Satellite Image Data is a Growth Business for Amazon Web Services

Satellites generate a lot of data, and Amazon Web Services (AWS) is providing a delivery and storage platform for new applications from its satellite data customers.

AWS highlighted its Space Accelerator portfolio in Washington DC last week during the company’s annual DC Summit, showcasing new services from satellite operators and data analytics companies.

Once viewed as a rare and expensive commodity only available through governments, earth observation imagery is now moving into mainstream availability with commercial companies such as Satellogic driving down pricing and increasing access through a cloud-centered e-commerce model.

“Strategic imagery was the priority of the day,” said Clint Crosier, Director, Aerospace and Satellite Solutions, AWS. “The space industry was born largely out of the need for governments to have strategic intelligence about adversaries and potential adversaries. That’s where the space game began, and there was a classic conundrum. If you wanted high-resolution imagery, you’d have to build a very exquisite satellite that would be very, very expensive and therefore you could only afford a handful of them and so you’d only be able to revisit an area on the globe once a day, or maybe once every three days.”

Satellogic and other “New Space” service providers are leveraging lower-cost launch services, mass producing lower-cost spacecraft and launching lots of them to generate drastically more imagery at lower cost, processing, storing and analyzing the terabytes of data on AWS with final products available on an AWS-hosted web site or shipped directly to a customer’s AWS presence for rapid availability.

“Fast forward and what Satellogic has been able to do leveraging the AWS cloud is bring (high-resolution images and high revisit rates) together and they’re actually complementary, because they’ve been able to leverage all the ways the cloud can reduce costs,” Crosier said. “They can build a cheaper satellite, therefore (customers can) afford to proliferate more of them into orbit, driving a higher revisit rate. So no longer do you have to choose between high resolution or high revisit rate – you can add them both now ,and that is a game changing capability.”

More Satellites on the Way, Bringing More Data

Closier knows a bit about game changing capabilities from his previous career as the United States Air Force general directly responsible for planning and implementing Space Force from scratch. Cloud services enable satellite Big Data companies to steadily scale services as they place more satellites – more data collection points – into orbit.

A single Satellogic satellite generates an average of 50 gigabytes (GB) of data per day, high-resolution 70 centimeter imagery with the ability to capture up to 60 seconds of full motion video. The company’s current constellation of 17 satellites provides up to four daily revisits of any point of interest on the Earth’s surface, but that’s just the beginning of a business roadmap to put more than 300 satellites into orbit by 2025.

Collecting around 850GB per day with 17 satellites, Satellogic and Amazon AWS will grow to handle 15 TB per day of unprocessed imagery from the 300 satellite constellation expected to be in orbit by 2025. Additional storage and processing growth will come from AI/ML-driven analytics processes to spot changes over time or detect and measure increased activity in an area of interest.

A Satellogic image of farmland in Brazi. Space imagery is becoming a bigger business for cloud players like AWS. (Image: Satellogic)

Ease of access and the massive quantity of satellite imagery is expected to create new products far beyond today’s corporate and government applications.

“The real customer market we see is (unlocking) untold amounts of use cases to the broader enterprise market and small businesses, state and local and provincial markets around the world,” said Matthew Tirman, President of Satellogic North America. “There are markets and use cases that we haven’t even thought of yet. That’s the real exciting piece.”

More Imagery Means More Data Crunching and Analysis

Data analytics companies Descartes Labs and Ursa Space systems, also AWS Space Accelerator participants, are leading the charge to leverage the Big Data flow from space. The two companies ingest terabytes of satellite data per day from multiple sources, including optical imaging satellites, radar satellites, RF monitoring satellites, and AIS/ADS-B ship and aircraft tracking services. Those data streams go into AI/ML models for analysis, generating insights for government and enterprise customers.

Descartes Labs (DL), based in Santa Fe, New Mexico, uses terabytes per day of satellite imagery to provide insights into agriculture, consumer goods, and mining and metals. Founded in 2014, the company goes far beyond simple storage and processing tasks when it starts crunching data. Descartes Labs has demonstrated how to use off-the-shelf AWS services as a supercomputer, winning places on the Top 500 list in 2019 and 2021.

“New Space” service providers are leveraging lower-cost launch services, mass producing lower-cost spacecraft and launching lots of them to generate drastically more imagery at lower cost, processing, storing and analyzing the terabytes of data on AWS.

“What we’re doing at this time is migrating our federal government work to AWS and we’re going to continue to migrate other critical workloads in the coming months,” said Alex Diamond, Head of Marketing at Descartes Labs. “We’ve used other cloud providers, and have been pretty much a cloud provider from the start. The beauty of AWS is that you don’t have to spin up the exact same computer. You can spin up what you need and when you need it. Any sort of tightly coupled application like weather simulation… even more loosely coupled applications that still require cloud computing expertise.”

Diamond estimates the company’s current storage archive is in the range of 15 to 20 petabytes, with AWS providing the combination of flexibility and large-scale computation power DL needs based on the types of applications required. The company’s analysis work includes building local weather models for agricultural production and forecasting how the impact of weather will affect commodity flows and yields for foodstuffs like palm oil, soybean, and sugar, with factors such as energy and shipping prices factored in.

Cost Improvements Drive Long-Term Growth

Ursa Space Systems, located in Ithaca, New York, has been working closely with Amazon for several years and is now expanding the relationship, seeing long-term opportunities as satellite imaging data grows and becomes more affordbale.

“From a ground perspective, we’re trying to accelerate the time in which we’re able to deliver our analytics to our customers,” said Nicole Robinson, President of Ursa Space Systems. “As of two weeks of weeks ago we’ve made our ready-made data sets available on AWS Marketplace. It’s kind of an upstream and downstream partnership on the tech side and the market side.”

A yearly subscription to one of Ursa Space’s ready-made data sets detailing global oil storage, auto manufacturing, or port monitoring is around $50,000 a year, with more detailed services monitoring areas of interest and customized and tailored solutions available and costing significantly more.

At a recent industry conference, Robinson described an automated oil pipeline monitoring solution that Ursa Space created using multiple sources of satellite data, with AI/ML trigging follow-up actions. RF information collected by a Hawkeye 360 satellite cluster would be analyzed for hand-held radio activity around a pipeline. Geolocated RF activity – radio traffic — would be used to “tip and queue” optical and synthetic aperture radar satellites to collect imagery in the vicinity around the area of interest.

Once the imagery arrives, an AI process looks for a barge or ship close by the area since pirates need a way to move stolen oil. In addition, satellite-collected AIS ship tracking information is used to see if watercraft in the area are broadcasting identification as required by international law or if there are “dark ships” that have turned off their beacons. If there’s a dark ship and unexpected radio traffic in a remote area next to a pipeline, odds are good that it’s a piracy operation.

Ursa see the combination of more satellite imagery coming into the market and cloud power being able to open larger downmarket opportunities as pricing continues to go down.

“We do actually very much want to crack the code and the e-commerce strategy where we’re able to drive down the price point per dataset just by having maximum scale,” Robinson said. “We want to make it usable, actionable by anyone sitting at home with curiosity about something going on in the world and they can pull up a credit card, make that purchase, and not have to refinance their house.”