There are many very cool high-tech ideas bouncing around the specialty crop industry, moving forward at various levels of speed, jumping one hurdle after the next. But there are also success stories and real agtech being utilized on the farm. Below are two companies that are well past concept development and are offering equipment that materially reduces the need for labor.
Bear Flag Robotics could well be the poster child for agtech startups. Co-Founders Aubrey Donnellan and Igino Cafiero started the company in 2017 in Silicon Valley, focusing on the development of autonomous farming tractors that are compatible with existing farm machinery. The emphasis was on retrofitting existing tractors with patented artificial intelligence technology to increase their efficiency and productivity. They soon joined a startup accelerator program and received seed money. They also became a resident of the Western Growers Center for Innovation & Technology. In 2018, Bear Flag partnered with a venture capital company and over the next couple of years received significant funding to move its concept forward. They first put an autonomous tractor in the field in 2018 and launched a commercial version of their technology in late 2019. By 2021, they were involved directly with many growers, including partnering with Church Brothers. Along the way, Bear Flag Robotics began working with John Deere as part of its Startup Collaborator program. The collaboration was a success, and the Bear Flag Robotics was acquired by John Deere about a year ago. The agreement has been valued at $250 million with Bear Flag staff still in place and running the operation, albeit with significant help from John Deere.
Cafiero said the agreement has resulted in a significant acceleration of the development and implementation of its farm automation. He recently told WG&S that while Bear Flag Robotics has successfully launched and been acquired, he has learned a lot in the process and would do things a bit differently if he had it to do all over again.
In creating an autonomous tractor, Bear Flag concentrated on tillage, reasoning that the operation was a time-consuming process that could be automated, saving time and money. The startup had a goal to deliver value to the end user as quickly as possible. Bear Flag had this philosophy and executed it well, but Cafiero said the move to the field was still too slow.
Farmers, he said, are not looking for perfection. They are looking for something that works. “My advice is to get out into the field with what you have as quickly as you can and work with farmers as soon as you can.”
He explained that the trap some startups fall into is that they gather a lot of smart people in a garage, and they develop and build a high tech piece of equipment that’s really cool but doesn’t necessarily do what the farmer needs it to do. “When we started Bear Flag, we were a robotics company trying to create an autonomous tractor to deliver to the farm community. Now we are an ag farming company trying to deliver value to the farmer through the use of robotics. That’s not just semantics. It’s a big difference.”
Cafiero said the great thing about farmers is that they are not looking for perfection, they are looking for value. If you can provide something that does the job better, they will quickly embrace your technology. “They are very quick to adopt new technology if it has value,” he said.
At its core, the Bear Flag concept was pretty simple: create an autonomous tractor that could do a job in a more efficient way. Cafiero said tillage is an operation every farmer employs. Bear Flag has now put real equipment in the field that has done tens of thousands—even hundreds of thousands—of passes without a driver. That is clearly creating value.
And Cafiero said the beauty of the high-tech equipment on the tractor is that those passes are also generating a mountain of data. “The more passes we make, the more that data becomes meaningful.”
For example, he said the passes are basically mapping every field and a farmer can quickly see where he has compaction. That data, he said, can be used to work those areas and increase yield and productivity. But Cafiero was quick to add that “data for data’s sake is useless.” It has to actually do something that benefits the farmer.
The company is not resting on its laurels, and, in fact, Cafiero said the acquisition by John Deere has allowed the firm to speed up development on other ag technology. The common thread is that the company is focused on “developing the best-in-class technology for autonomous agricultural equipment.” Its goal is to reduce the need for labor in terms of “mundane tasks” and free them up for more revenue-generating operations.
Speaking of the ag space, Cafiero said farmers are not trend chasers, and in any endeavor, there will be those early adopters and those that trail behind waiting for the concept to get a bit more mature. “But we have found plenty of early adopters,” he said. “If you show value and they see something that will move the needle, farmers are quick to adopt it.”
FarmWise Labs, Inc., an American agricultural technology and robotics company and another resident of WGCIT, also is moving forward on many efforts to mechanize. But it’s also already operating in the space with an automated mechanical machine that uses a combination of AI, computer vision and robotics to pull out weeds in vegetable fields.
The company currently has 12 of the weeders in the market and is offering a weekly weeding service to many growers in the Salinas Valley this summer. The company, founded by Sebastien Boyer, is well past the concept stage. “We charge a simple fee per acre for the service,” he said.
He added the service saves the grower money as the fee works out to be less expensive than the cost of the labor it replaces. As a rule of thumb, Boyer said one machine replaces about 10 workers doing the job manually. Depending upon the crop, a FarmWise weeder can complete the task on 5-10 acres a day.
FarmWise continues to offer trials and is marketing the service, but Boyer said the 12 machines now in service are solidly booked for the next 12 months. “Demand is very high, but we are in the process of adding more units,” he promised.
He estimated that by early 2023, the company will have expanded its capacity with additional units. The current FarmWise business model is to offer the weeding as a service, providing the equipment and the operator. All a contract customer must do is order the service via a cell phone app, specifying the date and the field to be serviced.
“Eventually we will have a purchase option as well,” Boyer said. “In fact, we are taking orders for delivery next year.”
Boyer noted that there are quite a few companies that are currently touting automatic weeders. He expects that two or three, including FarmWise, will survive, but he does expect there to be consolidation. And he also expects automated weeding to be quite prevalent in a relatively short time span.
“What sets us a part is the team that we have built,” he said. “We have two sets of workers. We have a group with technology expertise such as AI (artificial intelligence). And we have another group of workers with farming expertise. That helped us loop in farmers in our development process very early.”
Boyer said the early involvement of people that knew exactly what happens in the field and what a weeder had to be able to do and the challenges it faces helped FarmWise develop a workable unit very early in the process.
He said the business strategy has worked. FarmWise has created an automated weeder that does the job, does it economically for the user, and the economic model works for the company. “Each machine is profitable,” Boyer said, noting that the company isn’t yet profitable because of the high cost of development. But he said as FarmWise scales up, the financials will pencil out.
In addition, Boyer has greater ambitions for the machines the company is continually upgrading. They are developing this automated equipment so that one machine can do multiple tasks and work on multiple crops. In the future, he expects both the service and purchase options to be available. When a grower buys a machine, Boyer envisions that he or she will be able to weed a leafy greens field one week and a broccoli field the next. The same machine will be able to be programmed to spray multiple crops, the week after it weeds a ranch.
Boyer said he wants to “democratize” ag automation by building equipment that allows the user to determine how it will be utilized. This, in a nutshell, is the promise of AI. The machine will continually learn offering new services and new data.
Already, Boyer said a machine that moves through a 10-acre field weeding also is gathering data that the grower can access. For example, after weeding a broccoli field, a report can be generated that will tell the grower how many plants are in that field and their stage of growth. The data generated is an important aspect of how this new technology will compound its value to the user.