A New Kind of Robotic Farming Moves the Work Indoors and Dispenses With Tractors

Speaker 1: Let me take you down on the farm, but a very different farm on this one. There are no farm hands, no acres being tilled. There are no tractors. There are no crops sitting under the sun. They do things differently here at a place called iron ox. What's happening here. That's different [00:00:30] from any other place where little green things are being grown. Speaker 2: This is where after the plants have been seeded and have had a few days of germination, they come for what we call propagation. This is the, the really early stage of the plant's life cycle, where they start wanting that water and nutrients, but they're not quite ready to be hydroponic yet. And Speaker 1: I, I see lighting is in here. You're controlling that as well. Yep. And it's all L E D you don't need the sun to hit these plants Speaker 2: At, at this stage. They don't need that much light. They're really packed in tight. Yeah. So you're using very little [00:01:00] energy and you can just sort of pack them in. Speaker 1: Now I see, you know, we're still indoors, but we've got a lot of nice filtered sun coming in. And what do you call these big square trays? Speaker 2: So these are our grow modules and it's one of the, the key interve innovations that we think will really set us apart. And so each module will hold several of these rafts. Uh, this, these are rafts. Yes. Okay. And these have just been harvest. Yeah. Speaker 1: They're floating. These are RAFs. Okay. Speaker 2: Mm-hmm <affirmative> and this is where the plants will spend the majority of their [00:01:30] life. And what's under here is the, uh, the water that the roots are suspended in and all of the nutrients the plants need are in that water. Okay. And the reason we grow in these modules is it means that each group of plants can get water with just the nutrients that it needs, uh, rather than having pipes that go through a whole large facility and just pump out the same water to every single plant, we can give each one of these modules how much water it needs, and exactly the nutrients that they need and when they need them. Speaker 1: Okay. So you've got the addressability down [00:02:00] to this scale, which is pretty small as farming goes, right? Normally it's a huge field, but how do you put water and nutrients in here? I don't see any pipes or hoses or anything. Speaker 2: We have to be able to bring these modules to where the water is and to where sensors are. And that's where our mobile robot called Grover comes in. Now, these modules, when they're filled with water are gonna weigh over a [00:02:30] thousand pounds. Oh, that heavy. So it will drive under and it will align itself. And then this will lift up and it can lift the whole module and then drive it where it needs to go. We have the same types of sensors. You'll see, on a lot of other autonomous vehicles or mobile robots. Yeah. There's there's lidars lasers. We also have, uh, this forward camera here. And then we also have this upward facing camera, which isn't for navigation. This is for asset tracking. When we drive under a module, we can look at, you know, some tags and make [00:03:00] sure we know oh, so Speaker 1: Which points on the bottom of those things. Yep. <laugh> interesting. Okay. Where would this deliver a grow module full of plants and what happens when it gets delivered there? Speaker 2: So there's many different work cells that we can bring it to. Uh, one is, is this unit here, which we call a photo booth. Speaker 1: So what happens in here? Speaker 2: So this is where we can get 3d scans of a module of plants. Grover will bring one of the modules into this center right here. Yeah. And then we've got all of these cameras that can take pictures of the plants. Speaker 1: [00:03:30] I see four pair here. Speaker 2: Yep. And from all those different angles, we can combine them and we can get a really high resolution scan of all the plants. And this lets us understand quite a bit about how those plants are growing. It lets us make sure that they're growing kind of on track. It lets us predict how much we're going to have when we harvest it. Uh, it lets us see if an intervention is needed. Maybe it needs a little bit more fertilizer. Maybe it needs a little bit more water. And by taking it in this way, we can get all that information into our systems and sort [00:04:00] of close that loop. And since Grover is picking it up right there, it can take it where it needs to go for the next step. They're much higher resolution, uh, than, uh, you or I could see and not to mention, we can train our models on, you know, specifically on the things that really matter for each different crop type. Speaker 1: One of the key things you're either solving for and or challenging is the human eye. A lot of this today is done with very experienced farmers and growers and field staff that are [00:04:30] running around and saying, I've got lots of experience I can touch and smell and see Speaker 2: That's how we train the system is with those knowledgeable experts, teaching the, the computer, what to look for. But once we've got this system trained, now it's robots bringing it to one location. Speaker 1: All right. So you've shown me a lot of interesting tech robotics and sensing working together. What's the big point. Why does this need to be done these sorts Speaker 2: Of facilities? When we can do them at scale, they let us be a lot more efficient in all the [00:05:00] different resources that we're using. And they let us put these facilities in all sorts of places and grow during periods. We otherwise would not be able to grow in. Speaker 1: See, I like that idea of sort of not local, but regional agriculture. Speaker 2: It doesn't have to be hyper local. It doesn't have to be, you know, one of these facilities in the center of a big city, it can be regional. It can be, you know, hours and hours away. That's still much, much closer than being shipped. Thousands of miles. Yeah. Yeah. You know, being shipped, uh, you know, by a plane or by a boat, as you know, agriculture [00:05:30] has grown up in the world, it's very resource intensive. It uses a massive amount of the fresh water. Tons of fertilizer fertilizer requires a lot of energy to produce and fertilizer. Emitts a lot of greenhouse gases as well. So if we want to reduce greenhouse gases overall to where they need to be, we have to look at the agriculture sector Speaker 1: When you're done. And you harvest all this basil, there's still a lot of water in there with nutrients in it. Yep. Do you dump it or what do Speaker 2: You do do no, we can reclaim that. We can take that water, [00:06:00] treat it a little bit. The nutrients that were not used by the plant that was growing there, we can get them right back into our system. We can do our leafy greens. We can do fruiting crops. We can do vines. And for each one of these, they can make use of the systems we've already built. And we can build new systems that can cover these entire class of crops. We don't want to design an entire greenhouse for just one crop that can then only grow just that one crop. Speaker 1: Each [00:06:30] plant is cared for by our expert growers. And our growing process is empowered by robots. You don't see that on every package of basil. What intrigues me and impresses me beyond just what we've literally seen here in this iron Oak growing center is something I've heard echoed by the plant-based protein companies by Monarch tractor. We visited recently over and over. You hear a theme about really getting efficient about the inputs, getting a closed feedback loop about how you apply the water, the nutrients, the light, and getting very smart [00:07:00] about inspecting the progress of plants as they grow. And at the same time, reducing your overall energy and greenhouse emissions footprint. When I hear echoes like that from several aspects of any given industry, that's when I start to think we're seeing the future.