Farming’s New Frontier: AI Steers Path to Green Future

This is Dana Perkins and you're listening to Switched on the BNF podcast. Pesticides are considered an indispensable part of modern agriculture for many farmers. They're used to prevent weeds, insects, and fungus from attacking crops and decreasing yields. But the use of these chemicals comes at a cost because pesticides are a driver of environmental damage and soil pollution, and they have been linked to declines in bird and insect

Thank you for having me well.

Alex is very kindly saying good morning to me, But it's actually evening where you are, so I'm located in London, you're located in Australia, and we're here to discuss pesticides. It would be great if you could give us a little bit of context as we head into this regarding why we think there might be the opportunity to have

So we've got all these new policies emerging, like the Global Biodiversity Framework, these big supernational targets to reduce the

So it would be great if you could give us a little bit of color on the current way we are using pesticides in most places in the world so that we can understand what it is that actually needs to be fixed.

So agriculture has a trilemma. We need to feed ten billion people by twenty fifty on potentially less land as it's protected by nature, and to halt the biodiversity crisis which we're currently experiencing. As I said, we need to

So currently we spray.

Chemical pesticides at very high volumes, very indiscriminately, and that can either be via aerial with a crop duster, which is a small plane with a payload that just drops the chemical onto the field, or we do it via.

A big boom spread which is.

An effect attractor with a tank attached to it with two large wings which then deliver that pesticide along the rows of crops. Now, what this means is that the chemical can actually then either blow away in spray drift as the air picks up, or more chemical is released then is actually necessary to kill the pest. And that pest can either be a weed, it can be a fungus, it can be an insect or something even more obscure like a nematode which exists in the soil and can

So why don't you give me a little bit of context around just how bad the problem is in terms of the overuse or inefficient use of pesticides, be those insecticides or herbicides.

Absolutely so when we spray these chemicals in these big broadcast methods, ninety eight percent of insecticides and ninety five percent of herbicides don't even reach their target pests.

So what that means is.

It that chemical can then leave the cropping zone and the cropping system or the farm and then cause knock on consequences to nature and ecology, whether that be disrupting breeding cycles of aquatic organisms, killing native plants, affecting the behavior of large animals, that neurotoxicity as we call it, and even up to the human level. There was traces of glyphosate found in human urine at very, very shocking levels recently in the United States.

I think there is a general understanding that for overall health of organisms, we want to reduce pesticide use. I mean, these are micals designed to kill things, including insects, which we do need for biodiversity. So the urgency is so well laid out by the fact that ninety eight percent I mean, I'm just going to repeat that, because that is such a huge amount of overall pesticides not reaching their desired location. But Aha, we are here to discuss

Absolutely, so there's a leading technology which is emerging that we call as a blanket variable right technology, and it kind of does what it says on the tin. It varies the right in which we apply these chemical pesticides. Now, the pinnacle of that the best performing technology is green on green optical spot spray and that essentially is cameras that amounted to a conventional broadcast boom spray that then uses artificial intelligence algorithms to detect the weeds in the

So, if I'm understanding this correctly, this is basically a very advanced facial recognition technology but for weeds, so it can tell the difference between the plant that you want to have there and the plant that you don't want to have there.

In essence, yes, okay.

So this is interesting. I wonder if there'll be any mistaken identity moments there. But the point is that this seems like a technology solution to detecting the pests that are within the desired crops. But why is this considered an artificial intelligence application? What about it makes it AI?

So what they do is they feed a model, a computer model, annotated images of weeds thousands of thousands of times, and the model begins to learn the shape of the weed, the color of the leaf compared to a crop, and it can eventually manage to do that at speed and make the decision by itself which is the weed and

So also to simplify this, it enables precision application, so it's able to really make the location of the pesticide that's being used quite specific to address the specific weed. Now, how much of an improvement would this lead to, because it's all going into the soil in this general area. Is this actually going to dramatically reduce the amount of pesticides that are used in a commercial farm?

Absolutely, the best available technologies are promising an up to ninety seven point five percent production in the amount of chemical that is used.

Wow, So that's a huge improvement. And then that begs the question does this then have a net economic benefit for the farmer? What are the financial savings for the farm in terms of reducing pesticides and pesticides essentially a really expensive part of farming.

So yeah, with our modeling that we've done here at Bloomberg NIA, we've found that this employing this green on green technology can be up to sixty percent cheaper.

Than current broadcast methods.

So that takes into account the fuel, all the changes in cost of labor, the chemical itself, the herbicide, and the cost of the actually implementing and fitting one of these new green on green machines.

So presumably the machines themselves have a pretty substantial capital outlaid the beginning, and then the savings that you make on reducing your pesticide use is going to help you pay for that over time. But let's get into that to begin with. How expensive is it? How much cost is it going to be for the farmer to start using this technology to begin with.

So there's two models that you can go for here. You can either buy a new sprayer which is equipped with this technology, or you can be buying a retrofit part, which is where a secondary manufacturer will fit this to your current existing sprayre Now, most farmers already own a spray, they've already outlaid that capsule expenditure. Most of these models sit around one hundred and twenty thousand dollars.

And are these sprayers stationary? Are you capable of moving them around? And the real question I'm getting at is whether or not you need to have a lot out of them on a farm or you'd be able to invest in a smaller number and actually move them around.

Typically a farmer will have one sprayer and that'll be the one spread that they use. Now, obviously, in the really really large operations, when we get up to the thousands upon thousands of hectors. You may need to have a couple of machines, but for the most part, one machine will definitely take care of all of the land that you have.

So where in the world are these being adopted right now? Are there more technology friendly farming communities that we should know about somewhere between where you live and I live on this beautiful blue planet?

There certainly are there, certainly are.

Actually Australia, where we're sitting right now, has had the greatest adoption of variable rate technology or precision application. In the earlier stages, there was a technology called green on Brown, not quite as advanced and can only be used under

So you mentioned the conditions being right in Australia for early adoption of some of these technologies. What are the conditions that need to be in place in order for this to work.

Older technologies were only able to spray when they sensed a plant and not use AI to detect the difference between a plant, a crop and a weed. So that means they could only be used when the field was being fallowed and that is when the field is empty of any crops. So we were only spraying during that period between harvest and sewing, again to avoid the cost of accidentally spraying your crops. Now with the newer technology, we can use it year round because it's able to

When we're thinking about the sort of crops that this would be most used for, whether they're the economically beneficial ones because certainly some of them make more money than others, or the way that they're harvested does it make a difference. I mean, I'm from Napa Valley, which is well known as a wine growing region, and there's multiple different ways to go about growing grapes, and some of them are dry farmed, but most of them are in rows with

So the most gains that we're going to see here are typically in row crops. Now, these are big, broad acre operations that are growing grains, oil seeds, things like corn, wheat, canola. Now that's not to say that it can't be used and it isn't used in tree crops. The amount of gains that we're going to see are going to be best from these larger row crops, big operations. We think here in Australia there's some farmers who can go for

So you've painted a picture in my mind of fairly large farms monocropping, and you know, the things that are actually reaching most people on their dinner plate. And there have been concerns not just regarding pesticides, but actually the vulnerabilities of monocropping and also what it does to the soil to constantly be rotating the same crops on the

Certainly.

So, a lot of these chemicals reduce the health of the soils, not only in the microorganisms, but also those larger organisms that are existing, the beneficial insects and invertebrates. So by reducing the amount of damage that we're causing to the soil by overspraying pesticides, we actually stop the vicious cycle, which is damage soil less beneficial insects and therefore more chemical needing to be sprayed.

The pesticide business is a big business. By reducing the amount that are actually being used, there could potentially be a lot of money at stake. What is the current market size for this and how much are they doing? I guess annually in terms of how many pesticides are being sold around the world.

So the global pesticide industry or crop protection industries is also known, is around forty three billion dollars a year. Now that's completely under threat at the moment, both due to legislative pressures but also due to the farmer being very squeezed. The farmer doesn't receive very much of the share of the dollar, so wherever they can reasonably cut costs, they will. Now this technology is allowing them to reduce both their cost base but also to improve the productivity of their farm.

How are the pesticide companies viewing this new technology that's emerging and is it something that they're interested in and potentially developing a house or even acquiring, or is it something that they're threatened by and trying to figure out how long they can avoid it becoming mainstream.

The crop protection industry has been able to move very agile, and they've learned from previous legislative knocks such as in the EU the amount of pesticides that are being restricted for use on a yearly basis. So they've really grabbed the ball by the horns and they've jumped in at the deep end. For instance, there's a startup called the Eco Robotics ARA, which is a precision delivery unit which promises to reduce crop inputs from herbicide insect side and

So when you're talking about this AI technology being used to spot those pesky pests, you specifically referenced it being put onto a boom and kind of coming across the entire field. But I'm thinking now of drones that are not actually connected to the ground flying over. That is something that has been i think excitedly talked about. And the question for you, is someone who's looked at this in more detail, is whether or not you think that that's got legs No pun intended.

So there absolutely is an application for drones in agriculture. However, the spraying of large sways of cropping land is probably not it. The payloads are just too small at the moment and we haven't really seen advances in swarming technology, and that's when we deploy multiple drones. So I think the largest drone which is currently available is the Guardian Agriculture sc one, which has around seventy to eighty liters of payload capacity. It has also been invested by big

Not able to get to as easily.

Alternatively, and areas of sprayer can't enter, like greenhouse agriculture or glasshouse agriculture. So there certainly is a case for spray drones. The technology is just not there yet at the moment to spray big swathes of agricultural arable land. Another application of the use of drones will be in

Water for your cropping.

So this will enable us to work smarter using imagery mapping software that then can be loaded up to farm management software to deploy chemicals smarter. So the case for drones is either as surveillance and mapping imagery technology or as in small payloads in areas that we can't get a boom spread to.

So the drone technology is actually a compliment to the boomsprayer as opposed to an alternative to it, at least in the picture that you've painted of the future, where we're looking at the field and being strategic and then the boom sprayer has that ability to actually do the end application in that quite precision way. This seems like a great solution. You've got me very excited about what

Absolutely so, we already have companies offering drone surveillance to be able to detect other types of pests, such as Xavio for instance. However, what we haven't yet seen is a machine that is actually capable of seeing and detecting the other types of pests in real time and then therefore delivering the precise dose of required pesticide, whether it

So when I'm imagining the future, I should not be imagining some sort of robot, whether on a boom or as a drone, flying around chasing insects in some sort of a field. But who is coming up with these technology solutions. Are they Silicon valley tech companies, Are they smaller startups or are they the pesticide companies themselves thinking about these new and innovative ways to essentially make farming better.

So it's actually a mixture.

What we've seen is we've seen two roots to market, and that is from either in house R and D and bolt on acquisitions, or it's from parts ships with larger companies and smaller startups. So John Deere has gone the complete in house route. It's kept its doors shut, It's acquired a couple of interesting companies along the way,

Now let's just pivot again to adaption. Going back to what you were saying earlier about some of these units being over one hundred thousand US dollars to install. This is a big capital outlay, and we know that farming is a tough business to be in and the margins are not amazingly great, so this may be a fairly big hurdle for some farms to actually cross. What are governments doing and maybe we can use the case of Australia where we've seen good farmer uptake. What it's driven that.

That's a great question.

And actually the geography with the highest adoption in Australia has seen most of its adoption of infact all of its adoption driven by savings to the farmer. Now, with nineteen percent of the variable costs for a grower in each year being comprised of crop protection reducing that can really cause a benefit to the bottom line for the

So really the farmers are very much motivated by how much money they're going to save and the payback period in terms of money saved on pesticides. It's going to make it worthwhile pretty quickly to buy these machines. They may even be able to get loans for them.

Absolutely, So now that does bring a little bit of a question into it, and the limitations are that in countries where or with a grow who is not large enough to be able to outlay the capital expenditure to actually obtain the chemical, then what does that mean for them? But we are hearing certain calls, whether it be tax breaks or tax incentives or subsidies for purchase of precision agriculture equipment.

So we're thinking about the future of farming here, and this AI application is certainly one important way for us to think about tackling a problem pesticide use. So tell me a bit about the future. How soon and what does it look like in these modern farms that are going to be much more efficient.

So what we'll see in the future is we'll see fungicide, herbicide and insecticide all being able to be delivered from the same machine. But not only that, the machine will be able to make decisions in real time to detect all types of pests, but also be able to learn year on year what the trends are, what does the local area look like, what does the agronomy request or require to be applied, and so essentially what we're doing is we're building a large data model of the farm.

On farm cameras.

So as they learn, the machines get better and they get better and more precise.

So in this increasingly technology reliant agriculture future, what are the potential downsides? You know, with the exception of the obvious ones, which are that there is required energy use for operating any sort of machine, as well as energy use for the servers that will very much be required in order to facilitate this AI technology the energy transition. This makes it intrinsically linked to the changes that may

We haven't spoken at all about autonomy on these machines, and actually what we're moving to as a smart vehicle that can actually work and operate without the requirement of a human. But what that means is then the farmer can spend more time monitoring their crops and actually paying attention to the local agronomy. Even if these machines moved to being fully autonomous, being able to operate by themselves.

So, Alex, my final question really comes down to the fact that when I think of technology now, I think about all of these subscription models that we all subscribe to. Will this apply to this technology and will there essentially be software updates to the AI technology where it becomes more sophisticated and can do a better job of figuring out what the precision sprayer needs to do. And as a result, is this both a hardware and software conversation?

It absolutely is.

So there's essentially three models that the machinery manufacturers are employing. Firstly, is a one off fee to purchase both the software and the hardware. Then we have a second where you would purchase the hardware and then choose to turn on the software for a year, and that would enable you to spray the amount of hectares per year paying for

So this is really interesting because not only is this technology going to help save the farmer money on its pesticide use, but there actually is a long term revenue stream for the companies that are actually selling the technology. So I am very interested to see how these business models evolve and invariably where it ends up in the world,

Thanks for having me, It's been great.

Bloomberg NF is a service provided by Bloomberg Finance LP and its affiliates. This recording does not constitute, nor should it be construed as investment advice, investment recommendations, or a recommendation as to an investment or other strategy. Bloomberg n EF should not be considered as information sufficient upon which