US imagery analysis startup Descartes Labs has raised US$5m to develop AI-based solutions for the agriculture industry. CEO Mark Johnson tells SatelliteFinance how it is taking advantage of the flood of satellite-based data coming onto the market.
US imagery analysis startup Descartes Labs has raised US$5m to develop AI-based solutions for the agriculture industry, taking advantage of the flood of satellite-based data coming onto the market.
The series A financing comes hot on the heels of its first product release in August to provide corn yield forecasts across the US during the growing season.
The company was incorporated only a year ago when its core team was spun out of the Los Alamos National Laboratory, having been working on remote sensing for more than a decade.
Its CEO and co-founder is Mark Johnson, whose previous ventures include Powerset, now part of Microsoft’s Bing search engine, and personalised magazine app Zite, which CNN later sold to Flipboard.
Agriculture-focused VC Cultivian Sandbox led its most recent financing round, and as a result its managing director Andy Ziolkowski will be joining the Descartes Labs board.
Crosslink, which led its US$3m seed funding in December, also participated, along with TenOneTen, Data Collective, Correlation and ValueStream Labs.
In an interview with SatelliteFinance, Johnson said the proceeds will help grow its team as it looks to expand outside the US, particularly in countries where agriculture forecasting is not as reliable.
Almost all of the group’s analysis is currently sourced from free datasets from the US government, although it has a partnership with startup satellite operator Planet Labs and, as more smallsat imagery players announce LEO constellations, Johnson said increasing competition will keep prices low and give it more data for new innovative solutions.
“It’s a great time to be in satellites,” he said.
“There’s an incredible historical record of data and tons of satellites going online in the next few years, which will give us an unprecedented view of the planet. Someone is going to need to process and make sense of all that data and Descartes Labs is well positioned to do so.”
Some of these new operators, however, believe providing in-house analytical solutions will help differentiate them in the increasingly crowded sector. California’s Hera Systems, which recently raised several million dollars to begin launching its constellation next year, is looking to create a one-stop shop for its customers, delivering pre-packaged data directly to mobile devices.
A variety of data sets from multiple sources is key if customers want to be confident they are seeing the whole picture, according to Johnson.
“The danger of building a complete end-to-end product within a satellite venture is that you may be missing other valuable pieces of data from other sensors” he said.
“Once you’ve built a hammer, you start looking for nails. By analysing data from many sources, we’ll be able to pick the right constellation and datasets for the problem we’re trying to solve.”
As for rival satellite analysis firms on the ground such as Orbital Insight, which raised US$8.7m earlier this year to develop its AI-based analysis tools, Johnson believes Descartes Labs’ ability to process vast amounts of data – quadrillions of pixels to get its corn yield estimates for instance – will give it the edge.
He said it is still early days for the industry, but that its focus on understanding global crop production will help it stand out from the crowd.
“I believe that what separates good startups from great startups is focus,” he said.
“There are a lot of interesting problems in agriculture and we intend to continue applying our brainpower to those problems for now. That being said, I expect that we’ll start exploring applications in oil & gas, mining, and shipping at some point in the next year.”
Another potential market is the finance industry, where traders are always looking for unique data sets to give them an edge.