Orbital Insight Raised $8.7 Million To Use AI To Analyze Satellite Data

Orbital Insight, a company that provides data analytics of satellite imagery, announced this week that it had closed an $8.7 million Series A round led by Sequoia. Bloomberg Beta, Google Ventures, citizen.vc and Lux Capital also participated in the round. Sequoia partner Bill Coughran of Sequoia will be joining the board of Orbital Insight as part of the deal.
Orbital Insight was founded by Dr. James Crawford, an artificial intelligence researcher and entrepreneur who has experience building intelligent systems for NASA and other organizations. He was also previously the Engineering Director of Google Books, an experience that helped lead to the idea for his new company.

 

 

“What this company really is is a Google Books like pipeline – that is, an automated AI pipeline – designed for understanding and processing satellite imaging at scale,” Crawford explained to me.
The past few years have seen an explosion of satellite imaging companies. Older companies like DigitalGlobe continue to put up satellites and produce reams of detailed data. However, they’ve been joined by a number of startups such as Planet Labs, Skybox and more who are putting up a myriad number of satellites.
The result of this influx is that between the various companies watching the planet from above, it’s possible to see changes being made in particular parts of the globe over time, and to track and monitor those changes. Those changes, in turn, can be used to inform decision makers and investors about the status of the economy, agriculture production, and more.
To analyze these images, Orbital Insight has developed machine learning programs that can be “taught’ to find and analyze data points of interest. During its seed phase, for instance, the company used a human to tag cars in parking lots while the program observed it. After a few hundred tags, the program “learns” to identify cars in other parking lot images without the human being involved.

This wasn’t an idle exercise, either. Once Orbital Insight’s deep learning program learned how to count cars, the company was able to determine a mathematical relationship between the number of cars in parking lots of retail stores in a quarter to the store’s revenues that quarter. That type of information is definitely valuable to investors. But for the company, it’s a proof of concept that will allow it to pursue other kinds of data.

This funding round is geared towards utilizing satellite images to find other sources of useful data. One project, for example, is examining the shadows of buildings in China over time. Using those shadows and a little trigonometry, the company aims to be able to figure out the rate of construction in China to see if it’s speeding up or slowing down.

Another project for the company is examining satellite data about oil storage tanks all over the world in order to try and figure out what oil supplies look like in a given moment, instead of beinga few weeks behind like current methods.
The company is also moving beyond its initial customers, which were primarily investors, to work with other types of businesses, governments, and non-profit organizations in order to analyze data of interest to all of these groups. The net result, Crawford hopes, will be revolutionary.
“In long term, I see this as a way to understand the world at scale,” he said. “The combination of satellites and software allows people to really see and assess the world around them.