Forty years after humans first saw pictures of a blue and white marble taken from space, it’s remarkable how few new images of Earth we get to lay eyes on. Of the 1,000 or more satellites orbiting the planet at any given time, there are perhaps 100 that send back visual data. Only 12 of those send back high-resolution pictures (defined as an image in which each pixel represents a square meter or less of ground), and only nine of the 12 sell into the commercial space-based imaging market, currently estimated at $2.3 billion a year. Worse still, some 80 percent of that market is controlled by the US government, which maintains priority over all other buyers: If certain government agencies decide they want satellite time for themselves, they can simply demand it. Earlier this year, after the government cut its imaging budget, the market’s two biggest companies—DigitalGlobe and GeoEye, which between them operate five of the nine commercial geoimaging satellites—were forced to merge. Due to the paucity of satellites and to the government’s claim on their operations, ordering an image of a specific place on Earth can take days, weeks, even months.
Because so few images make their way down from space every day, and even fewer reach the eyes of the public—remember how dazzled we were when Google Earth first let us explore one high-definition image of the planet?—we can fool ourselves into thinking that the view from space barely changes. But even with the resolutions allowed by the government for commercial purposes, an orbiting satellite can clearly show individual cars and other objects that are just a few feet across. It can spot a FedEx truck crossing America or a white van driving through Beirut or Shanghai. Many of the most economically and environmentally significant actions that individuals and businesses carry out every day, from shipping goods to shopping at big-box retail outlets to cutting down trees to turning out our lights at night, register in one way or another on images taken from space. So, while Big Data companies scour the Internet and transaction records and other online sources to glean insight into consumer behavior and economic production around the world, an almost entirely untapped source of data—information that companies and governments sometimes try to keep secret—is hanging in the air right above us.
Here is the soaring vision that Skybox’s founders have sold the Valley: that kids from Stanford, using inexpensive consumer hardware, can ring Earth with constellations of imaging satellites that are dramatically cheaper to build and maintain than the models currently aloft. By blanketing the exosphere with its cameras, Skybox will quickly shake up the stodgy business (estimated to grow to $4 billion a year by 2018) of commercial space imaging. Even with six small satellites orbiting Earth, Skybox could provide practically real-time images of the same spot twice a day at a fraction of the current cost.
But over the long term, the company’s real payoff won’t be in the images Skybox sells. Instead, it will derive from the massive trove of unsold images that flow through its system every day—images that, when analyzed by computer vision or by low-paid humans, can be transmogrified into extremely useful, desirable, and valuable data. What kinds of data? One sunny afternoon on the company’s roof, I drank beers with the Skybox employees as they kicked around the following hypotheticals:
- The number of cars in the parking lot of every Walmart in America.
- The number of fuel tankers on the roads of the three fastest-growing economic zones in China.
- The size of the slag heaps outside the largest gold mines in southern Africa
- The rate at which the wattage along key stretches of the Ganges River is growing brighter.
Such bits of information are hardly trivial. They are digital gold dust, containing clues about the economic health of countries, industries, and individual businesses. (One company insider confided to me that they have already brainstormed entirely practical ways to estimate major economic indicators for any country, entirely based on satellite data.) The same process will yield even more direct insight into the revenues of a retail chain or a mining company or an electronics company, once you determine which of the trucks leaving their factories are shipping out goods or key components.
Plenty of people would want real-time access to that data—investors, environmentalists, activists, journalists—and no one currently has it, with the exception of certain nodes of the US government. Given that, the notion that Skybox could become a Google-scale business—or, as one guy on the roof that afternoon suggested to me, an insanely profitable hedge fund—is not at all far-fetched. All they need to do is put enough satellites into orbit, then get the image streams back to Earth and analyze them. Which is exactly what Skybox is planning to do.
The most important thing to understand about Skybox is that there is nothing wonderful or magical or even all that interesting about the technology—no shiny new solar-reflecting paint or radiation-proof self-regenerating microchip, not even a cool new way of beaming signals down from orbit. Dozens of very smart people work at Skybox, to be sure, but none of them are doing anything more than making incremental tweaks to existing devices and protocols, nearly all of which are in the public domain or can be purchased for reasonable amounts of money by anyone with a laptop and a credit card. There is nothing impressive about the satellites they are building until you step back to consider the way that they plan to link them, and how the resulting data can be used.
Such bits of information are hardly trivial. They are digital gold dust, containing clues about the economic health of countries, industries, and individual businesses. (One company insider confided to me that they have already brainstormed entirely practical ways to estimate major economic indicators for any country, entirely based on satellite data.) The same process will yield even more direct insight into the revenues of a retail chain or a mining company or an electronics company, once you determine which of the trucks leaving their factories are shipping out goods or key components.
Plenty of people would want real-time access to that data—investors, environmentalists, activists, journalists—and no one currently has it, with the exception of certain nodes of the US government. Given that, the notion that Skybox could become a Google-scale business—or, as one guy on the roof that afternoon suggested to me, an insanely profitable hedge fund—is not at all far-fetched. All they need to do is put enough satellites into orbit, then get the image streams back to Earth and analyze them. Which is exactly what Skybox is planning to do.
The most important thing to understand about Skybox is that there is nothing wonderful or magical or even all that interesting about the technology—no shiny new solar-reflecting paint or radiation-proof self-regenerating microchip, not even a cool new way of beaming signals down from orbit. Dozens of very smart people work at Skybox, to be sure, but none of them are doing anything more than making incremental tweaks to existing devices and protocols, nearly all of which are in the public domain or can be purchased for reasonable amounts of money by anyone with a laptop and a credit card. There is nothing impressive about the satellites they are building until you step back to consider the way that they plan to link them, and how the resulting data can be used.
by David Samuels, Wired | Read more:
