PITTSBURGH — The rocks are as dumb as ever, but the people behind TerraClear are getting smarter every day as the ag-tech startup fine tunes its AI-powered machine that helps farmers pick those rocks out of fields.
Based in Bellevue, Wash., and Grangeville, Idaho, 4-year-old TerraClear was started by Brent Frei, the former CEO of Onyx Software and co-founder of Smartsheet. The company uses mapping technology, computer vision and AI to help guide its robotic Rock Picker, a device that can be mounted to a piece of farm equipment to grab an average of 400 rocks per hour out of the soil.
TerraClear President Trevor Thompson said the company has sold every picker that it’s built since last fall. But supply chain concerns are constraining how many and how fast new ones can be built.
“Each iteration, we’re kind of tweaking the design to where we’re ready to really deploy this thing at a larger scale,” Thompson said.
Thompson got off the farm and traveled to Pittsburgh this week to bring word of TerraClear’s work to the Cascadia Connect Robotics, Automation and AI conference. The conference is organized by Seattle-based Cascadia Capital, which is underwriting GeekWire’s independent reporting on the topic.
In a panel discussion titled “The Smart Field: How is RAAI Reshaping Agriculture?” Thompson said TerraClear provides farmers “freedom from rocks,” a very old and recurring pain point for the industry that is costly in terms of labor and the impact on expensive machinery.
The panel discussion, which also included ag-tech innovators working on how crops are grown and harvested, highlighted the role automation will play in the near future and how startups are innovating in the mix with just a few giant companies that control all key aspects of the industry.
“We have yet to meet a farmer who did not want an autonomous Rock Picker to do the thing that they would have to do otherwise,” Thompson said. Getting to that point has involved meeting farmers where they are.
At a conference and in a city where self-driving cars are often the talk of the town, autonomy on the farm could happen faster than on the streets — while presenting its own set of challenges.
Modern-day combines and tractors already employ a good deal of automated features, so agriculture isn’t just waiting for the concept to be perfected on roadways. But Thompson said getting the farmer out of the cab elicits at least one concern.
“It’s fine when they can see it,” he said of farmers watching a self-driving machine on their land. “But as soon as it goes over the horizon there’s a certain unknown fear that it’s going to continue down the canyon and they’re going to lose $500,000.”
Innovation will also come quicker to tasks that are more low risk, such as rock picking, Thompson believes. Farmers certainly will be quicker to jump at a machine that can handle such grueling work, but they might be slower to let a machine pick delicate fruits and vegetables.
Some on the panel have been underwhelmed by how slowly the giant agriculture companies and equipment manufacturers have been to innovate. But when it comes to harvesting, big tech got a mention from the panelists. The harvesting of data is what’s attractive to Google and Microsoft.
Renee Vassilos, director of agriculture innovation for The Nature Conservancy, called out Amazon Retail in India and the agronomy services it has launched, helping farmers improve yield and quality of fruits and vegetables with technology.
Whether TerraClear ever becomes a tech giant of its own remains to be seen. Frei has previously said it has the potential to be bigger than Smartsheet, which is an $8 billion company.
Last year TerraClear raised $25 million in new funding, and to date it has raised $38 million. The startup now employs 35 people, having hired a bunch on the engineering side in the last year.
In the world of ag-tech, Thompson said TerraClear is “kind of the popular kid in town” because the startup is solving a problem that is visceral and right now, and farmers like technology that can solve a problem immediately.
“We’re using technology that’s going to make things better over time,” Thompson said. “But it’s not like it’s a predictive problem. It’s not a future problem.”
