOs are a major American innovation. The U.S. government should take the lead now in spearheading a change in the regulations to accommodate LEOs. This is no minor matter. As the dimensions and promise of Teledesic loom more starkly, the Japanese or Europeans are certain to make similar proposals. "When they do," Craig McCaw predicts, "they will immediately have their government on board. They will be able to go to the ITU right away. My greatest fear is that we will have the technology all ready, and foreign companies will beat us out because they can get their governments in line."

The U.S. government was on board for Apollo 25 years ago and the U.S. won the first space race. This space race is just as important, but the government is treating it as some sleepy-time infrastructure project. In fact, it is the information superhighway going global and ubiquitous. It is the ultimate promise of the information age, says McCaw.

Sustaining The U.S. Lead In Technology

McCaw explains: "It'll mean ecological disaster if China mimics what we did_building more and more urban towers and filling them up with people who queue up every day on turnpikes into the city, emitting fumes into the air, and then building new towers and new highways when you want to move the company, and then digging up the highways to install new wires."

McCaw waves toward the window, out at Lake Washington. "Look at that floating bridge. It took $ 1.5 billion to cross Lake Washington, then it got busted in a storm. Cross this lake, any lake, any ocean in the world with broadband wireless. That's the promise of Teledesic. All you do is to reconfigure the communications in software at zero incremental cost. No wires for the final connections. It's what we do in Hong Kong and Shanghai, where everyone uses a cellular phone."

President Clinton, Vice President Gore and other members of the administration continually ask what they can do for technology. One thing they can do is vastly streamline the process for approval of communications projects. At the moment, Congress is determined to retain bureaucratic dominance over the most dynamic enterprise and technology in the world economy_what they like to term the information superhighway. They see it as a possible source of congressional power, campaign finance, employment and pelf, like the Baby Bells today or like existing construction projects. Rather than turn telecom into a vast porkbellied poverty program, however, the administration should deregulate the field. Communications companies must be permitted to compete and collaborate wherever the technology leads.

Whether the administration knows it or not, these technologies are its greatest political asset. The high-tech industries unleashed in the 1980s by venture capital and junk bonds are now the prime fuel of the economy of the 1990s. Comprising perhaps 60% of incremental GDP and 48% of exports, the momentous upsurge of computers and communications is even compensating for the mistakes of the Bush and Clinton regimes and making plausible Clinton's continuing claims of economic success. But now Clinton, Gore and FCC Chairman Reed Hundt must make a choice. If they want to maintain this redemptive U.S. lead in technology, they must be willing to forge new alliances in Congress to get the politicians and bureaucrats out of the way of the future. A good start would be to open the floodgates for the global onrush of low earth orbit satellites dedicated to computer communications. If they do, they can help make the world, as McCaw's Alberg puts it, "a truly global Internet in an ever-expanding ethersphere."

And The Winner Is...

Globalstar is the easy winner for current offering of mobile phone services under a CDMA regime of spectrum sharing. But Teledesic can add phone services to its broadband computer system. Over time, Teledesic's 840 satellites will outperform Globalstar's 48. Big question: When will microchip technology advance enough to allow broadband applications over CDMA? When that happens, Globalstar has a shot at the grand prize.

Iridium is both too expensive to compete in mobile phones and too narrowband for data. Today's champ Spaceway is maturing. Big winner for the next decade is... Teledesic.