“We didn’t need to look at many companies to establish that there was one clear leader that was way ahead,” John Teeple, director of advanced technology for John Deere tells AgFunderNews.
Teeple is talking about Blue River Technology, the Silicon Valley agricultural robotics startup that’s using computer vision and machine learning to identify weeds in the field and spray herbicide on them autonomously.
Deere & Company announced plans to acquire Blue River Tech for $305 million last week as the next step in its quest to build autonomous equipment for agriculture.
“As we thought about what we needed and the next iteration of in-field challenges we wanted to solve — where we expect machine learning to be able to sense, make decisions, take an action, and then learn from that in real time in order to optimize yield potential, reduce costs, and create environmental savings — that next set of capabilities was something we didn’t have,” said Teeple.
Deere is famous for its wide range of green tractors, many of which use state of the art technology such as computer vision to guide its sprayers and Global Navigation Satellite System (GNSS)-based machine guidance to help tractors steer autonomously. And while the company is constantly innovating internally, there are limitations to what can be achieved “organically” in-house, according to Teeple.
“We always view ourselves as a tech company and think about being at the cutting edge of our industry, but for us to do what Blue River has over the last six years would have been a very long and expensive process, very difficult in terms of access to talent, and something we can’t really emulate at John Deere.”
Not only is John Deere’s computer vision technology “rudimentary” compared to Blue River’s, according to Teeple, but the startup “has a collection of unique talent from academia, and a diverse set of experience which, together with technology advancements in last couple of years in hardware tech and GPUs where you can now do processing on the edge [without the internet] as opposed to in the cloud or on super computers, has created a whole new set of capabilities in machine learning, deep learning, and computer vision that Blue River has optimized and focused exclusively on for ag applications,” he said.
The largest agricultural machinery company in the world spent over a year talking to Blue River before deciding to purchase it, as well as looking at over 50 other companies working on robotics and automation in the field. But no other companies came close to offering Deere what it was looking for, according to Teeple.
“To be able to characterize and identify weeds from plants in real time and over a million plants an hour is really stretching the capabilities of today’s applications of computer vision,” he said.
The acquisition was the next logical step for Blue River too, according to Teeple and Blue River’s CEO Jorge Heraud.
“As you think about where they were going next, they would have been building out their manufacturing capabilities and distribution networks, many things that John Deere has been successful at around the globe for 180 years, so there’s a nice match of capabilities; where each of us saw a need, the other fits that need. Jorge has described it as a hand in glove relationship where both companies have equal benefit.”
Deere is keen not to stifle this innovation, as can often befall startups acquired by large corporates, and will keep Blue River in Silicon Valley as an independent subsidiary reporting to its precision agriculture department.
Blue River will, therefore, continue on the same path it had before the acquisition; to bring its see-and-spray robot to the market next year. It will also explore other areas of John Deere’s business where its technology can add value, such as new crops or different types of machinery. “We will help make all John Deere machines smart and not just in ag; there’s an opportunity for Deere’s construction machinery to act more precisely too,” said Heraud.
Teeple is keen not to get too greedy too soon, however. “The challenge will be to manage our appetite to ensure the Blue River team stays focused, so we will be selective with the connection points as the Blue River team grows.”
Blue River is planning to expand the team immediately and add 12 new members of staff to its engineering and machine learning teams. Blue River’s ability to attract high-quality talent was another major appeal of the acquisition, according to Teeple who said it’s a struggle for a large company like Deere.
Will John Deere look for more acquisitions in the near future?
“Whether through recruiting or potential inorganic plays [like acquisitions] we are always looking at the next generation of solutions so will keep our ear to the ground,” said Teeple, adding that he plans to tap into Heraud’s connections to the agtech startup community.
Deere is also invested in the Iowa Agtech Accelerator, which Teeple says is showing some interesting startups from across the globe.
Agricultural robotics has typically remained a small part of the growing agtech startup universe, as a capital intensive and challenging segment. Last year the category experienced a pullback in funding to just $100 million for 17 startups. In the first half of this year, 12 farm robotics, mechanization, and equipment startups raised $87 million, according to AgFunder’s upcoming AgriFood Tech Report, which could represent some growth.
*Make sure you receive a copy of the AgFunder AgriFood Tech Report as soon as it’s released by signing up for the weekly newsletter here.*