At the beginning of 2012, a new startup moved in to the ground floor of the building across from the Clock Tower in San Francisco’s South of Market neighborhood. A cryptic, index-card-size sign appeared to the right of their smoky glass door, which I walked past every morning. “Cosmogia,” it read. An amalgam of cosmos and logia, was my best guess. Cosmos: universe, obviously. Logia: communications of divine origin.
My office was on the floor above them, and a tinted glass atrium penetrated both floors, so I could glimpse what was happening down there near their windows, at a very obtuse angle, especially after dark when their lights were on. It looked like a hardware lab, with unmistakable engineering workbenches and clear, softwall strip curtains—clues that they’d made a low-grade pressurized clean room to purify particulates from the air. This was intriguing; in 2012, serious hardware startups in SoMa were rare.
I first met Robbie Schingler and the Cosmogia team that April, when a burglar got into the building from the roof door, broke a lot of glass, and took a few laptops, and everyone checked on one another to try to figure out what had happened. Schingler told me that his company was making little solar-powered satellites—a type called a CubeSat—that would be deployed as low-cost, secondary payload on rockets. It was outfitting them with parts you could find on $300 smartphones, radically bringing down the cost of these starcraft from as much as $30,000. Today, this type of satellite can be built for as little as $7,500—or less, depending on its configuration.
This would be the moment when anyone with a single SpaceGeek gene in their body would have freaked out and melted in euphoria. I’ve got friends who have that gene, and I’ve got friends at NASA, too, who have it in spades. But alas, I have zero SpaceGeek in me. I was just glad Cosmogia hadn’t been robbed.
A couple of years later, Schingler took me into their clean room and showed me how they were assembling their next-generation spacecraft, called a Dove 2, right there on the ground floor of that building across from the Clock Tower. By then, they’d ditched the name Cosmogia and become Planet Labs, and then simply Planet. They had deployed a few dozen satellites in a ringed fleet that circled our planet. Every day, Earth rotated underneath these flying cameras. Planet’s goal was to take a picture of every spot of land on Earth, every single day, and sell them.
Now I was quite fascinated. To be honest, if you strolled down any South of Market street and passed one of these Doves on the sidewalk, you’d barely notice it. Especially with the solar wings tucked in—it’s a metal brick smaller than a shoebox with the antennae sticking out like a tongue depressor. You’d think it was a motor rack set that had fallen out of a used drying machine. Or a server fan that had fallen off the back of an electronics-recycling truck.
But I understood instantly: this modest little gadget was an ethical black box, gargantuan in scope.
Planet had made it both massively cheaper to be Big Brother and massively easier to catch evildoing in the act. They’d made it massively easier to spy on industrial competitors, and they’d made it massively cheaper to monitor deforestation and farmland crop health.
All technology is an ethical black box. “Technology is neither good nor bad; nor is it neutral,” historian Melvin Kranzberg wrote three decades ago. The technologies that have the potential to do the most good for the world also have the potential for maximum evil. CRISPR gene editing, AI robotics, social media, nuclear fission, pesticides. Over three decades as a tech journalist, I’d come to the conclusion that only one thing determined how any company would use technology: the moral philosophy of its founders.
From their company’s earliest days, Schingler and cofounder Will Marshall had publicly declared that they’d formed it to use satellites to help humanity. They spoke about coral reef monitoring and overfishing and saving forests. And they seemed to be really good guys. Their pedigrees as NASA engineers suggested that they weren’t after money. But it was also quite obvious that to pay for all this do-gooderness, someone needed to pay Planet. Someone who could use the daily images to make loads of money. It was an ominous and inevitable equation.
I was rooting for Planet. But I wanted to see how they really turned out.
WRITTEN IN THE STARS
In the competition for Earth imagery, Planet had bet on image frequency. At least daily, if not multiple snaps a day—for every square foot of land on the marble.
Competitors bet in other dimensions. They had far fewer satellites but far more powerful optics in their cameras. One direction was image resolution; the pictures were prettier, with a higher pixel density, and the satellites could see through clouds. Heck, even Google Earth still makes Planet’s imagery look like Lego bricks if you’re looking at just a single image. (But the image Google Earth gives you is typically one to three years old.)
Competitors were also far better at capturing light rays outside the visible spectrum. Their cameras offered an extra band of green and five extra bands just beyond red—red edge, near infrared, and so on. These reflected wavelengths of light could see past the chlorophyll of a soybean leaf into its internal structure, picking up water, nutrients, and carbon. They could capture photosynthesis activity live. They could tell when almond hulls were about to crack, or when vineyard grapes were peaking in Brix—something you’d normally need a handheld refractometer device on the ground to measure.
Still other firms went after stereoscopic, three-dimensional images, synthesized from multiple photos of the same target, taken from different angles as the satellite passed over.
Make no mistake—Planet’s Doves were bare-bones, stripped-down hunks of metal compared with competitors’ fancy spacecraft. Their avionics were sorta primitive. They had no thrusters. Their solar panels doubled as wings to create drag during deployment. They merely rode Earth’s magnetic field until they fell from the sky and burned up. Inside the Doves were little electric motors that spun in tight circles, connected to nothing—but they generated just enough counterforce to aim the satellites’ cameras down at Earth. When they flew over the Atlantic Ocean and hit South America, the Doves would heat up on the side facing the sun, causing a thermal field that blurred their images. The Doves had almost no redundancy built in, violating a cardinal rule of aerospace engineering—because they were essentially disposable. The Doves were vulnerable to solar storms, cosmic rays, rocket shock waves, and thermal differentials. When any fell from orbit, Planet simply sent up new ones.
Seriously, there are high school kids who assemble CubeSats and send them to space as auxiliary payload on a rocket.
Anyone could see where it would all converge. Between 2015 and 2017, Planet and other low-orbit-satellite companies would compete heavily. Eventually, the one with the most valuable satellite constellation would emerge a winner and buy up the satellites of many of its competitors.
So who would win? Take a retrospective second to think this through. Some of the biggest investors in the world were making their bets, including Google, which scooped up Skybox Imaging and renamed it Terra Bella. Any of its seven high-resolution SkySats could shoot live, black-and-white, full-motion video of your backyard barbecue, if it happened to be overhead.
If Earth didn’t change a lot, then daily photos wouldn’t be meaningful. If the human eye mattered, then high-resolution 3-D images would win. If it was all about agriculture, then the invisible spectrum was where to put your money.
Guess what? You might have noticed, but the world is changing really fast. And with the implementation of AI image recognition and spatial mapping that turns images into data, the human eye’s perception of images no longer matters. Also, because its satellites were cheap to make, Planet could send up two dozen better ones every year.
Hedge funds didn’t want terrific, high-res photos of parking lots outside Walmart four days before Thanksgiving. They just wanted enough resolution for a computer to be able to count the cars on Black Friday. Hedge funds didn’t want an update a few times a month on how much oil Saudi Arabia was storing in reserve in its floating roof tanks—they wanted to know every day, and trade on it. Nobody wants yesterday’s news. Especially the agricultural industry, which needs to know which day to harvest. Not which month.
Planet’s bet on frequency paid off. It became the one with the most valuable satellite constellation. Even Google caved, and in 2017, it sold Planet its small fleet of high-resolution SkySats.
SEE SOMETHING, SAY SOMETHING
In 2017, Planet had enough Doves in the air to meet its target of photographing every bit of Earth’s landmass every single day. It was creating 1.2 million images a day. The algorithms it was developing would soon be able to look at any 100-square-foot picture and not only count houses, cars, roads, cattle troughs, and storage tanks, but also measure aridity and photosynthetic activity. It had beat out many of its competitors and was getting paid for its images. Planet was at that critical upward swing on Silicon Valley’s J-curve where the cool, inspiring, exciting company so frequently goes “corporate.” Fans of tech startups have seen the script play out so many times. A company we root for—because of its daring—suddenly matures and becomes a safe, industrial blob, ensconced in its network of financial industry and corporate partners. It gets boring, and prints money.
But Planet, somehow, didn’t turn into a blob.
Instead, the moral philosophy of Planet’s founders really emerged. Planet was woke.
The favorite tool of hedge funds and Big Food turned into…well, kind of a technological Batman. Yeah, that Batman—the Caped Crusader. Bruce Wayne, friend to industry, became the Dark Knight, looking out for the common people, catching bad guys red-handed, and monitoring the machinations of evil profiteering.
It seemed to start in the fall of 2017. All of a sudden, Planet started churning out news of global wrongdoing. North Korea was testing nuclear weapons in tunnels deep beneath Mount Mantap—Planet caught the landslides triggered by the explosions, which allowed experts at Johns Hopkins to estimate that North Korea’s weaponry was six times more powerful than it had been a year earlier. About a week later, Planet and Amnesty International were publicizing images of Muslim Rohingya villages being systematically burned to the ground in Myanmar.
Brad Adams, executive director of the Asian division of Human Rights Watch, told me, “The Rohingya were being slaughtered and raped. Villages were being burned down. But a lot of governments and diplomats were ‘knuckle draggers.’ They were so committed to Aung San Suu Kyi and a successful reform process that they were denying or challenging the facts on the ground. Showing Planet’s images ended their game.”
Planet then captured images of a Chinese aircraft carrier and air strike division joining a naval flotilla in the heavily contested South China Sea, where Beijing is bullying Vietnam, the Philippines, and Malaysia.
While the Facebook blob was being dragged before Congress to explain how it had all gone wrong, Planet was demonstrating how technology could still be a force for good. “We weren’t responding to what was happening to other tech companies,” Schingler explained when we caught up after all those years—shortly after he’d showed me the Dove 2 in the clean room, Planet had moved to another SoMa location. “We were just executing the vision we had all along. We started small, with pilot tests, and kept going.”
And the news just kept coming from nonprofit global watchdog groups and academic institutions using Planet’s images to both detect and visualize crookedry. A secret Iranian long-range missile facility was discovered, contradicting the claims of Iranian leaders. When President Trump began separating children from their parents at the southern border, Planet’s starcraft were above the detention camps in Tornillo, Texas, showing that while the administration was claiming that not many kids were being held there, rows upon rows of 20-person tents had in fact been set up to expand capacity.
When the Camp Fire in Northern California broke out at 6:33 a.m. on November 8, a Planet satellite was snapping photos within four hours, allowing emergency officials to spot burned and as-yet-untouched properties.
All of those were individual, discrete events. Then Planet unleashed a systematic approach. It was working with the Amazon Conservation Association to automatically spot new roads being bulldozed in national forests in Peru; if one of its Dove 2 satellites found something, engineers summoned one of the high-resolution satellites Planet had bought from Google to the location, to confirm the presence of logging trucks. In the Ivory Coast, a similar partnership helped target illegal cocoa plantations that were using boys brought in from Burkina Faso as slave labor. With the Paul G. Allen Philanthropies, Planet began building a global coral reef monitoring system that detects the nitrogen-fertilizer algae blooms and ocean acidity that harm coral. Then there’s Carbon Tracker, which measures the plumes from China’s massive coal plants to estimate their carbon dioxide emissions. The tracking of emerging missile systems in Iran and Syria is done in partnership with Middlebury College and its Institute of International Studies in Monterey, California.
If it seems like a lot, the vision for what’s to come is even bigger.
Planet is now getting a lot of national attention. As of last fall, it had launched more than 300 satellites, and about 140 of them were active. As I was writing this essay, the company was featured on 60 Minutes and in the New York Times. But what it’s really up to—its grand vision—is clearly not being told and, thus, is not understood.
Put your brain in a TED Talks frame of mind. Get your Davos/World Economic Forum thinking hat on.
It’s increasingly hard to see a rosy future.
We’re worried about not just climate change but, with melting permafrost releasing methane en masse, a massive extinction of human and animal species. As if that were not a big enough problem, at the same time we can see that artificial intelligence, robotics, and software are displacing human labor at such an alarming rate that in 10 years we may have more people not working than we have working. The world’s fishing fleets pull a trillion fish out of the oceans every year, and we are at our limit. But two billion more people are coming, and they need to eat. Cyberterrorism and online impersonation and data theft are growing exponentially. Nationalism and ethnocentrism feed off the fear of the future, the fear of what’s happening now—driving trade wars and the dissolution of long-standing multilateral agreements.
In Silicon Valley, we call this convergence. All these massive changes, all at once.
If there’s hope, it’s knowing that ProPublica’s tools are fighting back against Facebook’s invasion of our privacy. It’s watching a lone-wolf Twitter account, Sleeping Giants, destroy the advertising base that fuels Breitbart’s hate speech. It’s seeing Instagram’s Diet Madison Avenue hold so many Mad Men accountable for how they’ve mistreated women. It’s watching swarms of volunteers walk around the state of Pennsylvania pointing optical instruments at natural gas facilities to measure methane leaks (it turned out the state had a five-times-bigger methane problem than the industry had reported to regulators).
Planet is part of the social-good ecosystem, a diverse but thriving movement to use data for good. It offers a spark of hope, but if we want more than hope—if we want to actually fix our mess—this detection net needs to be vastly more robust and far more sweeping.
Take the international greenhouse gas accords, which will reward countries that lower their emissions and reduce deforestation. Planet’s images can be a tool to measure compliance.
“Under the 2016 United Nations Paris Agreement, 67 countries are up for very significant financial incentive payments,” explained Tara O’Shea, director of forest programs at Planet. “But measuring all of a country’s forest is hard, and at best we’ve had only annual measures. With our satellites and system, ecologists can easily calculate, automatically, a daily indicator of forest cover.”
A commodity-like index, of sorts, for what Planet calls “big indicators.” “If we can measure things that society values, we can use public policy to create financial incentives to change behavior,” said Schingler. Industries can be held accountable. Lawsuits against violators will have teeth. Governments can create far-less-arbitrary taxes or penalties.
“In the global economics system, there are a whole set of things that are externalities,” explained Andrew Zolli, VP of global impact initiatives at Planet. “They’re not counted, so they’re not valued. So humankind damages those systems.”
California is the single best place to develop this vision, said Schingler. Planet has partnered with the state to create policies around automated, real-time climate indexes. “Our legislature in Sacramento is one of the most forward-leaning in the world. So many of our companies, funded and scaled in California, are mission-driven. It’s a natural place to experiment with new approaches to solving challenges that fall between the responsibility of corporations and governments.” He’d like to help California monitor everything from methane leaks to forest dryness to water quality to lead emissions and more.
When I asked Schingler how this ecology-monitoring system would evolve over the next 15 years, he reminded me that we don’t have 15 years. “We have 12 years to live up to the challenges the U.N. gave us. Of that, it will take a half decade to get to real measuring and reporting systems for sustainable development goals.”
“We didn’t launch 300 satellites and point them at the stars,” said Zolli. “We pointed them back at Earth. Together, they are a kind of mirror, to see things we normally can’t see. A mirror is an incredibly powerful tool for moral discovery. It takes phenomena that are on the ginormous scale down to the range of our cognition—increasing salience, moral urgency, and drive to action.”