Robotics in Organic Farming w/ Craig Rupp

Interview with Craig Rupp, CEO of Sabanto

Show Notes

Robotics and self-driving tractors may sound like a pipe dream, but for Craig Rupp, his company is already doing it for farmers across the country. In this episode, we hear from Craig Rupp, CEO of Sabanto, as he shares the latest about robotics and automation in the ag industry.

Learn more about Avé Organics: www.aveorganics.com

Learn more about Sabanto: www.sabantoag.com

Connect with our guest on LinkedIn

#agriculture #farming #organicagriculture #organicfarming #robotics #selfdriving #tractors

Podcast Transcription

INTRO: Welcome to Organics Unpacked, a podcast for the business-minded organic grower — an interview podcast where we hear from the top experts in the commercial organic industry, with a focus on the business elements of organic growing both in and out of the field. You will gain insight and grow your operation. This show is brought to you ad-free by Avé Organics, a Wilbur-Ellis company. To learn more about Avé Organics, visit our program notes. In the meantime, enjoy the show. 

TOM: Hello, everyone. Thanks for tuning in today. Welcome to a new episode of Organics Unpacked, a podcast focused on organic farming from a practical standpoint. I am your host, Tom Buman. Today, I’m joined by Craig Rupp. Craig is the CEO of Sabanto, a company that really focuses on robotics and has a special emphasis in organic farming. I think Craig’s work is going to really prove to be eye-opening for a lot of people. Craig’s going to share some of the newest information on robotics and crop production. Craig, welcome to the show. 

CRAIG: Thank you, Tom, and I’m glad to be here. 

Background in Autonomous Farming

TOM: Thanks. Craig, give me a little bit of background on yourself. How did you get to where you are today? What is the process that got you to where you are today? 

CRAIG: For starters, I think what really helped was I grew up on a farm in northwest Iowa, Cherokee, Iowa. When I graduated high school, I went on to engineering as opposed to farming. I guess I was the spare, not the heir. I went to Iowa State, got an electrical engineering degree. Then I went to Chicago and started working at Motorola. So I spent the majority of my career in the wireless industry, and, given my agriculture background, throughout my career, I’ve kind of been pulled back into agriculture. I was living in Chicago, and I moved back to Iowa. I took a job with John Deere, and I developed the StarFire receiver and the GreenStar display for them. Then I went back out in the wireless industry. Then a friend of mine and I decided that we’re going to start a company called 640 Labs, and we’re going to start collecting data off of tractors, combines and applicators. We headquartered that in Chicago, and we were acquired in 2014 by Monsanto. Then I just started working hand-in-hand with farmers and understanding a lot of their labor issues. When Bayer bought Monsanto, I left Climate. I was actually working for Climate. So the FieldV drive, I was the guy behind that. Then I decided that I wanted to do something about this labor problem in agriculture, and I decided that I’m going to start a farming-as-a-service company using autonomous farm equipment.

Mission of Sabanto

TOM: Tell us about Sabanto, what kind of products and services you have. Just give us a general overview of your company.

CRAIG: So we were founded in October of 2018. Then we decided that we’re going to do something audacious in the spring of 2019, and we decided that we’re going to go out and autonomously plant throughout the Midwest. So I rented a JCB 4220, and the only reason we got that tractor was it was 102 inches wide, and I could put it on a drop deck trailer and move it from farm to farm and autonomously plant. Then I bought an 18-row, 20-inch Harvest International planter, full up precision planting, seed meters and down force. Then we started autonomously planting in Iowa, Nebraska, Minnesota, Illinois and Indiana. Then I caught the attention of some investors, and I was funded in 2019. I started building a team, and I went back to Chicago. I was in Ames, the Iowa area, at the time. Then I went back to Chicago and hired four guys that could help me pull this off. Then, in the fall of 2019, we started doing cover crop seeding. We took our tractor and put a drill behind it. Then, in 2020, we decided to go smaller. It turns out that JCB 4220 is kind of an expensive platform, and we went with the smaller Kubota tractors, 60 and 90 horsepower, and we built up four little planters. We took that 18-row planter, and we chopped it up and made four five-row planters and started autonomously planting with multiple systems throughout Iowa, Nebraska, Minnesota and Illinois again. 

What’s interesting is when we came back from planting, then, lo and behold, all these organic growers called us. It turns out that it’s kind of a unique industry in that, the conventional guys, they have a labor problem, but the organic guys have a big labor problem because once they’re done planting, they have to tine weed, rotary hoe, cultivate. They’re spending a lot of time in the driver’s seat, and we started going out rotary hoeing and cultivating for farmers throughout Iowa and Illinois. So we did that the majority of the summer, and it was kind of interesting. I always tell people there’s this software development process called agile, where you work in two-week increments, and it’s a sprint. You try to get as much as you possibly can get done. It turns out this one farmer — he wanted us to cultivate every 10 days. It worked out really well because we had to cultivate his field five times every 10 days. So we went out and cultivated, and we came back and said, well, we need to improve this, this, this. We need this feature. Then the software guys would spend, instead of two weeks, they would spend 10 days updating their software, and we would go out. Just those five iterations really, really helped us.

Uses of Autonomous Equipment

TOM: So you talk about autonomous equipment. Tell me. I equate that with robotics. Maybe that’s right or wrong, but tell me, specifically, what your equipment does that’s different than, typically, what’s done in the landscape today.

CRAIG: Let me give you an example of what we did in terms of rotary hoeing. The first thing we do is we strike the boundary of the field. Then what we do is we take the farmer’s data, and we know his planting data. So we know where the rows are. What we do is we take this little tractor out there, and it’s got GPS receivers. It’s got a safety system, cameras on it that we monitor if there are any safety issues. We control the steering. We control the throttle. We control the hydraulic transmission. What we do, then, is we create these path plans of which this tractor will drive. It follows this path within a certain margin, and what we do is we perform the field operation, and we monitor it remotely. So there’s nobody sitting on the tractor. It goes out, and it runs 247. This last fall, we ran one of our tillage rigs 39 hours straight. 

Fuel for Autonomous Equipment

TOM: Wow. So how do you deal with the fuel issue? If you’ve got a planter, filling the planter, are there traditional ways of doing that? Or do you have new and improved ways? What does that look like?

CRAIG: So let’s just look at fuel, for example. I mentioned before that I worked on, or I developed, the FieldView drive. I spent probably three to four years of my career debugging, reverse-engineering CAN bus messages on the tractor. So that fuel gauge on the display of that tractor comes from a CAN bus message. I have engineers who have gone through and found out where that message is, and what we do is we intercept that message. I can tell you just what that gas gauge reads, or the fuel gauge reads, at any given time. So we monitor that.

TOM: Great, so you monitor the tractor. You know when it needs something. You know when the planter needs something. You make those stops. You do what needs to be done, and then you push a button, and it sends it back on its way?

CRAIG: Exactly. One thing we learned this fall was we have enough fuel to run about six hours nonstop. So we’re adding more fuel capacity, such that we can run longer hours without having to attend it. 

Learning Curve for Autonomous Farming

TOM: I always think we’ve got these autonomous things, but what’s the worst accident you’ve ever had with some of your autonomous things in the learning curve that you went through?

CRAIG: In the learning curve? We have safety systems on board, so if there’s an obstacle detected in front of it, it will stop. We haven’t had any accidents per se. We do have the ability, if we go off path by more than a certain amount, it stops. If it goes, we geo-fence the field. If it gets within a certain margin of that field boundary, it stops. If we lose communications with it, or we lose communications to the back end, or if we lose RTK, we stop. So we’ve gone through. Also, hardware, we have multiple systems. So there are really three systems working in parallel. If one of those systems acts flaky, then the entire tractor dies. So we haven’t had any accidents per se, but we are monitoring. We’re pretty up on safety, and we make sure we follow procedures. We try to stay out of its way whenever we can.

TOM: So one of the things that people that are really kind of into soil health and organic farming tend to think about is what can I do to improve my soil? Compaction is just one of those things that is a killer in soil health, and it’s because we just have so much big equipment. With autonomous vehicles, are we going to see even larger vehicles and larger equipment?

CRAIG: I think we peaked in horsepower.

Sizes of Self-Driving Tractors

TOM: Are you going to see heavier equipment in the field with autonomous? You’re talking about Kubota tractors. Is that kind of for tests, or do you see equipment size getting smaller with autonomous tractors?

CRAIG: So let’s just take a look at where we’re at today. Tractors and implements have been designed such that they want to maximize the number of acres an operator can cover in a given amount of time, usually a 10-hour day, limited by DOT rules and what they can haul on a truck and what they can put in the machine shed and at the expense of cost and size and all sorts of things. Now, we have the ability to run 247, and we can deploy multiple rigs, and the operator is out of the equation. So look at it this way. If I had two small rigs, I would have to have two people in it, which would be a big problem for a farmer. Now, I can run multiple and do a swarm type of architecture with these things. There’s nobody on them. There’s someone monitoring it remotely, but they’re monitoring multiple of these little things, of these little machines. So I honestly think the autonomy will bring down the required horsepower and size and weight of the equipment and improve the soil health by lessening the compaction.

TOM: Yeah, I just really see, as we scaled agriculture, everything has gotten bigger, and, a lot of times, that’s hard on our soil. This is one of, to me, the promising things of, maybe, a technology that really could benefit soil health by, like you said, providing multiple systems in a field that can do 247 work and get the same amount of work done, maybe even faster.

CRAIG: Yeah, I think that is the case. We went out and deployed three units at one time in a field. In all three units, each one weighed. The tractor is about 5,500 pounds, and the planter is about 2,500 pounds. So it’s a really lightweight system, and I think there are some other advantages. I think we can probably go out earlier than some of this large equipment. There’s a lot of concern we get from farmers about the compaction that they have, and they really want to find a solution for that.

Future of Autonomous Farming

TOM: So you’ve talked a little bit about the things that you’ve tested — seeding, cover crops, doing some planting. Where do you think you’ll be in the near future, and where do you think this can go in the long run?

CRAIG: In the near future, as in next week, we’re going to be in Nebraska, working with a large organic farmer there, doing some tillage for him. We’re working with some organic growers and performing just some of the mundane field operations, like rotary hoeing, for example. You’re going to see us all summer doing field operations like that. We’re going to be doing some spring tillage, as well, with some high-speed discs. You’re going to see us this fall doing that, doing cover crop seeding and also doing some fall tillage, as well.

TOM: Where do you think this can go in the long run? How far do you think autonomous vehicles can take agriculture?

CRAIG: I don’t think it’s an all or nothing, meaning that I really can’t see an operation being 100% autonomous. I think there’s going to be a path where we get to 80% autonomous, and the other 20% is going to be done manually. I see that in the near future.

Scaling Autonomous Equipment

TOM: All right. Then, as we think of scaling this, do you see this helping the large farmer, the small farmer or is it independent of farm size?

CRAIG: I think it’s going to be independent of farm size. We get a lot of people reaching out to us, large farmers and small farmers. Some of the small farmers, they’re getting up there in age, and they want to have a succession plan. I want someone to take over the farming operations. I see a lot of that. You go to some of these large growers, and labor is a big problem for them. They want to scale. They want to get larger, and they see what we’re doing as a path forward for them to scale.

TOM: So, Craig, conservation is very near and dear to my heart, and I think a lot of organic farmers are. So a lot of people are still contouring and stuff. The autonomous system, does it fit well into anything but straight row patterns, or do you have to have pretty things laid out in a straight line in order for it to work?

CRAIG: Very good question, Tom. At this point in time, we do straight rows. However, nothing is stopping us from doing contours. Mostly everything we use hardware-wise is off the shelf, from the tractor, the GPS unit, the controller. All the data acquisition devices that we use are off the shelf. We did have to create a path planner. So we have a coverage planner as to where in the field that needs to be planted and the interior — the headlands, or end rows, versus the interior. Then we have a path planner, which takes the passes and then all the turns. This is where, at any given point in time in the field, we know exactly where it’s at, and we follow that recipe. For us to do contour, it’s not a problem because all the points in the path are separated by one meter. We can do contours, and we’re working with agronomists. So, if the agronomists would come to us and tell us, here’s how I want this planted. At least give us a drawing of how he wants the contours. We’re more than happy to do that, and we can do that. At this point in time, we haven’t done it yet.

Accuracy of Robotics

TOM: So, going across your field, what kind of accuracy? If you go across twice, how much variance is there in space and distance? Is it within inches? How close can you get, when you draw up a row pattern, to keeping your equipment in that line?

CRAIG: I can tell you, right now, we’re three centimeters. One sigma. We’re within an inch with a little Kubota tractor, and we can improve that by going slower, taking the front end down a little harder. There are tricks we can play, but we’re pretty good. We’re industry standard.

TOM: I remember as a kid, dad would put the sweeps on the cultivator a little further apart because his sons couldn’t drive the line close enough. So I assume that you probably get a little better weed control with robotics because you can control the row pattern so closely.

CRAIG: Exactly. What’s interesting is our system, to us, is open, and there are just possibilities that we have, or options that we have, to improve that above and beyond just had we gone out with an off-the-shelf system.

TOM: So kind of another question is I know that when dad would send his boys out and do some spot spring, sometimes, it was easier just to leave the sprayer on than turn it on and off all the time. I would assume that robotics has a pretty good fix for that. Once you program it in, it does what you tell it to do.

CRAIG: That is true. I’ve talked with a lot of farmers, and one of their complaints is the operator, especially tillage. This one farm was doing deep tillage, and it got to be Friday afternoon, and five o’clock is coming up, and the operator was kind of cheating. He wasn’t going as deep as what the farmer wanted. We can’t cheat. It’s programmed to plant a certain way or apply these operations. You can’t cheat it.

An Agriculture Technology Company

TOM: If you were going to describe Sabanto, are you a software company? Are you an equipment company? Where would you describe your sweet spot?

CRAIG: We’re an agriculture technology company. I don’t consider us an equipment company. I don’t see us building a tractor. I don’t see us building an implement. We’re using off-the-shelf implements and tractors and working with various manufacturers, equipment manufacturers, that is, to give us the pieces required for autonomy.

TOM: So, especially in organic farming, how soon do you think it will be before 25% of the acres, let’s say, are farmed autonomously? Can you look into the future and say, how fast is this coming? Is it incrementally slow? We’re going to do a few things. Or is it like, boom, we’re going to see something happening really fast?

CRAIG: That’s a really good question. It’s a really hard question. I think it’s going to be fastened to what people think because most every farmer I talk to, they’re really, really interested in it, and they see the vision. They say, this is going to solve my labor problem, and they understand. I like the idea that you guys are going smaller, lower costs, lower CapEx, lower OpEx. I really like what you guys are doing. So I think it’s going to come sooner rather than later. To answer your question, it could be five years.

TOM: Really, before 25% of the organic farms might have some type of autonomous equipment on them?

CRAIG: I could see that. I could see that in five years. Yeah, I can. You know, Tom, the tractors we use are commoditized for the most part, and they’re mass produced. The tractors we use, you can build them quite fast.

TOM: Maybe one of the biggest obstacles you’re going to have is having people sell their big tractors and downsize, huh?

CRAIG: Yeah, perhaps.

Sweet Spot for Autonomous Farming

TOM: It might not be a cheap thing in itself. So what’s your sweet spot, do you think? Where do you think is the most growth when you look across all the different crops across all the different geographic landscapes in the U.S.? Where do you think autonomous farming is going to happen fastest?

CRAIG: I would like to think it’s probably going to be on the organic side. What is interesting about the organic farmer is they’re very methodical. They’re very disciplined, and they know that they have to spread out their field operations because they can’t plant 2000 acres a day because, then, they’re going to have to cultivate 2000 acres a day. So they’re very methodical, and they like spreading things out, and that’s something that I had no idea. We were working with a lot of conventional guys. Then I got to know a lot of these organic farmers, and I’m like, well, this is something interesting. You go to the conventional guys. We were working with some large 20,000-plus acre farmers, and, I mean, it’s a sprint. When that ground is ready, I mean, they’re going 100 miles an hour. Then you work with these organic guys, and they’re like, well, wait a minute. What these guys do in a day, these guys do in a week, and it’s kind of interesting. I thought it was rather interesting as to how they operate.

TOM: Just because they do have that labor force, and it’s limited, and they have so many things to do, right?

CRAIG: Yeah. If they do three to five passes over the ground, you’re going to have to have a guy sitting in that seat to do that, at least in the Midwest. That’s where I see a lot of early adopters. Truth be told, organic farmers, their mindset? They’re pretty progressive guys. You know what I mean? They really are compared to a lot of the conventional guys.

TOM: So, Craig, if somebody wants to get a hold of you and learn more about your technology, what’s the best way to do it?

CRAIG: You can go to our website, saban​toag​.com. You can leave a message there.

TOM: Okay. I also noticed you got some great videos there of tractors running and stuff like that.

CRAIG: Yeah. We have a pretty neat culture in that a lot of even our software guys grew up on a farm, and the majority of them have worked in agriculture in one way or another. So they know the ins and outs of corn and soybeans, and it’s worked out fairly well for me. 

What Farmers Should Know

TOM: Great. Well, Craig, I want to be mindful of your time, and I really appreciate the time you took, but one last question. If you had a farmer in a room, and you had two minutes together, what would you want to convey to him or her about your company’s products or services? 

CRAIG: The first thing I would do is I would tell them I have a solution for your labor problem and your capital expense problem. 

TOM: Wow. Didn’t take two minutes, did it? 

CRAIG: No, it didn’t.

TOM: Alright. Craig Rupp, CEO of Sabanto, thank you. With that, thanks to the listening audience. Be sure to tune in every week when the Organics Unpacked podcast will unpack yet another facet of organic farming. Drop us a note if you’d like to see a specific issue covered or a specific expert interviewed. Thank you for your time.

OUTRO: Thank you for listening to Organics Unpacked. If you enjoyed this episode, please consider subscribing and giving this show a five-star rating and review, so we can continue to help organic growers improve their operations.