Seven years ago, three of the largest farm-equipment companies in America said they were close to rolling out technology that would allow farmers to use driverless tractors pulling grain carts during the harvest.

Today, no such product is widely available, and no near-term promises are being made by John Deere, Case IH or Kinze Manufacturing.

“Autonomous systems are a proven concept and are usable in the ag industry,” said Phil Jennings, service manager at Williamsburg, Iowa-based Kinze. “But at this point it’s truly a matter of timing to find the right balance of use case and cost.”

Driverless tractors have long been on the cusp, but they are still not widely used, bogged down by safety concerns, legal obstacles, the difficulties of offering technical support to a large number of farmers using a single product, and lingering doubts about the business case.

The technology exists. Tractors are already self-steering. A farmer in Indiana built his own automated system that he operates with a PlayStation controller, and the major tractor manufacturers have all developed some type of autonomous tractor. Representatives from John Deere and Case IH did not respond to requests for comment.

At the Farm Progress Show in August, the nation’s largest farm trade show, one of the exhibitors was SmartAg, an Iowa startup that is commercializing a system that allows a farmer, from a combine, to summon an autonomous tractor pulling a grain cart. The product, called AutoCart, costs about $40,000.

Colin Hurd, the founder of SmartAg, said transferring harvested grain from the combine to a grain cart is one of the key bottlenecks in farming. It’s tough to find someone to drive that extra tractor. Some kids come back from college to help. Retired neighbors drive the tractor. Spouses take days off work.

“The challenge is that if they’re operating for even 20 percent of the harvest season without a grain cart operator, they’re losing an incredible amount of money,” said Hurd. “Every time the combine fills up they’ve got to go drive it to the edge of the field and dump it. So it slows down the entire harvest process. You’re also going to burn more fuel because that combine keeps running that whole time, and every engine hour you put on that combine also reduces its value significantly. It’s a $500,000 machine sometimes, so every hour you put on that combine sucks a lot of value out of it.”

Two or three extra days of harvest can also mean two or three days less of tillage before the ground freezes. A field that was tilled in the fall warms up faster in the spring, so fall tillage is critical.

“If they don’t get that done in time, especially in Minnesota, that delays planting,” Hurd said. “There’s a ton of data that shows if you delay planting you get less yield.”

AutoCart consists of software and hardware that allows the farmer to summon a tractor pulling a grain cart, and allows the tractor to drive itself into position alongside the combine and stay there as the combine keeps moving and empties grain into the cart. It also includes sensors that can spot obstacles and people and avoid them. The system works with John Deere 8000 series tractors. Five AutoCarts have been installed. Ten dealers have signed up to sell the system, including one in Fargo.

For farmers, the idea is intriguing, said Zach Rada, a farm business management instructor at Ridgewater College in Willmar. “It’s one less person to pay. It’s one less person to find,” Rada said. “It’s a labor issue as much as anything.”

He was speaking on the phone from a tractor he was driving for one of his neighbors during the harvest. Older farmers don’t always know how to run the newer tractors, and hiring seasonal help is a minefield, Rada said.

“You could put an ad in the paper, but good luck finding somebody worthwhile,” Rada said. “Guys like me will get paid $20 to $25 an hour because the farmer has to make it worth your while.”

Kyler Laird, a corn and soybean farmer in northwest Indiana, posted a video two years ago of an automated tractor system he made himself. He nudges the tractor into place with a PlayStation controller and then locks it to a path and speed. His system doesn’t have the self-driving capability of the AutoCart, but it was impressive enough to draw the attention of people across the industry.

Laird said the challenge is to create a product that’s user-friendly for farmers who want to be able to use it once a year without having to think too hard about it.

“It’s one thing to make something for a factory floor, where people are coming in and they’re doing the same thing every day,” Laird said. “Farmers expect to pull the tractor out of the shed when it’s time to harvest and that grain cart needs to be ready to go.”

SmartAg is staffing up to be able to provide service to customers who buy its product, and that will be the largest challenge, said Laird and Stuart Birrell, a professor of agricultural engineering at Iowa State University.

“It is much easier to develop the system than the substantial resources needed for technical support if it becomes a product,” Birrell said.

Birrell predicted that SmartAg will most likely be acquired by one of the big manufacturers, or its technology will be licensed to them.

“While small-venture firms can move very quickly, I would not bet on them against the larger OEMs (original equipment manufacturers),” Birrell said. “The OEMs have much larger resources and effectively control the information on both the tractor and combine that is needed to coordinate the cart and combine.”

Hurd said SmartAg “fully intends” to launch AutoCart commercially in 2019, and he hopes to roll out a second product that would enable tractors to till fields autonomously, another bottleneck in the annual cycle of farming.

“To scale the company we will require significant capital and plan to raise that capital through private investors or venture capital,” Hurd said. “Our focus is not on selling the company, but we are open to considering the right opportunity should it present itself.”