S11E143 Elmar Mair/Neatleaf - AI Crop Management Powered by Neatleaf’s Spider

Harry Duran

0:00 - 0:05

So Elmar Mayer, CEO and founder of Netleaf, thank you so much for joining me on the vertical farming podcast.

Elmar Mair

0:06 - 0:08

Thank you so much for having me. It's a pleasure to be here.

Harry Duran

0:10 - 0:16Ignore this for outputs

So for folks that are not familiar with Netleaf, what part of the world are you calling in from today?

Elmar Mair

0:17 - 0:37

So Netleaf is, yeah, we are basically located in Scotts Valley in Santa Cruz in California, and that's where our headquarter is. But we do have to offices, the other office is in Munich in Germany, and yeah, currently targeting the us market, but we are about to deploy in Germany next week. So exciting.

Harry Duran

0:38 - 0:41

Okay. Where did you grow up?

Elmar Mair

0:41 - 0:48

Oh, I'm originally from the italian countryside, all the way in the north, in the mountains, the Dolomites, so.

Harry Duran

0:48 - 1:07

Oh wow. How did you get an interest, make our way into vertical farming, obviously, but it looked like from a lot of your background you had an interest in AI early on. And I'm curious if you thought that this is where you would end up when you're thinking about where you're going to end up after university studies.

Elmar Mair

1:07 - 2:14

Definitely not. No one really leaves that area. It's like the Alps, the mountains, it's all very enclosed and all my friends are still there from early on. But somehow my trajectory was just different. I fell in love with automation, with robotics, AI, and ended up studying that, working at space. And then I was always intrigued to leave and live in a country nearby the ocean. And so I got attracted to California and no one told me how cold the water here is coming from the mediterranean sea, that other change and yeah, and so yeah, now I worked in different companies. My last gig was at Google X, the everyday robot project. You had like this fleet of robots moving around with armst, very surreal. But then I learned about the agriculture and the fourth agricultural revolution. And somehow it felt like closing that loop, you know, coming from the countryside. Spending my summers at my artist farm and now again working in that space just feels like, yeah, I just learned everything on my trajectory and now I can apply it and come back home.

Harry Duran

2:15 - 2:20

What's your earliest recollection of your sort of love affair with technology?

Elmar Mair

2:21 - 2:58

I think really when the first time at the university, I was intrigued about computer science in general, like just the language, learning that, being able to talk to a machine. But then at university, the first time I actually programmed an embedded system, so I basically had impact in the physical world with my programming. That's where I realized, wow, this is super exciting. You know, you can distort like audio and make, you know like guitar distortion device or something like that. And it was one of the exercises. And then the first time I saw something moving around, that's really where I was like, I want to do this. I want to move things in the world and talk to these machines.

Harry Duran

2:59 - 3:33

And so as you were moving up in your career and you had the opportunity, I saw that you had some experience with autonomous vehicles at Lucid. And then obviously you mentioned that you're working at Google as well. Is that something that you had really taken a fascination towards in terms of understanding AI? And I? Obviously, learning technology and computers is one thing, and programming them is one aspect of it. But then when you start to get into AI, then you're really working with a different level of intelligence at that point. And I'm curious how you started to become more aware of the possibilities of AI and what was pulling you towards that.

Elmar Mair

3:34 - 4:51

Definitely. I think that, I mean, look at it, for example, the language, like for us, humans understanding languages, learning languages, the semantics of the world, what is each meaning, what's related and all that. And seeing now that AI can pick up on that and really can grasp that space and do a better job than humans can do in that sense, it's really fascinating. And again, it's to some extent like black magic, because a lot of what happens there is like empirical, again, like, it's started off obviously with a lot of mathematical and theoretical statistical foundation. But then to train those things, it's a lot of trying and understanding and grasping what data you need and how you have to formulate things, what works, what doesn't. So you have this black box which you push data in and it becomes smarter, and you try to make it as smart as possible, and it can pick up on things where humans are incredibly challenged, or it's a hard time for humans to actually do certain aspects, but the AI can understand these relationships and be very good at predicting things where humans struggle, seeing cancer in images where humans are challenged to kind of like, figure this out. And that's fascinating what you can do.

Harry Duran

4:52 - 5:04

And so what was your experience like working on the everyday robots project at X? You know, you spent some, quite a bit of time there. What would you say were some of your big takeaways or maybe your aha's from working on a project of that scale?

Elmar Mair

5:05 - 6:12

Probably it's just a culmination of, like, what I learned from my career working on all these complex robotic systems like self driving cars, drones, human lights. It's just to understand that the more the complexity of embodiment, like a machine moving around, interacting with the world, it's incredibly challenging. It's like you can use AI to do certain aspects, but to bring it all together, it's like playing a puzzle and bringing together all these pieces and still trying to make it work, understanding what matches what doesn't, it's really fascinating. So, yeah, for me, the insight was really the one thing you did, the power of different viewpoints. Like the robot tried to approach a bottle and grab the bottle and just stared at the bottle and often missed and just wasn't able until we moved ahead. And you get a different perspective and you get a better understanding of the world and the object itself and the relationships. And that was one of the learnings which we then also adopted at Neatleaf as we developed a system there.

Harry Duran

6:13 - 6:36

I'm excited to get into the origin stories of neat leaf, but I'm always curious what people's relationship is to food, because obviously we talk about vertical farming, we're growing food in these farms, and obviously because of the background you had growing up in such a beautiful countryside. I'm wondering if you can think about your upbringing. What was your relationship to food or fresh food or access to fresh food?

Elmar Mair

6:37 - 7:42

Yeah, there's one thing which I would call out if people say, like, why would I visit ever my home region? I would say it's food. It's just this mix of cultures where you have like the delinquent cuisine, the Austrians with german influence, plus their own alpine cuisine. So, like, you have this all very homemade dishes and I mean, it's just a cultural thing. You just like talk a lot about food. And food quality has an incredibly high standard. Spending money for food because it's good. You know, you don't buy the cheapest product, but you buy the one which is high quality or like good quality. So I think those things are definitely part of my thinking. And, yeah, so I coming back to that space again and really understanding the importance of being able to grow in a consistent way, in a reliable way and leverage, you know, like controlled environment, agriculture, greenhouses to afford, like, to accommodate that, grow nearby urban centers, not having to ship the product across continents, which is not sustainable at all. Those things are just definitely, which I 100% believe in, that this has to happen and it's the future.

Harry Duran

7:43 - 8:35

What's been the reaction to friends and family as you've moved careers and now you're back and you're working into with food and growing food? I tell my story a lot here about as I got into this podcast and I started learning about everything that was happening I was paying more attention to the grocery stores. I was actually turning over the packages, seeing where my lettuce comes from. 3000 miles away, New York. From Salinas Valley to New York, you have to travel and the lettuce is good for maybe two days before it starts turning Brown. Maybe we need to look locally. I'm here in Minnesota now. Now there's revel greens and that's almost all that we buy now because it's greenhouse and so. But people's habits need to change. And I think it starts with education, awareness that there's a better choice. And I'm curious, you know, if you've had to explain some of what you do to friends and family in a way that's maybe enlightening for them.

Elmar Mair

8:35 - 9:51

Yeah, it's very interesting in a sense that the greenhouse technology itself is still pretty young. You know, like the concept of growing a greenhouse and travel does, like aerobonics. You know, it's even like going beyond that. It comes originally. Originally, like, a lot of it comes from the Netherlands. That's where, you know, that's a whole other story, how that evolved. But, like, it's really impressive if you drive through some of these towns, like vassal land, etcetera, but it's like packed with glass houses and they're all very profitable, making, you know, lucrative. So it's the future and it hasn't really reached a lot of the areas. Also, the US is still picking up on that. Now. The beauty with that is that the more we get technology to really improve the outcome, the yield, the result in these facilities, the more we're going to see the acceleration of such facilities, then you're going to save all the traveling across the continent to get from Salinas to New York. And 60% of the letters you buy there is just gas money in Japan. But so for Albina, talk about it, we don't have many greenhouses in the area where I come from because. Not that established yet, but I think everyone kind of is intrigued by it. It's like, oh, okay. It's just, I think it's a lack of knowledge and understanding of what you can do in these greenhouses and how you operate them, but it's going to happen.

Harry Duran

9:52 - 10:13

So I saw that you started neat leaf in late 2020, which is a very challenging time to start a company. So I'm curious about the origin story. How did that start and then take us into that mindset, maybe even pre 2020? And as you were thinking about the idea for what would become neat leaf, what was happening in your world.

Elmar Mair

10:14 - 12:41

I think it's a mix of curiosity, hype and naivety. I was very intrigued. Again, I really enjoyed working in small teams like lucid motors. We got 100 people building a cardinal, like, mind blowing, it's nuts. And it was very exciting though, with the right people. I was so impressed what you can actually do with the right set of people and coming from a big company like Google, again, like a real zillion treaty, kind of like you want to work and again in a small environment and heard about, you know, agriculture, like fourth agricultural revolution, data driven automated cultivation, and I was kind of intrigued. That sounds totally up my alley and want to learn more about that. And as I then looked at that space where you have like a greenhouse which is fully controlled, I went to visit some universities. The professor was pointing at the sensor dangling in the space and like, oh, this is what controls you, everything, and just move it to the right, to the sweet spot, which is the average. And I just looked at it coming from, you know, the everyday robot project. You can do so much with data, but this thing is not going to cut it. And it just hit me, I was like, in any other industry, if you think about it, automotive production lines or any other processes, production processes, the first thing you do is you try to capture the information that's happening in that automatically. Your sensors to account stuff, quality insurance, troubleshooting, it's all trying. Based on census and automation, we have been doing that since the fifties. And it came from Japan, where they tried to withdraw minimum fuel resources. They had to optimize the production. And in cultivation, one of the biggest markets and the oldest markets, we rely on humans, humans to walk these aisles, look at the crop, understand what's going on, assess the situation, interpret the language of the plants, and it's such a complex space. You have all these factors, temperature, humidity, CO2, lighting, irrigation, nutrients, which are all correlated and impacting each other and fighting each other. And again, the human goes in there and subjectively assesses the situation. Has to remember like, oh, it looked the same yesterday or last year. We had the same situation, it looked different, it's not where it needs to be and how it should be. And I was just like, why is that the case? It's a huge market. Why don't we have the same concept technology as we have in any other industry? And that's what triggered native, being naive about it in that sense, like, what's going on?

Harry Duran

12:41 - 12:57

And so what were those early challenges or those early pain points that you were looking to solve for farms did you have an idea about who an ideal client would be back then? Or is that something where you thought you had a concept for what the product would be and then kind of see who the audience might be for that?

Elmar Mair

12:57 - 14:25

Yeah, I think the first step was to understand the problem at hand. Like, why do we rely on humans and can we automatically collect that data? How could it work? The first conversation I had with Agfunger back then with Rob Lerker, I still remember it very well. He said, like Elmar, Agtech is not a technology problem, it's an adoption problem. That's obviously a huge gap, but how do you get people to adopt it? It's kind of stuck in my brain of like, okay, whatever we gonna do to make this problem, to solve this problem, it has to be low cost, it has to be easy to adopt, transparent to the operation, like anything which you can do that people just, you know, try it out and do it. So we started with like single sensor boxes. Within a month of like, inception of the company, we basically had like single sensor box it systems in the space, recording images and taking measurements of the environmentals. And we realized basically a lesson from the everyday robot project, that a static view of that blend is not going to cut it. If you want to apply machine learning AI, you really need different viewpoints of that blend, of be able to adopt reliable machine learned models and not have to rely on humans anymore. And that's very okay. We have to have a system which allows us to move that camera around and look at the plant from different views as a human does. A human doesn't go in there and looks at the plant like, oh, this plant is stressed. And they keep staring at the same plant from distance, but it goes there and moves around, checks out the different spots of that plant, and that's how they understand really what's going on.

Harry Duran

14:26 - 14:29

So is this your first time leading a company as CEO?

Elmar Mair

14:29 - 14:31

Right? Yes, it is.

Harry Duran

14:32 - 14:59

So we've had several first time CEO's on the show. And actually because it's a new industry and we have people coming either from other industries or in that startup mindset, or having had experience in adjacent industries such as AI and robotics. So who are some voices or mentors that were helpful for you in this process of first time CEO understanding? What is it that I don't know about leading and starting a company for the first time?

Elmar Mair

14:59 - 16:18

I think that's where I would highlight the Navy dealer. I mean, I think it's, I mean, at least for me, I definitely did not know what it means what it takes to start a company and then also like what it takes to start something as, like, neatly with the complexity we have where you have to do hardware, software, firmware, electronics, cloud infrastructure, AI blend, science, and you have to web development, you have to bring it all together to actually make it work. And so that's really where. Yeah. What's the challenge? And so basically I had a network of people, like other founders, they gave us a lot of guidance. And you have investors, then you talk to them. So you start building your network in that space and it's really, really valuable. There's so much you don't know and so much you need to understand. At the same time, everyone keeps telling you, like, slightly different things. Yeah, to some extent. You just have to make, yeah, learn your own lessons and they're painful and you're gonna get it. But it's, there's so many different trajectories a company can evolve and can grow that it's hard. There's no one path you have to take. It's all about you following the right concepts, the right strategies as you navigate that crazy roller coaster which every company is.

Harry Duran

16:18 - 16:25

So was the intention to get it funded from day one in order to do the work at the scale that you're looking to do with Neatleaf?

Elmar Mair

16:25 - 17:13

Yeah, so we self funded it originally just to kind of give us the time to understand that space first so that we know what do one actually want to do to approach the investors? We started having investor conversations which, like talking to like, Agfunder, for example, which understand that space. And they gave us feedback on, like, this doesn't make sense. And this would make more sense. And then at some point, obviously, we also deployed these prototypes and we came up with the Spider system, this cable based robot. And at that point, just the reaction was so different. Every time before we have other ideas, we've said, okay, clients, would you like to try this out? And they was like, yeah, sure. And which basically was like, I don't care. But that the reaction with the spider system, like, oh, my God, I want this. I need this. I want to be first. And that was kind of like, that's a different reaction.

Harry Duran

17:14 - 17:15

Yeah, yeah, for sure.

Elmar Mair

17:15 - 17:16

So it's like, okay, this is it.

Harry Duran

17:16 - 17:48

Yeah, well, I imagine that might have something to do because you can sense that it may be a testament to some of the guidance you had early on in terms of advisors or your board. But just like everything that I've seen so far around neat leaf in terms of even like, the press, you've been getting over the years. You know, there's a concerted effort to give you the proper visibility, and I imagine you made sure that you had something, a good story to tell before you start turning that on. Right. Because obviously, you know, if you're not ready, you don't want attention. And it seems like you tell the story only once.

Elmar Mair

17:48 - 18:36

It's like, you know, if they don't like it, you have to wait quite some time before you get another chance to tell another story. And you've definitely got earlier. If you wait until you think you have the perfect thing, it's too late. So you have to go out there with something, which is. Which embarrasses you, basically. And I mean, the first system in the field was 2021, December. So nine months after God had a spider, we kind of built this prototype super small team and came out, put it in the field. We were all super proud, but looking back, it was incredibly embarrassing of what we did, obviously, but very scrappy, very. Just getting the concept out and working with clients who are believers who got the vision and the mission, and you work with them. And then obviously, yeah. As you evolve and iterate and improve, then you can approach more and more clients and you have a story to tell about that. Yeah.

Harry Duran

18:37 - 18:38

Who came up with the name Spider?

Elmar Mair

18:38 - 18:50

I think it was, when we looked at it, it was just like, okay, it's like the 4k or this thing in the middle. It looks like something moving around in this web. And so we just adopted that name. And I think it. Yeah, it suits it well.

Harry Duran

18:51 - 19:39

Yeah. And it's one of those sort of like things that it almost becomes like a brand identity because it's when people see it in action, it almost looks like a spider that sits and spinning its web around the farm. Yeah. And so as you were thinking about the different, all the different moving parts that are involved in starting a company, who did. How do you think about this idea? I like this idea of like, getting people to talk about how they think about thinking. So you have to build a team. Right. And you have so many different moving parts. You mentioned R and D and robotics and, you know, marketing and staffing up a team. So how do you think about where and when to deploy and who should be your first hires? And so I'm curious about that thought process, and I love having this conversation with first time CEO's because it's helpful for listeners who are in that same boat.

Elmar Mair

19:39 - 20:58

Yeah, I think that's always the biggest challenge for startups is recruiting, being able to hire or convince great talent to join you on your mission, knowing that you can't pay them. The seller is probably which they're gonna get a in other more established companies. And for that, yes. You need to feel like you need to find like minded people. I mean, obviously you have equity which you can, you know, share the share, have them part on, like the benefit once you're successful, and. But it's really about, you don't need many people. It's about finding the right people, which means they have to be experienced, but also versatile, so you won't be able to fill all the gaps. You need to have people who can cover multiple bases and really ensure that it's not just very narrow minded engineering skill, but more broader skills. And it's obviously a balance because the more senior you become, you often become more abstracted away from the hands on work, or you are very into the details. So finding these people is definitely not easy. And we got very lucky being able to recruit some friends from our, my network before which we are just excited about the mission I joined. I think that's the key. Early investors, early stage investors, they always invest in the team. So I think that makes a lot of sense.

Harry Duran

20:58 - 21:02

So how have you grown as a leader since starting the company to current day?

Elmar Mair

21:02 - 22:14

Well, first of all, I didn't have gray hair when I started the company in many ways, but no, I think it's. I was just thinking about that. I think it was the most intense learning experience over the last four years for me, that's worth way more than any other university or education or experience I had, in the sense I was just exposed to everything in a company. And your faith or your future is on a very thin flat. Early on, everything has such a huge impact. You get this investor that is jump off. Does the client like it? No. Buy into it. Everything is, you know, you have kind of the best team, the best idea and all of that, but you still need luck. And that's the one thing which, you know, people have to consider as a starter startup. They should not go in there and say, like, oh, I want to be successful because the chance for that is very small. I think that you have to go in there. I believe in this mission. I believe that this is, I spent my time here on something valuable, which can be very successful as well. But, you know, that has meaning, and that should be the thing which drives you. Then it's worthwhile independent of what, whether you have luck or bad luck.

Harry Duran

22:15 - 22:42

You mentioned mission and meaning, and I'm curious what those words mean. For you personally. And a lot of times I can see this in the conversations that I have here, that these founders and these CEO's are putting really their passion into this project because it's really something that they believe in and they feel like it can change the world. And I'm curious, as you think about this and communicating that vision to your team, when you think about those words, what does that mean for you personally?

Elmar Mair

22:42 - 23:40

I think that's the foundation of everything. We would not be here without that, and we would not keep going without that. That's what drives everyone. And part of that is really the first part of the company. Like, when you develop the product, it's all about your beliefs, your mission, and then it's all about the customer feedback. And if we wouldn't get like this, exceed this excitement in that space and, like, enthusiasm about what we're doing, that's, you know, that's really what drives all of us to go above and beyond that, you know, like, put in all these hours. But it's really, you know, we started off like, well, we can't continue growing as we do. We need to change that. Why does it no one rely no data, as they call. Why do they run around figuring out what's going on? They should just deal with the issues. They shouldn't worry about what's happening. And so, yeah, believing that this system is really going to help or replace that burn and help them, that's what convinced everyone to give it all.

Harry Duran

23:40 - 23:43

What's a tough question you've had to ask yourself recently?

Elmar Mair

23:44 - 24:25

I think the same question which I keep asking myself since the beginning. What did I miss? Why wouldn't it work? And that's what you just go in there, okay, this is my mission is by believe. And then you ask yourself, okay, what did I miss? Why didn't it work? Why hasn't no one else done that yet? And then you can't just keep trying to validate and prove it. Every time you get like there's a hurdle there or something going on, you ask, okay, what? Why would it, why is it not working? Why doesn't it work? Why wouldn't it work? Is it just a hurdle? Did I miss something? Is a fundamental issue. And I think those are the things which I think also healthy to ask yourself, because at the end of the day, you don't want to waste your time or investors money, if you know that, or if you have doubts in something.

Harry Duran

24:26 - 24:38

So for the benefit of the listener who's just now learning about neat leaf for the first time. How would you explain what it is that neat leaf does? And who would make an ideal partner for you?

Elmar Mair

24:38 - 27:51

Right? So Neatleaf developed the neat spider, which is a cable based robot, like a football stadium camera for greenhouses or indoor facilities. So your dutch, we have, like four corner pieces in the corner of the greenhouse or in the bay, and then four ropes go to a center piece, which is like little box packed with sensors where they measure all environmental conditions, like temperature, humidity, cosmic. Do we measure plant height? You get the growth rate. You measure leaf temperature for leaf VPD, but also, like, transpiration in general. Like, you know, how much the plant is growing, how much it's processing, how much nutrients it's sucking up for, like, liquids. It's. No one has access to this information. We measure the light intensity on the back, you understand the daylight, integral, etcetera. And then we have cameras, RGB cameras, NDI cameras, and we measure that. We detect every yellow leaf, every necrotic leaf, every foliage, bass spout and mildew. And the system moves around the canopy 24/7 on top of it and just monitors your crop and tells you if something is going on or something is going south or even an issue and quantifies everything. You can now dock numbers. Instead of saying, like, it looks the same as yesterday or it looks worse than yesterday, and just start arguing with your coworkers. It's about. No, I have 3% more yellow leaves than yesterday. I have 5% less folding than yesterday. It's the same size than day 22 than last time we grew it a year ago. It's numbers. It takes away all that friction, and it's just also communication to management. It just goes away. That's how every other production line works. And in cultivation, we don't have that yet. And so what we are facing a little bit is, like, the smartphone problem. When smartphone came out, I was like, I don't need that. I have a computer. Check my emails, there, a cell phone. Why would I need a smartphone? But once you get a smartphone, it's like, oh, my God. Can't do without it. What did I do before? And so that's really where the system is the same thing. It's like people at the beginning are like, well, I have been growing this way many, you know, since now, like, forever. Why would I need this thing? And once they get this information, this data, it's just a really game changer. You can loop in experts remotely. They don't have to fly in anymore. You don't have to snap a picture and send it no, they can log in, go back in time, see what happened. All the conditions really understand what's happening. You have a burned leaf, which you have a certain pattern which you want to figure out, why is it that way? What happened there? You can go back in time, see when did it start, where the conditions and all that. So it's, yeah, that's how you want to operate. You can actually troubleshoot, you can understand, you can improve. And we talk about an industry where you have, depending on the crop, ten to 20% percent crop loss every cycle, ten to 20%. No other industry would allow for such a loss. It's just because we all accept it because like, well, it's just hard. And cultivation is freaking hard and you have to nail it every freaking day. And it's such a crazy burden and such a crazy risk. Culture are heroes in my mind, but just because they don't have the tool. And now that's why they really love the tool, because now basically they have a second pair of eyes more actually looking at the crop and constantly monitoring it and telling them, hey, something is going on. You need to deal here, need to deal with that now. And so they don't have to constantly pull and like scout and figure out which is, yeah, it just changes the whole way how you do cultivation.

Harry Duran

27:52 - 28:54

So that term, right, cultivators, I feel that term is evolving now over time because, you know, maybe 1020 years ago, if you're just working in a greenhouse and we don't have these technologies at our disposal, you rely on cultivators who have been around the block, so to speak, and know their crops, and they've been working maybe ten or 15 years. I remember someone saying once that they analyzed, like, the brains of the taxi drivers in London years ago. And because there's no grids, like in New York City, it's grids, right? You can quickly figure your way out. But in London, like, the streets are crazy and the cab drivers, you know, they somehow would map the entire London city in their mind. And that's their job, right? They have to figure this out. And I think about cultivators having to learn this, like, through doing it years and years, making mistakes and figuring it out. But now it feels like there's so much information and so many variables that they almost have to learn and use technology and partner with technology to become a better cultivator. So it feels like that's evolving over time.

Elmar Mair

28:54 - 30:37

I think this is a wonderful analogy, and just think about it. And now you take that London cab driver and say, no, you actually now have to drive in whatever Stockholm. And, you know, obviously, they know the concept of driving. They know what they have to remember, but they start from scratch, understanding Stockholm in the city and the problems there, and there is special situations, and the same is with cultivation. If a cultivator leaves, first of all, that knowledge is gone. That operator of that facility starts over trying to educate someone or teach someone, like, oh, they have to make that experience. What happened in the last year is gone. You know, it's not overdevelop. And that cultivator starts over in a new facility. Okay, how does this facility work now? Like, what are the kinks and the issues in the area? Like, how do I deal with that? And that's where, again, having a system which captures this knowledge and this information is really key to maintain that knowledge. And neatly, at the end of the day, is a data play. If you think about it, it's like we have now a system. We have access to a unique data set which no one else has. We can obviously offer down the road services which no one else has access can offer right now, you know, and again, one is maintaining that knowledge. The other one is really understanding microclimates. No one runs around. You can't measure microclimates. It's, like, incredibly hard. And. But our system just moves around that measures them. And you can add an understanding of whether your airflow is correct. The humidifier is failing. Your h vac needs to be a replaced filter. Issues which, in general, you realize once the blends are struggling and you have, like, an uneven candle beam, like, what's going on here, and you start trying to figure out the damage is done, you want to realize those things before you see the damage to the plant as soon as possible. And, yeah, and then you can really deal with them. And, yeah, that's really how it needs to be, in my mind.

Harry Duran

30:38 - 31:00

Do you find that farmers, when they start working with the neat leaf system, are surprised by the difference in microclimates that might exist within their own greenhouse or in their own farm, and not realizing, like, they're thinking this is just one climate that's happening within the whole farm, but because the spider is making its way across the entire farm, it's pulling in these data points that's showing you that there's some discrepancies there.

Elmar Mair

31:00 - 32:17

There's a lot of learning happening every time. It's really exciting on how, you know, like, really understanding the. Not just the microclimates, but also a lot of the growth rate. Oh, if all the plants are gonna just stretch differently. No, they're actually just, like, growing longer and, like, understanding those things, but being able to measure them and being able to do quick a b tests. Oh, I want to try this one substrate versus the other one, these nutrients recipe versus that one. You have all the values there, you know, that we pick up on flowers, flower sizes, the crop itself. We can predict yields. You can really give you the whole picture of the plant as it evolves. And so you can easily compare, you can have reference runs. Okay, this is my reference, and now I just compare it to that. 01:00 a.m. i. Better than last time. Am I worse? How am I trending? No one has this information right now. And I know in this big cultivation operations that one of the biggest challenges is knowing how much you're going to be able to sell. So, you know, sales orders, products like, we're going to need this and whatever. Three months, you need product a and like 30% and product b, 50%, and then after three months, they're going to get like 60% a, 40% b or whatever. These are the challenges these operations have to deal with. So being more precise, more accurate in these predictions, will change a lot on how they can monetize on it.

Harry Duran

32:17 - 32:43

Was there a conscious effort to do an outreach towards cannabis growers, or were they naturally attracted to neat leaf? Because obviously, because of the potential for what's possible in terms of the size of their operations, because I saw that you're even working with cannabis genetics companies. So it seems like you're making great strides in that industry. And I'm wondering if there's something specific about some of the challenges that they're having or typically have historically had where neatleaf is helping them.

Elmar Mair

32:43 - 34:25

Right? So the technology itself is crop agnostic. So the spider system can operate in any greenhouse. When we develop, like, spiders, what markets are we going to target? And the way, how we thought about is, like, in any other technology development setting, you're like, okay, well, what are the highest margin markets? Like? Tesla started with the roadster and the Model S, and then the model three hints started. Model three, obviously. And so that's how you start developing a technology, because that's where you have higher margins. You can justify higher cost. As you scale, you bring the cost down and it becomes, you know, add more bells and diesel, it just all works out. So cannabis, besides being probably the largest margin market in horticulture, it has other interesting advantages. It's one of the fastest growing plants on land. So you have a lot of growth cycle of turnover, which allows you to quickly iterate. The second thing is, it's a new plant, which we did a little research on because it was illegal and so on, the history of it, and so people don't really know how to grow it at scale. It's new. So having access to data to really help them understand the plant better and optimize the processes and the performance, it's really key. The third thing is, for compliance reasons, since you have this regulatory constraints around it, you have to report all the yield on the potency, the wet mask, dry mask, biomass, etcetera. So you have all that information which is going to be your cost function to optimize for. You get it already out of the box, you don't have to explicitly measure it. And so it's an interesting plan to start with, but we're already operating outside of cannabis. We have systems in berry nurseries, and we are talking to ornamentals and other places, and there is a lot of excitement to use that system.

Harry Duran

34:26 - 34:41

Where are you seeing the most engagement or excitement in terms of the possibility for changing how they're currently growing from farmers, whether it's a specific crop, or people that are measuring more than they were able to before?

Elmar Mair

34:41 - 36:24

I think the biggest excitement is different aspects which are very valuable for various sizes of operations. I think the bigger operation, the more it makes sense to standardize things and have automated reporting and have access to that data, access to forecasting, being able to have remote access, so that you can actually see what's happening and things like that. So I think for multistate operators, bigger, no operators, this tool is just makes sense. You do it in any other domain, in retail, in logistics, whatever, you always have tools which monitor, you know, where things are. You can always look it up. And then when it comes to cultivation, you have to wait for someone to send you like a handwritten report at some point with a certain delay. It's your mind boggling. But now we have automated reports, we have an interface where you see the plans, you can tag each other, you can highlight certain areas, you can comment, you have a whole archive of what happened. That's how cultivation has to be. In a sense, that whole like adopting what we do in other domains. So that's the value add. If it comes to size, plans wise, I think all the plants have the same challenges. There's pest pressure, there's the unforeseen environmental conditions, equipment failures, just consistency in your operation and understanding, optimizing for certain genetics. If you change your genetics, you start over again to find you and optimize things, and that takes a lot of time. We had now virus in tomatoes, which they had to start over with their genetics, and it's going to take years for them to optimize it again. For them, data is key, because then they can accelerate that and get to that same performance as before. We quicker.

Harry Duran

36:24 - 36:59

So, Emma, where are you seeing the most potential for innovation in this space? And maybe it speaks to something you may or may not be able to talk about in terms of your roadmap. But as you think about the future and this field of AI is constantly innovating, even if you just track what's happening with the chat dpts of the world and the clouds of the world, it seems like there's an update coming out every week. It's pretty amazing how fast things are moving there. When it comes to machine learning AI and making all this data available for farmers. Where do you see this going?

Elmar Mair

37:00 - 38:59

I think the obvious destination for this is that we more and more automate that whole process. Take the human out of the loop out of that first. It's like taking it out of the monitoring loop. They should not have to look at stuff and understand what's going on. The second thing is taking them out of their controls loop. You know, like, you don't want to have them choose what's best and optimize it, but start suggesting certain changes and then optimizing things, because it is such a complex space. You have all these parameters, and humans are just, like, not great at that. And I remember when doctor son called, delay was like, yeah, how do you, like, assess now whether this is working well with the environment or something is going on? I just go in there and look at the plans, and then you feel it. Something is off. And I was like, you feel it? But it is a feeling because your brain doesn't even perceive the cues anymore. There's so many different things going on. Your brain just does. Like, something doesn't match up, it just feels off. You might not know yet what the cue is, but that just means that those are the limitations of humans, where you can't also, like, look at all these plants, which our system can look once it's deployed all over the place. You know, it's the same with x rays. A doctor won't be able to see that many, as many x rays as the neural net could see when it was trained, based on all the data of all the hospitals all over the world. So it's just. Yeah, it just makes sense. And so having that system now understanding better how to optimize for certain genetics, for certain cultivars, that's really what I think the future is. There's still a lot of, you know, human labor involved. The job of a cultivator might change. It's less of really going in there and touching the plants and understanding and guessing and being stressed out, but it's more about making the right calls. Based on this genetics, there's data, this might be going on, this is how we optimize. You might want to grow these strains together and the other one's there and stuff like that. And it's. Anyway, it's a gradual transition, not a binary flip.

Harry Duran

38:59 - 39:53

Yeah. When we talk about medicine as well, and the support that AI is providing as an example of a related field, you know, a lot of the, I think the third leading cause of deaths in the US when it comes to medicine is doctor error. And it's, you know, it's. If you're working twelve hour, 16 hours shifts, you know, you get tired and, you know, you could be the smartest, most intelligent, high Q person in the world, but if you, you know, you're working 18 hours straight, you know, you're gonna get tired. And I imagine in this field too, like, if people are working, teams are stretched out, and I, you know, they've got limited resources, big farms with only a couple people to see. You know, there are people, some who have been doing this long enough that they can intuitively tell something. But at some point, you know, you really do need some help if you're really gonna take this to the next level. In terms of tracking everything that's available and all the different things that you monitor and all the different systems that you monitor as well, 100%.

Elmar Mair

39:53 - 41:01

I mean, human error is just unavoidable. It's just a nature of things. And we have so many times we have seen that there's, the irrigation nozzle isn't plugged in or something was not managed properly. And it just happens. This can be communication, it can be like being exhausted diet, all of it. And so that we have to accept it. Unless we are automated, the one thing we need we should not accept is that we don't detect these issues right away. And that's the other thing, because some of it is also not human error. It's just like equipment failure. But being able to pick up on those problems, that irrigation doesn't work and you know it right away, instead of like two days later when all the plants are dead, that's really the key. And we have seen issues where, you know, like it always happens on a Friday or before Christmas. It's mine, it's mindful, it's a curse. But then, yeah, it's just a team is a short theme. It's not, maybe not the a Team. And so, yeah, they just missed something. And the damage is huge. No, there's thin margins in cultivation and so if you have some yield loss, you're not making profit anymore. And so that's where it's such a struggle and such a challenge.

Harry Duran

41:01 - 41:28

Do you see opportunities to partner with other companies that are doing similar services for the farms? I'm thinking about robotics, obviously. If you think about, you've got the monitoring, you've got the AI and the machine learning, are there maybe opportunities to hand it off at some point to robotics to say, hey, it's ready to pick or information that you're gathering and now becomes available to other systems to act on?

Elmar Mair

41:28 - 42:23

Right. I think the way how we look at neat leave is that we have two pillars. One is the robotics, the automation. That's basically what we call a data harvester data monitoring system. And the second one is the data processing, the AI, the dashboard to another show like the results, the reporting, notifications. Now the second pillar is totally agnostic to the first one. The data can come from anywhere. So if there's already like a system in there or something which collects data, you can process it. You can slap a platform like this on a drone, like sensor platform on a drone, fly around outdoors, or, you know, any other system you want to use, or handheld devices, things like that. And obviously on the automation side we can do more. We could for example, say like now we add a little spray nozzle and detect powder, mildew and start spraying and keep going and minimizing the pesticide usage, be very reactive instead of preventive and minimize exposure of humans to these dangerous chemicals.

Harry Duran

42:24 - 42:33

I'm curious, with the applications you've had in greenhouses, have you had conversations or have you started to look at implementation specifically within vertical farms as well?

Elmar Mair

42:33 - 43:49

Yes, we have been talking to some partners there. It's a little too early to disclose, but obviously it makes so much sense. You even have like in a more smaller space now you have more plants and they all need to be looked at because also there equipment can fail. Best can show up like other issues and all clock the regular muscle and all that. So it's being able to monitor your plans is going to be always key. And that's also the thing which allows you then to really push your plans to the next level. Understanding it based on transpiration rate, for example. How quickly is it growing? When is it growing? When do you want to really stress it to get some results or something like that? It makes a lot of sense and I think with our data pipeline, AI and all that, I think we are in a really good place to actually help facilitate that. We're probably going to adjust the system a little bit to make it work in a stacked in a tiered approach. We have something on our roadmap for next year, which is very exciting. Yeah, definitely. I mean, that's definitely a place for vertical farming. Obviously was a little hit in the last years. I'm like after the excitement. But there is crops where this makes sense and which don't need the sunlight or that much light to actually do well and that makes. Yeah, and your space constraint, you don't be efficient, so there's definitely a space for it.

Harry Duran

43:49 - 43:53

You might have to come up with a different name. Another insect maybe, instead of a spider.

Elmar Mair

43:54 - 43:56

Definitely. Yeah. You already have some ideas actually.

Harry Duran

43:57 - 44:29

Yeah. So I like to leave some time towards the end of these conversations for any messages that you have for the indoor farming industry, for your peers. You know, we have a lot of like CEO's and founders that listen to this podcast. And you know, having been through this experience and working on this company and on this project for the past four years, you know, you've seen a lot of what's happening and a lot of the hype that's happening in this industry, but also a lot of the potential for what's possible. So I'm curious what comes to mind for you as you think about maybe a message for your peers in this space.

Elmar Mair

44:30 - 46:04

I think you're on the right track. I 100% believe it. I think what I've seen, what I've learned, it makes so much sense to grow in controlled environments, to be able to grow all year round, the same quality, the same product and push it to the next limit. When we brought strawberries from outdoors into indoors, we were able will achieve three to five times as much yield. Now imagine if you can really like understand the plant's performance and push it to the next level with, you know, all the different factors which we have, what we can gain out of that, not just like preventing crop blood, but like even pushing it further once we dive into breeding and allow us to like breed strains for certain conditions to push more. I mean, we currently agriculture, you know, relies on mules like growing crop, which is hard to kill. It's really resilient. But if you want to really feed all the people that you're going to be in like 2050 or whatever, you have to double our production. And we can't grow, can't use mules anymore. We need racehorses, but racehorse are fragile and you need to really know what you're doing there. It's challenging, and that's where automated data collection, that's my belief that this is the future. But in general, I think any growers should definitely hang in there. I think it's the future not just like we're going to keep evolving and growing, but I think it's actually really going to accelerate as we're going to get better profitability and show like the value add, etcetera, we're just going to explode. And explosion is going to allow us to bring costs down of everything, even learn more, get even better, and it's going to get more and more and more. So I think we are not, we don't see that acceleration yet really fully, but I think it's going to happen not too far out.

Harry Duran

46:05 - 46:41

Yeah, it feels like there's a bit of a shakeout there for the people who maybe didn't have the right business model or the right approach, and they had to really look at their expenses and some companies didn't make it. But I think the companies that have seen that and have learned the lessons from them, I think, are in a better position than maybe they would have been a couple of years ago had they not seen a lot of those failures firsthand from an industry perspective. So I think it's putting people in a position where they're making better choices about where their company is headed and where to invest in resources and how much cash. Maybe it's important to keep on hand all these things that I'm sure keep you up at night.

Elmar Mair

46:41 - 47:35

Definitely. Yeah. And it's. Yeah, it's a little bit the tail end of the first wave. There was a lot of hype, as you mentioned. There was what we call dumb money coming in, where people just put a lot of money and they're not really understanding the space. Ideas were supported, which should not have been supported. Like, there's other people which knew that space. They're like, oh, my God, really? Again, we did that like ten years ago, didn't work, but now we are at the end of that. There's like, I think the companies survive, and to come out of that, they're going to define that market, which is really real. That's huge. It's not the market is real. It's there. It's just something where now the companies made it through this and really shown that we can come up with something which provides value and is profitable. We can now ride this out and be part of that next scaling curve, which is not a high, but it's actually the real thing. Any technology development is like that and I'm really excited about that next curve we're going to see ahead of us.

Harry Duran

47:35 - 47:57

Well, I want to thank you for making some time to come on the show and share the stories. It's always interesting to see all the different technologies that are happening in the space and what people are doing and how they're tackling all the different problems and challenges that farmers face in order to have the best crop possible and have the most yield. And so it's really exciting to see what neatleaf is doing. It's neatleaf.com. is there any other place you want to send people to connect with you and the team?

Elmar Mair

47:57 - 48:08

No. Please check out the website. Contact us for the website or send us an email to infoetlift.com or me personally to almareatleaf.com. eager to connect and yeah, good to hear you start.

Harry Duran

48:08 - 48:36

Well, I want to thank you for your time. It was really inspiring, especially to see the arc and, you know, to see someone who's got that, you know, rich background and especially a background that's connected really closely to the importance of and the emotional connection that we have with our food. Right. And you have experienced that firsthand. And now to now get to experience and work with technology, which was a passion of yours. And now, in a way, bringing everything together now with neat leaf, I'm sure feels good for you that you're working on something that's really changing the world.

Elmar Mair

48:36 - 48:45

It definitely is incredibly rewarding and a great journey and a great time every day, even in tough times, it feels right.

Harry Duran

48:46 - 48:51

Yeah. Well, I appreciate your time. We'll make sure all those resources are in the show notes as well.

Elmar Mair

48:51 - 48:53

Thank you so much for having us, for having me. Appreciate it.