Join the Newsletter

Stay up-to date with food+ag+climate tech and investment trends, and industry-leading news and analysis, globally.

Subscribe to receive the AFN & AgFunder
newsletter each week.

artificial intelligence & robotics startup

How an Artificial Intelligence & Robotics Startup Got Acquired by John Deere [podcast]

October 3, 2017

In this episode, I speak to Jorge Heraud, the cofounder and CEO of Blue River Technology, a Silicon Valley startup bringing robotics to the farm. Blue River was recently acquired by John Deere, the world’s largest tractor company, for $305 million. This was extremely exciting news for the agtech industry as exits have been relatively few and far between.

Blue River was founded in 2011 and has some big investors behind it, including Khosla Ventures, and Innovation Endeavors, which is the VC fund of Eric Schmidt, Google’s chairman, as well ag industry giants Monsanto and Syngenta. The company raised just over $30 million in total since its founding.

Blue River’s main technology is a see-and-spray robot, which is a tractor attachment that uses computer vision and machine learning to identify weeds amongst rows of crops, and then spray herbicide directly onto them. Blue River says it can reduce agrochemical use by 10 times, with significant impact on the bottom line for farmers, but also on sustainability metrics.

In this podcast you will find out about Jorge’s engineering background, the decision to launch Blue River, the challenges of being at the cutting edge of robotics, deciding to sell to John Deere, and tips for entrepreneurs in raising capital.

I hope you enjoy our conversation — there’s an abridged transcription below too.

Louisa B-T: Being acquired by a leading company in the industry in which you work is the dream that most entrepreneurs have. So, how does it feel?

Jorge Heraud: Yeah it feels pretty good. As you said, John Deere acquired Blue River just a week and a half ago when we closed the transaction. And yes, I’m very happy. They’re a leading company with a very good reputation, and they’re acquiring us to accelerate us. We’re going to stay where we are in Silicon Valley, and stay innovative. They’re treating all our employees very nicely and, again, they’re acquiring us to accelerate us. That’s not always what happens in acquisitions. This is exactly what we wanted to happen.

Jorge Heraud

Louisa B-T: Fantastic. Let’s rewind a little bit and talk a bit about your background and how and when Blue River Tech started. So, you worked at Trimble for many years, and you’re an engineer by background. How did you come to the idea of setting up Blue River, and what was your initial mission?

Jorge Heraud: So, yes I’m an engineer by training with a master’s in Electrical Engineering and Engineering Management from Stanford. I worked at Trimble for about 15 years, for the most part in precision agriculture. I was the Head of Engineering for the agriculture group, then I was the Head of Business Development working in acquisitions, and we were successful at acquiring four companies that were in the precision agriculture space, so I ended up managing that as the head of the precision agriculture business unit. After about 15 years at Trimble, I had a really good appreciation for precision agriculture. I had also migrated from engineering to the business side, and so I decided to take a year to go to business school to do an MBA. So I went back to Stanford, and during that time I thought really long and hard about what I wanted to do next and decided that some of the entrepreneurs whose companies I had acquired were having way too much fun, and that’s what I wanted to do next.

Louisa B-T: How was that transition from a large, big corporation, and obviously to do MBA, but then becoming a Silicon Valley entrepreneur? That must’ve been quite a change for you?

Jorge Heraud: Yeah. There are many, many differences. Some on the personal side, some on the business side. It was me and my co-founder, Lee Redden, who was also at that time studying at Stanford. He was doing his PhD in Robotics and Machine Learning. And we had this realization that, hey, when you’re alone, all the decisions — all the good and important decisions and all the silly and small decisions — fall on you as an entrepreneur, and it was everything from, “Gee, where do we rent,” and what type of space we want, how much can we afford, to what salary we pay each other?” It’s a lot of responsibility, but it’s a lot of fun, too.

Another big change was financial. I was an executive at Trimble Navigation, had a good salary, good options, and well compensated. I started Blue River, and I started with no salary. So we didn’t have any money, but even more than that, I had been, as I was saying, studying my MBA for a year, so that means that I had not been receiving salary for a year, and not only that, but Trimble had given me some money to pay for school. Since I was not coming back to Trimble, I had to pay all that back, so I was starting in the hole. That was something that hit me pretty hard.

Louisa B-T: Yeah, I can imagine.

Jorge Heraud: And I remember probably when it hit me the hardest is when I was telling my awesome plan to my wife, and my wife was pregnant with my daughter, and I remember having that conversation, and she just said, “Can you say that a little bit slower and again for me. You’re planning on doing what?”

Louisa B-T: Wow, she must be very supportive.

Jorge Heraud: Yeah, she was incredibly supportive; it was a financially risky situation, but I knew it was going to a rewarding path, almost independently of what ended up happening. I was, of course, looking for a good outcome, but very cognizant that lots of startups, unfortunately, don’t make it all the way. But I knew it was going to be a lot of fun, and I had talked with a lot of entrepreneurs, both that had successful exits and ones that didn’t, and the common thing was that they all learned a lot and had a lot of fun doing it and were very proud of what they had done.

Louisa B-T: Yeah, and I think with agtech there’s usually a fantastic mission that founders will have when they’re launching their start-ups, and I know that in the past when we’ve talked, you’ve talked about sustainability and environmental awareness. Was that always your mission from the beginning? What really drove you, in the beginning, with founding the company.

Jorge Heraud: Yeah, yeah. That’s a great question. So, in the beginning, when we were still at school, we had started thinking about what areas we should focus on, and we went through this exercise of talking with lots of farmers. My cofounder comes from a farming family in Nebraska, so we did a big project, actually as part of a class, to talk with different farmers, understand their problems and what we could solve. We had decided pretty early on that weeding was an area that we wanted to focus on, and one of the reasons that made it very appealing was the environmental impact we could have.

Our machines, today, can reduce the amount of herbicide used by a factor of 10, which is a lot. About a billion pounds of herbicides are applied in the US, and we can reduce that by a factor of 10. That was very inspiring for us, and when we latched on to that, we really liked it, but not only that, there was a good financial return for the farmers using this technology. We thought it moved agriculture in the right direction. So we decided to name the company Blue River because herbicides, instead of landing on a weed, which is what they’re designed to kill, they end up on crops or in soil and then all these chemicals end up, ultimately, in a river, in a small waterway that ends up in a bigger waterway that ends up in a river or ends up in the water table. So the impact that we felt was going to be most noticeable is in rivers, and that’s why we decided to call ourselves Blue River, signifying a good, clean river.

Louisa B-T: So, weeding was the plan from the beginning. I was interested to know how different the business looks today from what you set it out to be, but it sounds like you remained on target with that vision?

Jorge Heraud: Yes, for the most part. However, we did take a little bit of a meandering course. We started with this vision of weeding, and I’ll tell you what, when we started the company in 2011, a lot of the techniques that now exist didn’t exist. For example, machine learning was in its infancy, this new technique called deep learning that everybody talks about and we use very highly didn’t even exist and hadn’t been invented. The computing power required for these computers required big desktop computers, and now everything runs in a GPU — GPUs didn’t exist as something that could be put in ag machinery equipment — so we decided, actually pretty quickly, that although we liked the area of weeding, the technology wasn’t quite there yet, so we decided to start, instead, on a problem that was a precursor to weeding with thinning, and we started in lettuce thinning, which was simpler.

Thinning is really a type of weeding, but it’s a simpler case where you’re just looking for spacing between plants, and that allowed us to use technology that did exist but still perfect our technology. It was a little bit of a deviation, but it allowed us to continue to move forward in the direction we wanted, and then, around two years ago, with deep learning, with mobile processors, mobile GPUs, with this better technology, better algorithms, we were able to say, “Okay, great. Now we’ve developed all these things, let’s put it into the problem we really want to solve,” which is this problem of weeding.

So, yeah, very happy. We ended up exactly where we had noticed the big opportunity was, but the path was an interesting, meandering one.

Louisa B-T: It sounds like you were very much at the forefront of what this technology could achieve. I mean, did you ever feel like maybe you were a bit too early when you had to meander towards thinning because the technology just didn’t exist for what you actually intended to do?

Jorge Heraud: That’s absolutely true. We had that conversation at least a couple times in the company “ are we too early, are we in the right spot?” And for sure the first time we had it was when we were starting the company and decided to go after thinning, and I can tell you, right now I tell it as a good story, and an obvious decision but I’ll confess, it felt a little bit like a deceit that we had to retreat to a simpler problem, because the technology was not ready and, gee, should we admit defeat and completely change what we were going to do? Or should we go after this simpler problem? It was a hard decision that I’m glad we made, but it for sure felt a little bit of a deceit at that point, but now, with the benefit of hindsight, it was a perfect decision.

I’ll tell you another time … When we started doing thinning, even then, when we had our fourth generation machines, which were these very advanced machines that are the ones that we’re still using in thinning. We came up with this very complicated machinery, and one of the issues we had is that, initially, reliability was a problem. We were using these very fast processes that needed power dissipation, and we were using a liquid-cooled system. For example, the fans kept breaking; the sprayers kept short-circuiting. We had a lot of problems, and again, at some point, I remember talking with my management team, “Hey are we trying to take a step too far?” The machines were breaking.

I remember talking it over, even with the board, and saying, “Hey, guys. You know what? The timeline we had laid out is something we’re not going to be able to meet. This problem is harder than what we thought,” and I remember one of our board members that had this phenomenal response: “Hey, this is how it should feel. What did you expect? You’re trying to work on something that is complicated and revolutionary to the world, and you expected it to be easy, too?” All startups go through this realization that solving things is more complicated than they thought.

I really remember that board meeting; I was a little bit nervous about what the board was going to say, but they were very supportive. You hit the nail on the head though when you asked if it felt a bit early because it did.

Louisa B-T: But you never felt that it was hopeless, that your startup was going to fail and you were going to close down?

Jorge Heraud: Well, yeah. Startups are fragile things; I’m not sure I felt quite that the startup was going to fail, but for sure felt, “Oh, crap. Did we make the wrong decision?”

It was a big call. For example, when I was talking with the board I didn’t know how the board was going to react, if they were going to say, “Okay, Jorge, thank you very much. We will find a CEO that can do that job,” or something like that. They just went and supported and me and said, “Just continue down the path.” We were lucky with really good partners along the way. They supported us really well on this path. I think part of that comes from the opportunity that we were going after being exciting.

Louisa B-T: Let’s talk a little bit about the board, then, and your partners and your investors. The amount of venture capital funding that’s going to agrifood technology startups is still just a couple of percent of the global VC funding market, compared to how much food and agriculture represents of global GDP, so it’s a challenging world out there for raising funds. How did you find it? I mean, you have an incredible lineup of venture capital investors, but I’m presuming it wasn’t easy?

Jorge Heraud: Yeah, it wasn’t easy. It takes time, but yeah, I’ve been very lucky with my investors. I think that they were very supportive and exactly the right ones. Maybe just to summarize them, we had some early investors, seed investors, friends and family type round, and we had a grant from the NSF that was really helpful, a SBIR grant. Then we had Coastal Adventures and Data Collective. They’re both big VCs thinking about big, important problems for the world, so they’re really good partners, and then the last round we got more into agriculture, so we got Pontifax Agtech that knows a lot about agriculture, and Syngenta and Monsanto, which are companies that are in agriculture. So, really good … If you think about the investors we had, really good technologies, big thinkers, big, bold VCs complemented with good agricultural knowledge.

It wasn’t easy, to your question; especially the first round. I remember at Series A, I probably got 25 or 30 “No”s before I got my first “Yes” from Khosla, and that takes grit. I think a lot of entrepreneurs hear about this, but living it is very different from hearing it, and to any new entrepreneur out there I think you need to be prepared for that: you’ll need grit and resistance, and you’ll have to hear that your baby’s ugly for 25 times before you get a good “Yes,” and I think that’s typical. It’s not that I was unlucky or whatnot, but I think it’s typical that you get a lot of “No”s and you just need to pick up yourself and learn from the meeting and learn from the questions and just go there until you find somebody that funds you.

I think what helps quite a bit is just understanding who you’re talking with and why they would be able to understand. For example, when I was pitching to Khosla it was all about the impact we could have and how big this could get. But maybe a different VC I would’ve used a very different pitch. Somebody that knows more about agriculture, it’s much more about getting into the details and the business model and why does this make sense and how does it fit with the rest of the business?

Louisa B-T: And when it came to this potential acquisition with John Deere, were all investors in favor of it?

Jorge Heraud: Yes, so it was a unanimous decision from the board, but of course during the negotiation process around what number we should accept, etc, there were quite a few opinions that I got from my board. We all had to, of course, have a long, long conversation about it. The first decision is, “Are we better off selling now or continuing to build and perhaps sell later, or do something else? Try to shoot for an IPO or some other, different exit?” So that was the first level of discussion, and not everybody agreed at least when we started that discussion.

At some point, after talking and looking at the possibilities and rationing for that, everybody ended up being in favor, but we did not start being all in favor that this was the correct thing to do. Neither did we have agreement on a reasonable dollar, so we started quite far apart, but through conversations over several months, and talking with Deere, we started understanding our options and whatnot, and we all ended up converging that yes, this was absolutely the right thing to do.

Louisa B-T: What would you say were a couple of the key things that got everyone comfortable with the acquisition? Because I imagine, for you, this ultimate decision to sell your baby while, of course, an exit is often the goal for an entrepreneur, it must be hard to let go there, so what do you think was the real, convincing point that made this easy to do in the end?

Jorge Heraud:  I think, for me, and probably for investors too, one of the things that resonated the most with us was after visiting John Deere to talk with their management and see their factories and what they were doing, and getting to know and maybe hear a little bit what they were thinking about us. When I came back from that trip I remember doing a debrief with the board, and one of the things I said is, “Hey, guys. This is really the best partner out there for us. If you think about what we’ve done, we’ve created these machines that do see and spray, right? We have built some prototypes, and we’ve tested them in cotton, and what we found is that customers just absolutely love them. This is the machine of their dreams – that’s literally what one of the customers told me: “I’ve been dreaming about a machine that works like that,” that reduces chemicals, that saves them a lot of money.”

This one particular farmer can go from $100,000 on over-the-top applied chemicals to about $10,000, so it’s a huge, huge change. So we knew that we had something really good, and we had built this prototype, and we have the technology working, but we knew that the next stage was going to be scale, and we knew from our experience in lettuce how hard it is to make a reliable, durable machine, and John Deere is the premiere company when it comes to building quality, complicated machines in agriculture that work day in and day out. Their slogan is, “Nothing runs like a Deere.” It is really their expertise. Then we’d have to manufacture it, and we’re in Silicon Valley, and it’s pretty obvious this is not where you manufacture big pieces of iron, while they are the number one manufacturer in the US of agriculture equipment, and they manufacture worldwide, so if you think about scaling that’s perfect. Then the third thing is distribution and support, and they have thousands of dealers, hundreds of thousands of customers that are all very used to buying and supporting this type of advanced technology, and I really can’t think about a better partner to design, manufacture, and distribute the product.

So, if you compare that with what we had … We had a group of engineers, one person in sales, and we had zero people in manufacturing. The next stage for us was gonna be to reinvent ourselves and turn from a company that is a great engineering company to a company that also can manufacture reliable equipment, distribute it, support it, and grow the company in areas outside our expertise. Or just partner with somebody, and we decided to do the latter.

Louisa B-T: Deere was an obvious acquirer for you. Did you have any other potential suitors? I mean, I presume you were still in touch with your friends and ex-colleagues at Trimble. Was there anyone else that was looking at you guys?

Jorge Heraud: Yeah, we talked with several companies, either for investment or potential exits. As I was saying, Deere was the one that was the best fit. But yeah, we considered both machinery/equipment manufacturers, technology providers, and even the chemical industry. Our technology has implications for all of them, so we ended up talking, at different levels of conversations, with several companies in those different areas.

Louisa B-T: Robotics startups are still relatively few and far between in agriculture, and it sounds like there’s quite a good reason for that in that the technology is just getting there for many of the functionalities that would be useful in agriculture. How have you seen that space change and develop? I’m sure that you’re very well aware of some other companies and startups in the space.

Jorge Heraud: Yeah, I love this space. I think robotics is going to have a huge impact on agriculture. If you think about agriculture, especially large-scale agriculture, it’s a factory. It’s a big outdoor factory, and it’s at the stage of the Industrial Revolution, where machines are starting to be used, and they are very repetitive like conveyor belts and steam engines were that just did the same thing. That’s the stage where agriculture is, where we have big machines that are just doing the same thing over and over.

It’s very mechanical in nature. There are starting to be some more electronics in there, but for the most part, it’s very mechanical machines that are used in agriculture. You think about a factory now; they are no longer mostly mechanical machines. It’s not only about conveyor belts at a factory right now. It’s all about robotics and automatic assembly and pick-and-place machines and welding and painting. Robots are doing a lot of modern manufacturing but that hasn’t had an impact in agriculture yet, and I think that that’s what’s coming. That’s what’s next. It doesn’t need to be R2D2 and C3PO walking through a field; it’s going to be much more about smart machines; machines that can sense the surroundings and act differentially. I encourage people that are thinking about robotics in agriculture to continue to think about that because I think that there’s plenty of things that need to be done.

Louisa B-T: Do you see a world where there will be a lot more choice for farmers? I heard recently that farmers are actually trading up more often and thinking about how the digital ag space on the software side is growing rapidly with increasing choice around farm management software programs, do you think that’s what will happen in robotics, or do you think a lot of this innovation will be absorbed by the bigger companies?

Jorge Heraud: Yeah, that’s a good question. I think what’s happening is that technology is starting to drive a lot of the purchases. I remember when I started in precision agriculture, you bought a tractor because of horsepower and how much it could pull, and now, when a farmer goes and buys a tractor and a piece of ag machine, it’s a lot more about what it can do that is beyond horsepower. Horsepower for sure is a headline number, but it’s, “Okay, how does this integrate with the rest of my equipment? Can it do variable rate, can it change the varieties?” If you look at the technology that is in a planter, it’s a lot, and now farmers are starting to realize that it’s really about the technology that comes with the tractors, not just the tractor. The technology is moving fast, so I think that it’s going to make for faster upgrade cycles. It’s going to be not quite like the cell phone that the next new model is something very attractive, but it’s not going to be as much about keeping the tractor for 10 years, because you will be less productive, less efficient, and not able to do the latest things, mostly because of technology, not because of the machine not lasting.

So I think it’s going be very driven by innovation. And there are very good business models that allow both big and small farmers to be able to benefit from this, everything from leasing programs to custom application, custom harvesting operations, but I do think there are some good alternatives to innovating the business model of agriculture to allow farmers to take advantage of the latest technology. But, for sure, using the latest technology is going to be very important for farmers.

Louisa B-T: Well, it’s really very exciting, and seeing your acquisition by John Deere was great for the industry at large, and I hope it does bring more entrepreneurs into the space. So, thank you so much for taking the time today. I really appreciate it.

Jorge Heraud: Thank you. Maybe one last thing I’ll say Louisa is that I’ve been in this industry for a while, both at Trimble and then as an entrepreneur, and when I started entrepreneuring, entrepreneuring in agtech was a weird thing. Nobody was doing it. I remember it was us and there was this company called Solum, which is now part of Granular, and there was nobody else in agtech, so we had to explain this to every single investor and what it is that we did and what is agriculture, and why did it make sense. That has completely changed now. Now every investor knows about agriculture, even if not everybody’s investing yet, everybody knows. One of the things I credit for that is AgFunder; you guys have been phenomenal in educating the market, both the VCs and the entrepreneurs on what the opportunities are, and I’m really thankful for the help that you guys are bringing to the space.

Louisa B-T: I’m glad to hear it and hope we continue to be a useful resource.

Jorge Heraud: Yeah, I’m looking forward to that.

Join the Newsletter

Get the latest news & research from AFN and AgFunder in your inbox.

Join the Newsletter
Get the latest news and research from AFN & AgFunder in your inbox.

Follow us:

Join Newsletter