Planet Uses Satellites and AI for Geospatial Intelligence
A picture is worth a thousand words. That’s one platitude that hasn’t aged well in the age of the selfie. Everywhere you turn – ancient temples, the beach, cocktail bars – people are taking crappy, crooked photos of their misshapen heads, flashing gang signs in a dyslexic version of sign language. More than 250 people have died over the last six years while shooting selfies from the top of a cliff, in front of an oncoming train, and while posing with an injured bear. The phenomenon even extends to planet Earth. Nearly 30 years have passed since Voyager shot the famous Pale Blue Dot photograph that captured our world in one tiny pixel. Now, with thousands of satellites buzzing around the Earth, we produce planetary selfies every day. A nearly decade-old startup that operates the world’s largest satellite fleet for geospatial intelligence is doing much more than taking pretty pictures.
What is Geospatial Intelligence?
Planet, formerly known as Planet Labs, currently has about 150 active satellites collecting imagery of the Earth. That’s by far the biggest fleet of satellites owned and operated by one company – at least until SpaceX begins to launch its satellite constellation Starlink this year to enable global internet. All of those eyes in the sky are focused on the goal of producing geospatial intelligence, a topic that we covered in our list of eight satellite data startups.
The term geospatial intelligence itself is only about 15 years old and actually came into existence during a series of bureaucratic whiteboarding sessions to rename the National Imagery and Mapping Agency (NIMA). NIMA had been created in the 1990s to integrate the United States’ various agencies involved in remote sensing and imagery analysis with mapping and charting. What emerged was the National Geospatial-Intelligence Agency (NGA). That bit of history comes courtesy of the U.S. Geospatial Intelligence Foundation, a nonprofit educational foundation that supports what it calls GEOINT, because everything in the government must be made into a cool-sounding acronym.
USGIF loosely defines geospatial intelligence “as the derivation of information from imagery, geospatial data in all forms, and analytics.” The GEOINT revolution encompasses the “commoditization of commercial remote sensing, powerful analytic software, ubiquitous precision geolocation data, far-reaching broadband, rapidly advancing processing power, affordable cloud storage, and advanced analytics and visualization.” In other words, geospatial intelligence turns space-based imagery into actionable data. And it’s becoming big business, with one estimate saying the geospatial analytics industry is already chugging along at $40 billion.
Planet Enters Geospatial (Intelligence) Orbit
San Francisco-based Planet hit the scene in 2010 when a trio of NASA scientists founded the startup, working out of a garage to build the company’s first satellite. We briefly profiled Planet in our list of small satellite startups. The company has raised about $183 million in disclosed funding, according to Crunchbase. However, we’ve found other estimates, such as at SpaceNews, which puts the amount at about $300 million, while EquityZen pegs total funding at $452 million. A valuation is also hard to nail down, but most sources agree that it exceeds $1 billion, with PitchBook giving Planet a value of more than $1.1 billion back in 2015, though CB Insights appears to snub the startup by not including it on the data firm’s unicorn list of companies valued over $1 billion. In MBA speak: Planet has raised a crap ton of money and is worth a pretty penny.
Planet sent its first satellite into orbit back in 2013, a nanosatellite called Dove. The Doves weigh only about 11 pounds with three-meter resolution, comprising the core of the company’s fleet. Just four years later in 2017, Planet sent a record 88 Dove satellites into orbit aboard one rocket. That was the same year that the company reached its goal of imaging the entire planet once per day.
Planet also operates two other constellations of larger satellites it inherited through acquisitions. In 2015, Planet purchased Berlin-based BlackBridge and its five-satellite RapidEye constellation. A couple of years later, Planet got its hands on Google’s satellite imaging business, Terra Bella, which the tech giant had originally acquired for $500 million in 2014. The deal included the high-resolution SkySat constellation of satellites and a multi-year contract with Google to purchase Earth imagery from its former assets. Google reportedly also made an investment in Planet at the time.
A World of Services
Planet is not merely a purveyor of pictures. In fact, it doesn’t even consider itself primarily a satellite company but a data services company first. Its recent acquisition last month of Boundless Spatial out of St. Louis highlights that fact. The startup, which had raised about $10.3 million, creates software and tools for customers to apply geospatial data in various applications. Boundless had scored a $36 million contract with NGA back in 2017. NGA is also a major customer for Planet, so there were some obvious synergies with that acquisition.
The company’s data services extend well beyond government work, however, to include agriculture, forestry, insurance, emergency management, energy and infrastructure, and more. In agriculture, for instance, Planet’s stream of daily imagery can help farmers manage crop health day by day. One of the biggest-funded agtech startups out there today, Farmers Edge, is a customer. Farmers Edge provides its clients with a farm management platform that includes satellite data analytics from Planet. Another customer using Planet imagery is HARA, the Indonesian blockchain/agtech startup we profiled in our article on Indonesia’s Big “Big Data” Problem.
In other cases, insurance companies can now spot fraud from 20,000 or so miles in space thanks to high-resolution imagery, as well as assess damage from natural disasters more accurately. And a treasure trove of archival imagery, combined with real-time updates, can help insurers model risk. That means they can spread the risk involved in covering millionaires who build in fire zones and floodplains to everyone else in the form of higher premiums. Meanwhile, companies with infrastructure like oil pipelines can monitor their assets and identify when the jihadists have sabotaged the supply. And so on.
Putting AI into Geospatial Intelligence
To serve these various markets, Planet has developed an array of products and tools, but the most exciting one from our perspective is Planet Analytics, which the company says moves the need from basic imagery to intelligent insights using artificial intelligence. This suite of products “leverages machine learning to transform global, daily satellite imagery into information feeds that detect and classify objects, identify geographic features, and monitor change over time.”
Last year, the company rolled out a beta version of Planet Analytics with five new analytic streams based on what it calls Areas Of Interest (AOI): building detection, road network detection, aircraft identification, maritime vessel identification, and deforestation detection. Planet expects its new product to improve exponentially over time. That’s because it produced about six terabytes of global data every day to feed the machine learning algorithms to make them better.
The bigger picture is that Planet anticipates its new analytics platform will make it possible to not only observe and understand changes, but predict them as well. Planet Analytics will also enable a related initiative – Queryable Earth. The concept is to index physical changes on the planet and make it as searchable as a Google query. You will be able to answer questions like “What’s the daily average number of aircraft not taking off from LAX?” or “How many Starbucks were built in New York City in the last five minutes?”
If AI is the new electricity, data is the coal (or, for the green-minded, solar power) that drives those insights. In that context, Planet owns the whole utility – a vertically integrated company that extends more than 20,000 miles from the Earth’s surface. SpaceNews recently did a deep dive into the startup’s subscription-based business model, noting that Planet’s recent deals revolve less around providing imagery and more about delivering data for different information feeds. The use of machine learning to provide context and insights around that satellite-generated data – a paradigm that mirrors what a number of AI companies delivering enterprise solutions are doing – is the next obvious evolution. The world is changing, and Planet claims it can see what’s coming next. How much will that be worth someday?
If you enjoyed this article, then sign up for our free newsletter - Nanalyze Weekly. About every week, we'll send you a simple summary of all our new articles. If you didn't enjoy this article, share it on Twitter and tell everyone how much you hated it.