Eccentric Orbits: The Iridium Story was one of the Wall Street Journal “20 books that defined our year” for 2016. It isn’t perfectly edited, but it is rewarding for anyone interested in technology.
The Iridium satellite phone system was developed starting in 1988 by three engineers at Motorola (Ray Leopold, Ken Peterson, and Bary Bertiger). What was unique about the system was that there was a lot of satellite-to-satellite communication, an idea lifted from the Star Wars program.
Bertiger was a veteran of Cold War military systems, having designed most of the microwave communications for America’s spy satellites, and Leopold had been acting director of the Milstar Terminal program at Hanscom Air Force Base in Massachusetts. Milstar was the $ 5 billion super-satellite warfare system being built by the Pentagon to link bombers, nuclear missiles, submarines, fighter planes, and troops on the ground. Given “highest priority” by President Reagan to support Star Wars, Milstar had cross-links among eight satellites—ten-thousand-pound “switchboards in space” that cost $ 800 million apiece—so Leopold did know a little about linkage issues.
In theory the whole system could have operated with a single base station for bridging satphone calls into the terrestrial phone system. More than $ 6 billion in investor money was spent on the system. It went live on November 1, 1998 and bankrupt on August 13, 1999:
It was the biggest bankruptcy filing in the history of the United States—and the quickest. The company had been open for business a little over nine months.
The author attributes the insolvency to (1) slow adoption by customers, and (2) Motorola sucking all of the money out into its own pockets.
Karl Marx expressed sympathy for the capitalists who risk a lot of money on a project and then discover that there is no demand or there is too much competition. He thought that everyone would be happier with a centrally planned economy in which if you build it the customers will definitely come.
Iridium was conceived at a time when cell coverage was ridiculously bad and roaming was impractical. The idea was to sell to business travelers. By the time the satellites were up, the European GSM system was in place almost everywhere that rich business people might go. The natural customers were in specialized markets such as aviation, shipping, and military. I’ll cover the rescue out of bankruptcy in a subsequent posting.
As with most revolutionary ideas, the idea was not revolutionary at the time.
And yet creating a constellation of communications satellites that would cover the whole planet was not an entirely new idea. When AT&T launched Telstar in 1962, they expected it to anchor a system of forty satellites in polar orbits, fifteen satellites in geostationary orbits, and twenty-five ground stations positioned around the globe. The system would have depended on ordinary trunk lines to complete calls, but it was still acclaimed around the world as the most sophisticated satellite system ever launched and the future of communications for the planet. That promise was snuffed out by President Kennedy, who was alarmed by the very fact that AT&T was prepared to spend $ 500 million on the system, and the result was COMSAT, which ended up not being interested in voice communications at all.
Two scientists at Aerospace Corporation, William S. Adams and Leonard Rider, had recently published a theoretical article in the Journal of the Astronautical Sciences called “Circular Polar Constellations Providing Continuous Single or Multiple Coverage Above a Specified Latitude.”
Details matter, however:
When the inventors first talked about using LEOs, they envisioned an orbit of six hundred to eight hundred miles up. Satellites at that altitude would have to fly through the inner Van Allen Belt, a mass of charged particles discovered by Explorer 1. In fact there were two radiation belts at those altitudes, the other one having been created artificially in 1962 when both the United States and the Soviet Union exploded nuclear bombs in space. Once again Aerospace Corporation had the data needed. George Paulikas, the Aerospace engineer who stood by at all NASA launches to describe “solar particle events” in real time, had studied every aspect of radiation belts, and his research indicated that small satellites constantly flying through that plasma would take enormous punishment every day, regardless of what material they were made of. Eventually Leopold decided to go lower—420 nautical miles—using a seven-by-eleven constellation. There would be seven orbital paths converging at the poles, and in each of those orbital paths there would be eleven satellites. This meant that they would need seventy-seven satellites to make sure the entire Earth was covered.
The need for worldwide spectrum allocation and regulatory approval disfigured the project so that there were gateways in various places around the world and all kinds of different pricing plans. The legacy carriers fought hard in the regulatory domain:
The nations of the world were gathered in the neighboring town of Torremolinos [in 1992] to allocate radio-wave spectrum, an event called the World Administrative Radio Conference that is held infrequently—this was the first major one in thirteen years—but has a lot to do with what the electronic future of the world will look like. … A decision had been made in 1990 to “proceed on the assumption that we’ll get the frequency we need,” but the world had changed greatly from the days when big American and British companies could tell everyone else how the airwaves would be used. Almost as soon as Iridium was announced, a chorus of “You can’t do that” had gone up from several other corporations, including every national phone company in Europe. … That’s why Motorola had spent the previous sixteen months lobbying the entire world to make sure it got what it wanted: radio frequency bands that could be used to operate the first point-to-point global telephone system, not to mention the first commercial switching system in outer space. The army of Motorola employees sent to Torremolinos far outnumbered that representing the U.S. government. The United States sent teams from the FCC, Voice of America, Department of Commerce, Pentagon, State Department, NASA, National Science Foundation, Coast Guard, U.S. Information Agency, and FAA—but all of those delegations combined were still smaller than Motorola’s team. Since Motorola had offices around the world, the company was able to identify political allies in advance, but the company’s war plan went one step further and made sure that Motorola employees were named as actual voting members: the United States, Canada, France, and Australia all had Motorola employees sitting in their official delegations. Add to this the fact that Travis Marshall, Motorola’s chief lobbyist, was the U.S. ambassador to the International Telecommunication Union, which administers the conference, and you start to understand why many of the WARC delegates were resentful of the pressure, regarding the Motorolans as crass salespeople determined to hold them hostage, treating them like reluctant participants in the world’s biggest time-share presentation.
Motorola prevailed primarily by getting poorer countries to sign up:
Iridium would be the greatest thing to happen to the Third World since . . . well, since the United Nations was formed. At last every country, and every village in every country, could be connected to the worldwide grid. In actuality the Iridium business plan would not be focused on the Third World at all, but on the well-heeled executive travelers in North America, Europe, and Japan, but for the time being it was better to talk about straw huts in Papua New Guinea, not ski chalets in Gstaad. A promotional video for Iridium featured the President of Mali, his wife, and his staff in acting roles. After a while, Motorola’s incessant statements of love and affection for the outcasts of the world started to wear thin. An observer for the U.S. Office of Technology Assessment drily remarked that the average citizen of the Central African Republic would have to work for four years to earn enough money to purchase an Iridium phone, then work seventeen hours more to pay for a one-minute call.
Does your country have an entrenched phone company that fears Iridium? No problem—we’ll tack a dollar on to the “tail charges” of every Iridium phone call made from your country and send it back your way. If idealism didn’t work, maybe greed would. “The fifty-four African countries were used to getting checks every month from AT&T,” said Mondale, “so we had to agree not to undercut that direct-dial service.” This stratagem would come back to haunt Iridium in later years, as national telephone companies routinely asked for kickbacks disguised as fees—tiny Madagascar wanted $ 500,000 a year—just to keep the Iridium license in place.
Incumbents such as Inmarsat fought like tigers.
In France it was worse. Heading up the Motorola diplomatic effort there was Leo Mondale, nephew of Walter Mondale, the U.S. Senator who had served as Vice President under Jimmy Carter. Mondale was a talented communications lawyer who had worked in the Paris offices of Fairchild Space and for the aeronautics division of defense contractor Mécanique Aviation Traction (better known as Matra), and he was the first hire at Iridium, partly because Motorola thought his connections could bring the big European telecoms aboard. The initial meeting at France Télécom turned out to be an elaborate farce, during which the French executives affected bonhomie for their frères from across the pond while fishing for competitive information—but that wasn’t the worst part of the experience. Someone had managed to place listening devices in the first-class cabin of the Air France flight that carried the Motorolans to the meeting, so the Paris executives knew exactly what Motorola was trying to do and how they were trying to do it.
At the WARC, France Télécom delegates were telling anyone who would listen that participation in Iridium was a violation of the Inmarsat treaty. Then, after the Iridium system was patented anyway, France launched a complex and expensive legal challenge that resulted in a series of hearings before the European Patent Office in Munich.
After they got the spectrum they still had to build and launch nearly 100 satellites (the constellation plus spares):
He kept reminding Lockheed Martin, Raytheon, and everyone else engaged in the project that the final assembly of a satellite built by NASA or the Air Force took anywhere from nine months to one year per satellite. But since Motorola needed a hundred satellites, including all the spares and demonstration units, “obviously we don’t have a hundred years to build this constellation.” If Iridium had been a government project, for example, Stamp would have had to perform a thermal vacuum test on every satellite, placing it in a chamber that simulates intense radiation and the other brutal elements it would eventually be exposed to in space. The process was time-consuming and expensive, so Stamp made an early decision: they would “thermal-vac” the first one off the assembly line and assume all the rest were sound.
Stamp was also innovative about his assembly line, setting up his factory in Chandler so that the satellites were built horizontally instead of vertically. That meant the assemblers could work at waist level, whereas on military satellites, the technicians stood on ladders and leaned over. He trained Lockheed Martin in the new technique, then later bragged about it to his friends at Khrunichev in Moscow, only to be told, “We’ve done it that way for forty years.”
Stamp started looking at the specs and launch histories of every rocket in the world and eventually told Motorola, “Look, I feel much safer on a Chinese or a Russian rocket than an American one.” All the American rockets were handled by either the Air Force or NASA, all of them had spotty launch histories, few of them had the power he wanted, and he didn’t trust the people in charge.
When he was first hired, Dannie Stamp shared an office with the Iridium inventors and was the only employee of the Space Segment Division. By 1995 he was supervising a hundred people and had executed contracts promising 90 percent of the cash on hand: $ 850 million to Lockheed Martin, $ 350 million to Raytheon, and hundreds of millions for rocket launches at Vandenberg, Baikonur, and Taiyuan. At the project’s peak, there were five thousand engineers working exclusively on Iridium. By the time his satellite factory was operational, Stamp was spending $ 1.5 million a day and was supervising more space vehicles and launches than NASA, the military, or any other nation.
Over the next twelve months, Iridium would break every record in the history of rocket science, launching 72 out of 72 satellites on 15 out of 15 rockets in 377 days. Nine of the rockets were Deltas launched from California, with three launches from Kazakhstan and three more from China. In one thirteen-day period, Stamp’s team ran launches from all three space centers—Vandenberg, Baikonur, and Taiyuan—putting 14 satellites into orbit. After the constellation was operational, there were an additional 23 spares launched without flawing the perfect record—95 satellites loaded, 95 satellites launched. The “infant mortality” rate (satellites that never become operational because they fail before being switched on) was an amazing 2 percent.
There was plenty of enthusiasm among investors:
When the first Iridium stock offering was made in 1997, working stiffs in Arizona bought five and ten shares at a time, certain that this was equivalent to the invention of the telegraph or the telephone.
The initial investment by the eighteen global partners accounted for less than half of the start-up costs. The rest was covered with $ 1.45 billion in junk bonds and with common stock that was being aggressively marketed by Merrill Lynch. The unpleasant truth was that Iridium would need a million customers in the first year just to break even, and Staiano wasn’t sure how to do that. “The marketing was scary as hell,” he recalled. The phone couldn’t even be pictured in advertising because it was so enormous.
The brightest minds in American journalism and financial analysis were also enthusiastic:
the media didn’t really question the issue of consumer demand. It was assumed that everyone would want one.
Merrill Lynch, on the other hand, was so wildly optimistic that analysts seemed to be pulling low-ball numbers out of a hat. Retail service would be only $ 1 per minute, it reported, and phones would sell for $ 1,000. Staiano blithely told the media that it didn’t matter either way: billable minutes would exceed all expectations because of pent-up demand. Merrill Lynch agreed, pointing to its projections of 4.7 million users by the year 2007.
In fact, the financial community as a whole was very bullish on Iridium. A report released by Credit Suisse/First Boston on December 23—right before the banking agreements were signed—predicted that Iridium would have $ 4 billion in revenues by 2002, with 2.9 million paying customers, rising to 10 million by 2008
The CEO himself was a true believer: “He bought $ 1 million of Iridium stock with his own money.”
Iridium and Motorola had borrowed money with conditions attached
[an executive] rebooted his computer and checked the dashboard again. Finally, a few figures started to change. He began to see signs of the occasional call here and there. Then he suddenly realized—those were all test calls. Those were all either free phones or calls being made by Iridium employees. What if a $ 6 billion company opened for business—and nobody came?
His immediate problem was a billion-dollar bridge loan from Chase Manhattan Bank that would be coming due on December 31, 1998. For the average person, it might be hard to understand how a company could raise billions from partners, borrow more billions from banks, issue bonds to raise more billions, issue common stock to raise hundreds of millions, and then end up facing a panicky billion-dollar deadline on opening day. That’s why Staiano didn’t try to explain it to the average person. He didn’t even try to explain it to the average Wall Street analyst. This was the nineties. This was how things were done. In order to convert that loan into something more permanent—and avoid defaulting on it and causing cross-defaults of another billion—the Iridium system had to be fully operational. Those were the rules. Back in the more collegial days of 1996 and 1997, the bankers at Chase Manhattan had done the math and spoken optimistically of a $ 1.7 billion “permanent facility”—funds that Staiano could draw down on—but only after the company was open for business.
the covenants were set at around 22 percent of stated projections. The bank just wanted to be assured that 1) subscribers were being signed up, 2) subscribers were using the system, and 3) subscribers were paying their bills. After all, if you failed to reach 22 percent of your own business projections, something might be seriously wrong.
But they couldn’t meet the conditions:
Still, no one really believed that Chase Manhattan would pull the plug, so the goal going forward was to make the bankers feel good by whatever means. That meant not sharing the fact that, as of January 22, 1999, after two and a half months in business and several months prior to that trying to sell the service, Iridium had precisely 3,637 customers. Even adding a zero to that number wouldn’t get the company to the level that would satisfy the bank covenants, which were extremely conservative (52,000 customers by March 31).
The executives and Wall Street offered stock to the public while some of the insiders were unloading their own:
On January 27  Iridium collected an additional $ 250 million in what was trumpeted as a successful public stock offering. Public ownership of Iridium was now 13.25 percent.
On Monday, February 1, Mauro Sentinelli [top sales executive] sold all his Iridium stock.
The banks accepted the company’s story:
Entering its fourth month of service, Iridium had 6,009 subscribers and $ 535,000 in accrued revenue. To put this in perspective, the revenue target in the current business plan, which had been revised downward, was $ 50.467 million, so the company was at 1 percent of that number. Among “cellular home subscribers”—the international business traveler—actual revenues were $ 4,000 against target revenues of $ 8.571 million, or 4/100ths of 1 percent of the goal.
On Wednesday, February 10, Roy Grant issued the certificate [to Chase regarding the loan covenants]. He’d decided to do things the easy way. “Roy Grant was a very competent financial guy,” said Leo Mondale, “but he had difficulty standing up to Staiano. Ed Staiano was a very difficult guy to disagree with.” Iridium would meet all its first-quarter targets, Grant told Chase. There would be about $ 4.3 million in revenues for the first quarter. And the company would have 355,400 voice subscribers by March 31. Grant had failed to mention several things. He failed to mention that the company had only 6,009 subscribers. He failed to mention that there were only 38,000 handsets available for those projected 355,400 subscribers. And he failed to mention that Iridium’s accrued revenue was $ 535,000, which was far less than the $ 700,000 that Grant would deposit into his personal bank account the same day he issued the certificate, which was the day he sold all his Iridium stock.
At the monthly Iridium board meeting on March 19, Roy Grant was nowhere to be found. He had left the company, citing “personal reasons,” and was replaced by Leo Mondale.
Motorola was prospering:
Meanwhile, across the ocean at the largest fairground in the world—the site of the largest computer trade fair in the world—Motorola was introducing a new phone called the Timeport. It was digital, it was small, it used the GSM standard, and it worked on every continent of the world except Antarctica. It had been developed by the Schaumburg guys who had coined the term “Iridiots.”
A week later the Pentagon issued a news release that should have been good news for Iridium—a $ 219 million contract to buy Iridium phones, pagers, accessories, and minutes. Unfortunately, the fine print indicated that almost all the money would go to Motorola, not Iridium, through a deal with Motorola’s wholly owned service-provider subsidiary in Chandler. Later that summer, the State Department would spend another $ 1.4 million for one thousand handsets, enough to supply America’s 260 embassies and consulates—but again, all the money would go to Motorola.
At the behest of the enraged banks, the CEO was replaced with John Richardson:
The answer, according to the gateway owners, who could now elect anyone they wanted, was John Richardson, the fifty-five-year-old Australian who had been at Barclays Bank in Hong Kong before being hired by the Saudis to run the African gateway in Capetown. … Randy Brouckman, who was essentially running the day-to-day operations at Iridium, was equally uncharitable: “John Richardson had the lifestyle of an upper-class Brit, the casual attitude of an Aussie, and the arrogance of a South African white male.”
Publicly Richardson remained blithely optimistic and dismissive of what he considered Motorola’s silly predictions of imminent disaster. He told the Wall Street Journal in late July, “What [Motorola] said about the alternatives [of liquidation and bankruptcy] was, from an academic perspective, absolutely correct, but from our perspective, it’s simply not on the radar screen.”
Bankruptcy may not have been on Richardson’s radar screen—perhaps because he was on safari in Tanzania during much of July—and bankruptcy may not have been culturally acceptable to the foreign gateways, but apparently it was pretty okay with the bondholders.
They tried to get Motorola to give them a break on the monthly fees for operation and maintenance, but Motorola wouldn’t yield, so Iridium filed for Chapter 11 protection on August 13, 1999.
It is a great story about technology, doing business in a government-regulated world, and just how much perseverance and how many connections one needs to succeed. I’ll write more about this book, but I recommend reading it!
More: read Eccentric Orbits: The Iridium Story
Philip Greenspun’s Weblog