by Eamon Javers
The first steps were to notify the government, the client expecting to take delivery of that day’s fresh images; and then to notify the company’s licensing authority, the National Oceanic and Atmospheric Administration in suburban Silver Spring, Maryland, which granted GeoEye—a private-sector company—the right to operate a satellite in the first place. Calls also went out to the National Geospatial Intelligence Agency, and to GeoEye’s customers around the world.
By the end of the day, GeoEye had a diagnosis. The problem was with the camera. And the camera was on the satellite, zipping around the edge of the Earth’s atmosphere. It had been designed with “single string” components, so there were no backup systems on board. There was no way to fix it.
This was a serious blow to GeoEye, whose stock was traded on the NASDAQ exchange. Wall Street securities regulations meant the company was duty-bound to disclose the event to the investing public, as it did in an 8-K filing with the SEC on Thursday, March 8. The document said the company could control the satellite, but couldn’t get pictures from it. GeoEye couldn’t tell the markets whether the satellite could be fixed, or when its fate might be known. In the meantime, GeoEye would try to service its customers with an older satellite, IKONOS.
Investors were startled by the news, and GeoEye’s stock fell sharply, dropping from $18.28 on Thursday to $16.25 by the time the markets closed at 4 P.M. Friday. It had been a lousy week for Bobby Ferraro, but such is life in the high-stakes satellite game.
Ferraro and his team spent much of the next month trying to think up a way to fix the camera. They worked on the problem over and over again until late April. Then they finally gave up.*
WHEN THE SATELLITE OrbView-3 went down, U.S. intelligence agencies weren’t the only ones inconvenienced. GeoEye, which has more than 400 employees worldwide and generated more than $180 million in revenue in 2007, estimates that almost half of its business comes from the private sector. Companies use GeoEye’s three satellites for all kinds of monitoring.
Oil companies can check on the status of their rigs in the Gulf of Mexico. Agricultural giants can generate false-color images of fields, in which red areas show healthy growth and yellowish areas show crops that need help. Developers can generate topographical maps of real estate they might want to buy. Google buys images for use in its popular satellite application GoogleEarth.
GeoEye also sells satellite images of the ocean to fishermen. The pictures identify heavy underwater concentrations of phytoplankton, where the fish will go to feed. Every day, GeoEye’s technicians e-mail satellite maps of these places to ship captains on bridges of trawlers around the world. It’s hardly sporting, but satellite images can save a ship’s captain as much as 10 or 15 percent in fuel costs. Instead of wandering around the ocean hoping to bump into a rich school of fish, the captain steers directly for the best spots.
The commercial satellite industry is still small. GeoEye has only one American competitor: DigitalGlobe, which is based in the Colorado Rockies. DigitalGlobe flies a QuickBird spacecraft that was launched into orbit 450 kilometers above the surface of the Earth in 2001 and captures 75 million square kilometers of imagery data each year.
Overseas, there are several more image providers, including the French company Spot Image, which boasts that its FORMOSAT-2 is the only high-resolution satellite that can take pictures of the same location each day—other satellite companies must wait days before the orbits of their spacecraft bring them around again to reshoot a given location.
But it is the American company GeoEye’s newest satellite that demonstrates the astonishing overlap between government and corporate intelligence technology. It’s not just that the private sector has access to the same types of satellite technology as U.S. spy agencies. In this case, companies will be using the same satellite as some of the nation’s most sophisticated intelligence operatives.
GeoEye participates in a program called NextView run by the federal government’s spy satellite operator, the National Geospatial-Intelligence Agency (NGA). The program is designed to allow the federal government to pick up much of the cost of developing the next generation of spy satellites, which cost millions to design, build, and launch. Under the NextView program, GeoEye developed a new satellite, and the NGA kicked in $237 million federal tax dollars to help build it. The contract allows GeoEye to use the satellite for its commercial clients, too.
Here’s what GeoEye said in its May 2008 quarterly filing with the Securities and Exchange Commission (SEC): “The Company anticipates that NGA will account for approximately half of the satellites’ imagery-taking capacity during this time, with the remaining capacity available to generate commercial sales, including sales to international ground station customers and municipal customers.”1
The company launched the new satellite, GeoEye-1, from Vandenberg Air Force Base in California in late August 2008. Boeing rockets were used. Technicians from General Dynamics prepped the satellite. GeoEye-1 spends much of its time taking pictures for the CIA and the Department of Defense, and the rest taking pictures for paying clients, which have included Wal-Mart, commodities traders, and commercial fishermen.
GeoEye’s corporate spokesman is Mark Brender, an amiable former ABC News Pentagon producer. Sitting in a conference room at GeoEye’s headquarters, he flashes a picture on a display screen of the first spy photo ever taken. It is a grainy image taken by the highly classified spy satellite Corona on August 18, 1960. You can barely make out an airstrip at a base called Mys Shmidta on the Far East coast of what was then the Soviet Union. A parking area is also visible, but that’s about all.2 Still, the image was revolutionary at the time. Military leaders could now see what the enemy was building, and where. Piecing together the observed activities, military officers could infer the Soviet Union’s overall strategy. Brender flashes more images on the screen. Today’s commercial satellites can see individual people walking on the ground. They can see cars well enough to pick out the make and model. They can see coral reefs underwater in the ocean.
“For forty or fifty years, the intelligence community kept their overhead intelligence capabilities highly secret,” Brender says. “This technology was the family jewels of U.S and Soviet intelligence. It was developed in order for two cold war rivals to be able to watch each other very carefully.”
The doctrine of the time, Brender points out, was called mutually assured destruction. Neither side wanted to attack first, because the other side had the ability to strike back with devastating consequences.
Now, though, the technology is in corporate hands. “Now we’ve moved into an era of mutually assured observation,” Brender says. “Governments used this technology to better understand the capabilities of the enemy. There’s no reason companies can’t use this technology to better understand the capabilities of a competitor.”
GeoEye plans to launch a fourth satellite in 2011. GeoEye’s IKONOS satellite, launched in 1999, orbits the Earth from the north pole to the south pole and back again every ninety-eight minutes—“It’s moving fast,” says Brender, about 17,000 miles per hour. That’s about four miles every second. Because the Earth rotates underneath the satellites as they orbit at an altitude of more than 680 kilometers above its surface, each lap around the planet takes place a little farther to the east. As a result, GeoEye’s satellites can see any spot on the planet once every three days.
The flagship spacecraft, GeoEye-1, is two stories tall; in orbit, this whole satellite can tilt, pivoting about fifty degrees in any direction to shoot specific targets as it whizzes by. It makes twelve or thirteen orbits each day. It maintains what’s known as a “sun-synchronous” orbit: that is, it can pass over a given area at 10:30 A.M. local time every day. In a single pass, it can capture two images of the same target from different points in space, and so it has the ability to create three-dimensional pictures.* It will be able to capture images of up to 700,000 square kilometers per day, and more than 225 million square kilometers per year. It is expected to la
st for ten years, but as GeoEye found out with OrbView-3, this doesn’t necessarily mean it will.
To download pictures from satellites in orbit, GeoEye maintains ground stations with satellite receivers in places as remote as Barrow, Alaska, and Tromso, Norway. It maintains an unmanned station at the Troll research station on the antarctic ice sheet. In this regard, GeoEye is better positioned than the U.S. government. As the result of international treaties, the military and intelligence agencies aren’t allowed to build their own ground stations at the bottom of the world. But GeoEye can.
Despite its advanced capabilities, GeoEye doesn’t know how its customers are using the images they buy. Brender says clients simply give GeoEye the location they’re after, and the satellite snaps the picture. The provider doesn’t know why the client wants the picture, or who the client is snooping on. All the company gets is the location. And that’s all it needs: whether the client is spying on a competitor or looking at its own assets is irrelevant to GeoEye—the company gets paid either way. To GeoEye, every location is the same. “It’s all Earth,” says Brender with a shrug.
He points to a satellite image of a strip mine in West Virginia. The mining company has chopped off the entire top of a mountain, clearing hundreds of feet of earth to get at the precious coal beneath. A picture of the mine’s progress would be somewhat helpful for the mining company, but the executives there already know how the dig is going—they walk the site every day.
The image would be truly helpful, Brender says, to the mining company’s competitor. Mining bosses at a rival company don’t have access to the site. “If you were the competitor to that mining company, you might be able to watch that mine to see how much progress they’re making,” says Brender. The competitor could figure out how much coal had been found, how much it would cost to recover the environmental damage done by the massive excavators, and how much profit would be generated. With a sufficiently knowledgeable observer, you might be able to get a pretty good sense of whether or not the mining company would hit Wall Street’s estimates of its quarterly earnings.
ONE CLIENT OF GeoEye is Lanworth, a thirty-person firm in Itasca, Illinois, about forty minutes east of Chicago.* This company uses images from GeoEye and other satellite providers to beat the U.S. Department of Agriculture at its own game. To understand how, you need to know a bit about how the commodities market is closely linked with the government.
Each week, the Department of Agriculture (USDA) releases several reports on the prospects for different types of farm products, such as soybeans, chicken, eggs, and farm-raised catfish. The reports detail prices, harvesting progress, inventory, and other information that goes into calculating the prices of commodities futures on Wall Street. (Commodities futures are simply contracts to buy or sell a given product.)
In a hilarious movie from 1983 starring Eddie Murphy and Dan Aykroyd, Trading Places, the climactic scene involved Murphy’s and Aykroyd’s characters intercepting a crop report on orange juice in an attempt to corner the futures market on Wall Street.*
Traders buy and sell futures all day long, without ever touching an actual farm product. Much like trading stocks, the game is all about predicting the future price of the contract. If you think it’s going up, buy. If you think it’s going down, sell. Or you can short a commodity, selling its futures contracts at a high price now, and then buying them back later when the price is low. You pocket the difference as profit. It’s not easy: if you guess wrong and prices increase, you’re on the hook.
In the movie, Aykroyd’s character explains the drama: “One minute you’re up half a million in soybeans and the next, boom, your kids don’t go to college and they’ve repossessed your Bentley.”3 The movie’s heroes get an advance look at the USDA crop report on oranges. With that information, they know which way the market will move and position themselves to make a fortune and wipe out their rivals.
In real life, people are trying to do something similar. They, too, want to be able to send their kids to college and buy a Bentley. Because the USDA crop reports are a key element in moving the commodities markets up and down, the ability to predict what’s in the reports has become an enticing prospect. And now that satellite technology is available to anyone, the government doesn’t have as much of an advantage in predicting crop yields as it used to.
Lanworth uses satellite data to put together reports on the prospects for various agricultural products. Using satellite imagery of the fields, weather data, and other information that the USDA forecasters themselves use, Lanworth works up predictions of what the government will say in its reports. The company sells those predictions to commodities traders desperate for any information on what the market might do next.
Lanworth’s cofounder Shailu Verma says his customers include giant agricultural companies, real estate investment companies, hedge funds, and financial firms. “It’s not our objective to steal USDA’s thunder,” Verma says. “But if we are good, we’re generating the same information as the USDA, just a little bit earlier.”
Lanworth reportedly charges $100,000 per year for its reports,4 but the information those documents provide can be worth millions in the commodity markets. “Our customers are trying to figure out if there’s a deviation between what Lanworth is telling them and what the market expects. If there is, that’s a monetizing opportunity,” says Verma. He, too, is a fan of the movie Trading Places: he calls Lanworth’s business plan “the Eddie Murphy strategy.”
Here’s how it works in the real world.
In June 2008, Lanworth gathered all the data it could on the corn crop in the United States. Disastrous flooding in Iowa that month had caused more than $200 million in damage, and had thrown into question the size of the nation’s corn crop for the entire year (since Iowa normally produces a large proportion of that crop). The USDA’s annual report on crop acreage was due to be issued on Monday, June 30, at 8:30 A.M.
As the report was being prepared, the mid-June floods threw the government’s counting process into disarray, and the diligent bureaucrats at USDA went back to reinterview 1,150 farmers to find out how much the floods had affected the crops. But while the government was working its way through that tedious process, Lanworth’s analysts were poring over images from satellites high overhead. The company uses relatively low-resolution Japanese, Indian, and U.S. government satellites, and calls in high-resolution, more expensive GeoEye and DigitalGlobe images for spot checks.
What Lanworth found was surprising. Despite the flooding, the corn crop appeared to be in good shape. Some areas that had been flooded weren’t a total loss. And although the markets were bidding up the price of corn futures—traders expected corn prices to go up as supply went down after the flood—Verma and his team knew better. The price should come down from where it was, because the corn supply wasn’t damaged as badly as most traders in the market assumed. Corn futures, which are contracts to purchase 5,000 bushels of corn, were trading at about $7.50. Verma and his team reported their findings to their clients, who made new bets in the market.
On Monday morning, right on schedule, the USDA put out its report. The first sentence of this press release was all most traders needed to know: “Despite the recent flooding in the Midwest, U.S. farmers expect to harvest nearly 79 million acres of corn,” said the USDA release.5 Prices began to fall. The government report meant that corn futures were overvalued, and traders rushed to dump them as fast as possible. Ultimately, corn futures dropped nearly $2 in value. But Verma’s hedge fund clients were well positioned to profit from the fall. They didn’t tell Verma exactly how much they earned—“These guys are very cagey,” he says—but such a sharp price collapse could have been worth millions to a trader who made a big enough bet on it.
It turns out that Iowa is not the best place to use satellite technology, despite Lanworth’s success there in 2008. After all, the state is readily accessible, and the USDA already does a rigorous job of reporting what’s going on there. Satellite technology comes
in most handy to measure crops overseas, where supply affects global prices but governments don’t do as good a job reporting the details as the USDA does. “In India, the question always is, How correct are the government estimates?” says Verma. “Are there political pressures that make them say one thing and do another?”
With a satellite, it doesn’t matter what kind of pressure the local bureaucrats are under. Verma says Lanworth makes some of its most valuable predictions on crop yields in India, China, Kazakhstan, and Russia.
And the company is moving on to other sectors. Verma is looking at ways to help the insurance industry. He’s held several meetings with investment banks hoping to use satellite surveillance to spy on Chinese factories. With a good enough picture, Verma says, his team could count the number of trucks going into or out of any factory in China, and get a good sense of that facility’s production rates. Those rates will affect the costs of products in the United States and the stock prices of any number of companies around the world. Verma predicts that Lanworth will be selling such analyses soon.
IN MANY CASES though, companies aren’t spying into foreign countries, or even looking at competitors. They’re spying on themselves. Take Wal-Mart, the world’s largest retailer. This company hired GeoEye’s predecessor several years ago to take satellite pictures of its stores. Wal-Mart knew where its best-performing stores were, but it wanted to know why they were doing so well. What made these stores so profitable?
The satellite company took pictures of the locations and analyzed the images.
GeoEye recruits many of its employees from the military and intelligence agencies. Walking down the halls at any time, you can bump into veterans who were trained in the military in battle damage assessment (BDA). The same skills they used to decipher images of blown-up Iraqi military hardware can be used to interpret pictures of Wal-Mart’s parking lots.