Resources

• Dryad Networks
• Silvanet: Dryad’s Wildfire Detection Solution
• Impact of Wildfires on CO2 Emissions

Host and Guest

Climate Rising Host: Professor Mike Toffel, Faculty Chair, Business & Environment Initiative (LinkedIn)

Guest: Carsten Brinkschulte, CEO of Dryad (LinkedIn)

Transcript

Editor's Note: The following was prepared by a machine algorithm, and may not perfectly reflect the audio file of the interview.

Mike Toffel:

Carsten, thank you so much for joining us here on Climate Rising.

Carsten Brinkschulte:

Thanks, Mike, for having me.

Mike Toffel:

So why don't we begin with just a brief introduction to how you ended up at Dryad and a bit about your entrepreneurship background.

Carsten Brinkschulte:

Yeah, I'm I guess what people call a serial entrepreneur. So I've had three startups so far that I started or scaled in all in the telecom space. So I'm a telecom guy with a technology background, software development for 15 years, and then started companies in mobile email synchronization, device management and 4G telecoms technology.

I had a company in London on the stock exchange for seven years. And then I sold the company to BlackBerry, sold one to Twilio, and then moved back to Berlin. And when you sell a company, you kind of fall into a hole. You don't know what to do. You lost your baby and purpose in life, and you search for a new purpose in life. And that was, for me, a key event when I saw lots of fire on the news, the Amazon literally burning on Black Friday in Australia and California was on fire. And I kind of felt like, okay, this is something that really needs to be addressed. Basically, that's how Dryad, the idea to Dryad started. And there was on top of that, the Fighters for Future movement over here in Europe, were children, teenagers going on the street protesting, including my daughter, actually. And I guess that's what kind of in the end, flip the switch to start dry and say, we need to do something serious rather than just help the next smartphone generation or build the next 5 or 6 G networks.

Mike Toffel:

Yeah, so you're aware of the risk caused by climate exacerbated by climate change. And you had even household salience with your daughter out there protesting. And then how did that come into the foundation of recognizing that there are some technologies out there that you could deploy to do wildfire detection, which is the area that you're in?

Carsten Brinkschulte:

Well, basically, once you have a hammer, everything starts to look like a nail, I guess. And I'm a telecoms guy, so I can't help it to do telecom stuff. And when I thought about the problem of wildfire, which is staggering, actually, most people don't know that wildfires are causing up to 20 % of global CO2 emissions. That's about 6 to 8 billion metric tons. And somebody needs to do something about it, and we hope we can help at least. And the idea was, of course, yeah, how can you do something about it? And I'm a telecoms guy, so I can't help but using telecoms stuff to try to address this problem. So we had the idea basically with a friend of mine, a bottle of whiskey actually, to do telecommunications in the forest and put sensors in there that can actually detect the fire. Because we have, you know, we've got fire sensors in every room these days, almost by law in most countries. Why don't we have these sensors in the forest? They work quite well in buildings. was the idea. let's take a, the first idea was to let's take a fire sensor from a DIY store and connect it to the network and we're done. That was the initial idea. And it turned out a bit more complicated than that. And now we're at Dryad Networks four years later.

Mike Toffel:

Yeah, so the idea of these fire detectors in households is twofold. One is to ensure rapid evacuation to protect lives and then also to prompt a call to the fire department to be able to suppress the fire more quickly than you otherwise would. So even if you're not home, your neighbor will hear it and call the fire department. Is that the same idea that you're conceptualizing in the wildfires in the woods? Or is it something just one of those two? Or where are you on that?

Carsten Brinkschulte:

Well, the basic principle is similar. Technology is different. But the fundamental principle is detecting the fire as early as you possibly can. Tell the firefighters about it so that they can come and do something about it while it's still possible. That's the same principle or one of the principles that why we have fire detectors in our homes so that, yes, people can evacuate. That's one aspect. But then also the firefighters know about it. It can come before the whole building is on fire.

And in our case, we want to stop wildfires as early as possible because, you know, when there is a fire starting on the ground, 85 % of wildfires are human induced. It’s arson, reckless behavior, technical faults, accidents. We are causing most of the fire starts. Most of them start on the ground and they start very small. Except if they're not if they're not recognized, if they're not detected early enough.

They soon become one of those raging wildfires that we see on the news that actually are almost unstoppable. But they are very stoppable in the first hour or two when they're still very small. So time is really of the essence, right? And the idea here is what we're doing with dry networks is to build a sensor that can detect the fire while it's still in its infancy, when it's still very small, when actually maybe there isn't even a flame. So we want to smell a fire just like the fire detector in your home. doesn't need to see a flame, it just needs smoke. And so do our sensors. When there is smoke, the sensor is detected and then they call the fire brigade's effect.

Mike Toffel:

And so how did you develop this technology? Did you start with the technology that we use in our homes and then test it out in the wild and then decide what tweaks were needed? Or did you start in a different way? How did you start?

Carsten Brinkschulte:

Well, we started a bit differently because we knew that you can't just take a DIY sensor from the store and strap a battery on it and connect it to the network. That was just the first idea. When you think a little bit longer about it, you need to put a bit more sophisticated technology on it. So what we started actually in 2020, we assembled a team of six co-founders doing hardware, the other one embedded software, the other one cloud software, a finance guy and me, and we created prototypes. So the prototypes were based on off the shelf DIY tinkering electronics. So you can buy some electronic boards, you can buy sensors, and you can put them together into a prototype. And you can write software on it. And so we had basically what we have now, but as a very early prototype, a gas sensor that can actually smell, gas, hydrogen, carbon monoxide, and VOCs, and a software that would interpret the signals. And if it reaches certain thresholds and sees certain patterns, it sends a message to the internet and effectively create an email with a map on it so that we create it as a prototype, basically to convince ourselves and also investors that this is a good idea.

Mike Toffel:

And so how do you train this technology? So this technology is basically out there like sniffing the air and you want it to be able to distinguish smells that are hints of fire and ignore false alarms and all that. So there's this type one, type two error issue. You don't want it ringing when there's not a fire and you don't want it to not ring when there is a fire.

Carsten Brinkschulte:

Yeah.

Mike Toffel:

How do you do that testing and development? Are you doing that in a lab? in forests?

Carsten Brinkschulte:

Well, it took us literally, I would say, three years to actually get that working and a lot of venture capital and engineering power. we basically, nowadays, we use artificial intelligence in the sensor. So when the gas sensor sends its signal, we have an AI looking at the data. And we train the AI in the laboratory where we're burning different species at different temperatures to simulate different climate environments that the burning can happen in. And then we also trained the second part of the AI for the normal smell that the sensors should expect so that they can distinguish between, that's a fire. no, that's just some rotten egg or something that smells a bit funny in the forest. So we now have about 20,000 sensors distributed all over the world. And they're all smelling all the time. And we're taking that data of the various forests in the world where they are, and taking the smell of that, the forests of the world, if you want, and putting that into the AI. So now the AI can distinguish between, we've trained how oak smells like if it's burning, or redwoods. And we've trained it how it smells like in the redwood forest when there is not a fire. And we've trained it how it smells like in an oak forest when there is no fire. So now it can quite reliably distinguish.

Mike Toffel:

Got it. And so for the listeners who know about AI and machine learning technologies, what's the jargon-light version of AI that you're using for this?

Carsten Brinkschulte:

We're using machine learning. So AI is the umbrella under which all of these technologies fall. Effectively, technically, it's machine learning. You basically stuff a lot of data into an engine, and it tries to find patterns by which then signals later can be interpreted. And that's basically it, right? So we throw tons of data from the forest all over the world and thousands of measurements in our laboratory into the machine learning. And out comes a library that can distinguish between fire and non-fire events quite reliably with very low false positive rates, which is the key thing, but at the same time with high sensitivity. So you want it really sensitive, but with a very low false positive rate. That’s what the machine learning engine does that we're using in the sensor. In our case, technically, a bit more challenging. We don't do that on high class, high powered computers in the clouds. We're executing the machine learning code in the sensor. if you see this here, it's a small handheld device, size of your palm, roughly. Very low powered device, costs $100, so it doesn't have a lot of computer power. It runs on solar power as well. It doesn't have a lot of energy, but it still does machine learning. We needed to do this because we have thousands of those sensors. We can't have them talk to the cloud all the time.

Mike Toffel:

Yeah, so that was a question I had, like whether the machine learning was operating within the sensor or in the cloud, and you made the decision to do it in the sensor itself.

Carsten Brinkschulte:

It is. It is which is the only decision that you can make if you want to scale this to hundreds of millions, hundreds of thousands of millions of sensors, because otherwise you would overload the networks with the data transmission to the cloud. And that's key. So in order to scale, to really go very large scale, we needed to swallow some hard pills. And one is to put machine learning into the sensor. Because the point is, and that's why we call it Dryad networks, because we also create the networks for the sensors. Sensors are great. And there's a lot of work that went into it with the AI, ML, and everything else. But the key thing is, how do you get the data out of the forest? The forests are typically not connected with 4G because the mobile operators don't bother to put $100,000 base stations in the middle of the forest to give some connectivity to the bears. So they don't. And that means that there are no telecommunications in the forest.

So what we do at Dryad Networks. We're building low cost, large scale, solar powered mesh networks that we hang in the forest for our sensors to communicate. And these are low powered networks. So you can't transmit videos or high stream data, or high pipe, a large amount of data. That's why we had to push the machine learning execution to the edge, to the sensors. Because we need to do it over very narrow band networks that we're building for the forest.

Mike Toffel:

Got it. So you're developing the sensors with the AI machine learning piece embedded in it. Then you're also developing a low powered network. And one of the things about machine learning, as I understand it, and it's most of its deployment, is that as you use it more, it can get better, right? You set it up to self-train and get even better at detecting, at reducing type two and type one errors.

Carsten Brinkschulte:

Yeah.

Mike Toffel:

Now in your case, when the technology is in the sensor, does that limit your ability to do updates or did you build into it an update feature?

Carsten Brinkschulte:

Absolutely critical for our solution, because we do the machine learning in the sensor, is the ability to update the firmware in the device. So we did that from the very beginning. And that makes it even more difficult over the low-powered networks, the LoRaWAN networks that we're using. And on top of that, we're building mesh networks, which are based on LoRaWAN. So it's really difficult to do firmware updates, but we do that, absolutely. And it's absolutely critical that we can update the and with that in particular the machine learning libraries in the sensors that are distributed over tens of thousands of sensors. Because yes, we do indeed update the machine learning model, and it gets better and better and better over time because we get more and more data from all of the forests of this world. With every deployment we learn, or with every second deployment or so, we learn a new forest type.

We installed in South Africa. Guess what? The forest smells different than the woods in California. right now, I'm talking to you out of Bangkok in Thailand. And guess what? They've got different types of forests that smell differently over here. So we're learning how they smell and improve the machine learning library with the data from here. And then we will update the sensors over the air after they're being installed.

Mike Toffel:

Got it. And these sensors you mentioned are solar powered. So you have a battery installed that the solar charges and allows it to operate.

Carsten Brinkschulte:

No battery. Categorically, no. We absolutely hate batteries. Why do we hate batteries? Because A, you know what? They are a fire hazard. Have you ever seen somebody putting a nail through a lithium ion battery? You don't want that to happen in the forest when you're trying to prevent fires, right? So none of our sensors nor our gateways have anything with that smells of lithium ion in them. No batteries. We use supercapacitors for energy storage instead of lithium ion. And I think that's really important. I don't think that the fire departments of this world would allow us to install fire starters in the forest when we're trying to prevent them.

Mike Toffel:

Got it. So these super capacitors allow for energy storage as an alternative to lithium-ion batteries. And these allow you to have these sensors work overnight and for a couple of cloudy days.

Carsten Brinkschulte:

Correct. Exactly. So we've got the solar cell charging the supercapacitors during the day so that they can survive the night while doing their job. And yes, we do get enough energy even under the tree canopy in the shade. Makes it really challenging for the electronics and the software to be extremely energy efficient because the solar panel we have, if you look at this device that I'm holding up here, size of your palm has a very small solar panel, 0.45 watts for anybody who's interested in it. So that's really nothing. And if you put it into the shade, it gets 10 % of that. But still, our electronics need to work, and the charging needs to work.

Mike Toffel:

Yeah, so your device has roughly looked like a two inch by two inch square of solar panel, something like that.

Carsten Brinkschulte:

Yeah, I'm sorry, I'm not familiar with the inches, but I think it's about right. Yeah, it's six centimeters.

Mike Toffel:

Haha okay, six centimeters squared. And so how far apart do these sensors can they be? Or do they have to be, can they be several kilometers apart or do they need to be several per square kilometer? What's their range?

Carsten Brinkschulte:

The question is how much money do you want to throw at the problem? The more the better. The more sensors you put in an area, the faster the detection. But realistically, think the sensor can detect from... Sorry, not knowing in yards, 100 meters, 150 meters distance. We can protect about three acres that I know in imperial units or about one hectare with the sensor, the size of a football field, roughly. You can, of course, pace them out more widely. That's absolutely possible.

But then the fires have to be a bit larger before they can be detected. And the whole point about our system is the ultra-early detection. We want to detect fires the size of a campfire or even smaller. And that means we need a lot of those sensors, which also means the price of the sensors is critical. So it needs to be low cost. our sensors are really low cost so that you know the customers can afford to flood the forest with them to put a lot of them in the forest because only then it does make sense because imagine if you had a sensor, even if you build a sensor, maybe that costs like a thousand dollars or two thousand dollars, right, that can detect because it's so full of super electronics, I can detect a fire from say a mile distance. That's really great. Theoretically. Except if you then think about where does the fire is, where is the fire actually? Because a gas sensor like we do, it cannot geolocate the fire. It just knows, hey, there is a fire. It's like a fire. But where? We don't know.

And if you have one for every football field, for the firefighters, it's quite easy, right? You send them to the geolocation of the sensor that detected the fire. And they have to search the area of the size of a football field in the forest, that's quite easy, I think. But if you imagine 10 times or 10 football fields in the forest or even more, then it becomes really challenging. So it's really key that we put a lot of those sensors into the forest, very close to each other, because only then can we geolocate the fire, which is the key thing, right? We won't tell the firefighters, here, here's the fire, go there, extinguish it. It's still small. don't, you just need a helicopter and that's it.

Mike Toffel:

That makes a lot of sense. So who are your primary customers? These are private sector, public sector, some of each?

Carsten Brinkschulte:

Well, we have three custom groups. One is the private forestry timber companies that want to protect their assets, right? If they grow trees for 30 years and then within a week or one or two days, it's all gone. So protecting the assets is key here. Then we have municipalities. They basically want to protect infrastructure and people living in proximity of the forest. So it's public safety if you want that the use case. And then we have utilities, which are power line companies and railroad. Where you have two use case, two motivations for that. One, they want to protect their own assets, the infrastructure from fires coming in, destroying the power lines or the railroads. But then also they have the problem that the infrastructure may cause fires, which has huge liability issues. And so we have a dual role here. These are the three customer groups we have.

Mike Toffel:

And are you teaming up with or coordinating in some way with insurance companies? Because it seems to me like building fire suppressant systems where when a builder or an occupant of a building procures that, it reduces the risk of disaster. And so they can achieve lower liability or lower property casualty insurance, which can sometimes pay for the technology itself in just even one year sometimes. Are you finding that in your markets as well, that the reduced liability or the reduced insurance costs can help subsidize the deployment of your technology?

Carsten Brinkschulte:

Yep. In principle, our markets are global. We sell to the US, to California, to Canada, to Europe, and now to Asia as well. Yes, in principle, insurance is a great market because, for example, in California, you've got whole areas where you can't even get insurance anymore for your house because the white fire risk is so high. So obviously, there is a link to insurance here. We could help potentially to reestablish insurance. And that could solve the problem that many California communities are facing. The problem, I guess, that we're seeing with insurance is that that industry is extremely conservative and very numbers driven. So that they obviously want to see proven track record over several years. How many fires have you detected, prevented, and tell me about those numbers over the last five years. And I guess that's why we're not yet there because we just started to sell, starting to sell our solutions since last year.

Mike Toffel:

Got it. Are there stories within the year or two that you've been in the market that you can point to where your technologies provided earlier alerts and reduced the scale of the fire that otherwise would have occurred?

Carsten Brinkschulte:

Well, mean, we've detected hundreds, if not thousands of fires already in particular, because every customer that buys our solution, the first thing they do, they start a little fire to see if it actually works. So we now have 50 installations. So we've detected at least 50, probably 100 fires or so. But they are, of course, intended, now, small, controlled burns. But we did detect, in fact, an unwanted fire in the Lebanon, where I guess they now have even more fires for other reasons.

The fire we detected was in an area that had spilt forest, and they don't have a lot of forest in the Lebanon. And we deployed there, protected an area near a village. Then there was during the high fire season was a farmer that was burning grapevine, which he shouldn't be doing during the fire season. And we detected it. They called the authorities, and I guess he his wrist slapped or something. And then we… I guess we might have prevented an actual fire there.

Mike Toffel:

Now, the alarms come to Dryad and then you connect with the fire department or the authorities, or do the authorities get the alarm directly? Or does that vary by customer?

Carsten Brinkschulte:

Well, it varies by customer. Obviously, technically, they go to us and we relay these alarms with the geolocation of the sensor that detected the fire. But the customer can decide where the alarms get sent to. You can either, in the simplest form, you can define an email address that we send a mail to. But we also have an API that we can integrate with command and control systems where we send a geolocation to a commander control system that's then used to dispatch first responders to the location that we detected. It really varies from customer to customer, to be honest.

Mike Toffel:

Got it, got it. And what are the competing technologies to your sensors in the forest? We've talked before, you talked about satellites or cameras and other sensors. What's the competitive landscape look like and what are the pros and cons of your solution relative to those?

Carsten Brinkschulte:

So I think everybody who thinks about wildfire detection, the first thing you think about is satellites and maybe cameras. And obviously, there are good solutions out there already existing. And there must be like 50 camera companies that produce wildfire detection systems. And there are a few satellite solutions as well out there, either using existing satellite data or even starting dedicated satellites just to detect wildfires. And they all make sense. I would say it's more a co-petition between these technologies, because in the end, every technological approach has its advantages and disadvantages. And in the case of wildfire, as it is often in safety-related technologies, if you deploy multiple solutions, you get a better product in the end, a better service. So it makes a lot of sense to deploy sensors, because they are just the fastest way of detecting a fire, faster than cameras, much faster than satellites.

But you need a lot of them. And if you combine the sensors with a camera solution, for example, it has the advantage of overlooking a very large area, which is difficult for sensors to do. Then it makes a lot of sense. have the sensors in the high risk, high value areas, and you've got the camera to cover the rest where the sensors are not positioned. And then, of course, cameras cannot really see what's happening under the tree canopy, so they detect fires later. Yes, the smoke plume has to rise above the canopy before they can actually see and identify it. But still they're good, right? They detect it probably earlier than humans would. And combined with sensors, you get the human-induced fires that happen along roads, along highways, along hiking paths, around camping sites. You catch those with the sensors. And if there's a lightning strike, you catch those maybe a little bit later, but you can catch that with the camera.

And then on top of that, you should throw in probably the satellites because satellites, while they're not good at detecting fires early on, take hours before they can actually detect anything. But they're very good at tracking where fires go, and they have spread out of control. So if you can't stop a fire early enough, then you end up having one of those mega fires that we're seeing. You still want to know where is this fire going? Do you have to evacuate people? And how do you respond to it? And for that, are invaluable and much better than cameras or sensors would ever be. So ideally, you would want all three.

Mike Toffel:

Yeah, interesting. So they all provide somewhat complimentary services in a way.

Carsten Brinkschulte:

Yep. Yeah, that's why I would call it a co-petition rather than competition.

Mike Toffel:

Now, the demand for wildfire detection, I imagine based on the maps that I've seen in the prediction of dryness and droughts all over the world, is only going to grow in the coming decades, no matter what we decide at COP and so on. It's just a question of how quickly the demand will grow for these services. Where are you seeing the most enthusiasm for adopting your products? What sectors, what geographies? Because, you know, there's dry areas all over the world that have railroad lines that go through them or electric utilities, but not everyone's able to essentially afford insurance or perhaps face the same liability risks and so on. So, where you seeing the uptake today and then where do you see that changing in the future?

Carsten Brinkschulte:

Right, there's an impact for profit. So yes, our main mission is to have an environmental impact to prevent wildfires and with that CO2 emissions and the destruction of habitat for basically the fauna and the flora. But the point is, of course, because we also want to be a profitable company and not just a charity, that doesn't scale really. We need to look at where is the highest willingness to pay.

And at the same time, where is the biggest problem? So we're obviously focusing on those areas where we're seeing the most wildfires and where we see the highest willingness to pay at the same time. that because we're an impact for profit, that's our focus. So it's the United States, California, Oregon, other regions which are far prone. It's Canada. We're seeing a lot of uptake in Canada, a lot of demand. No wonder if you look at that they lost.

Think it was 14 million hectares of forest last year, but more than the entire country of Germany has in terms of forest in just one summer. So there's a lot of take up in California and demand for solutions like ours, but also in Southern Europe, Greece, Spain, Portugal, they're literally on fire. And there's a willingness to spend too. And then we're seeing Latin America, in particular Brazil, Chile, prime targets for our solution with a lot of fire and rather developed economies. In Africa, it's South Africa predominantly that has this intersection between willingness to pay and problems with fire. And here in Asia where I am right now, we see Thailand, Indonesia as two focus areas and of course Australia. We've just opened a subsidiary in Australia to go into that market.

Mike Toffel:

And is the sales approach the same across the three customer segments you mentioned earlier, public lands, private lands like timber companies, and then you group them into what you call utilities, the power lines and the railroads?

Carsten Brinkschulte:

The sales approach that we use at Dryad is to go mostly using resellers because we have such a globally diverse customer base. It's difficult to serve them all from Germany. So what we do is we've licensed our technology to now more than 20 actively reselling partners that represent our Dryad in the various countries and speak the local language and know the legislation, often know the end customers already. And they actually selling our solution to those end customers and we sell to the resellers.

Mike Toffel:

And what technologies, now this bucket of technologies you call Sylvanet, right? That's both the sensors and the network. What else beyond wildfire is on your radar for other types of use case?

Carsten Brinkschulte:

Yeah, I mean, we're calling this thing silvanet, which is Latin for forest and net is of course for the network because we're creating networks in the forest. That's the enabling technology. The sensors that we have, the wildfire sensors, there are applications in the network. So basically what we do is we build a digital network infrastructure in an area where there isn't any. And that's about four billion hectares on this planet that forested area but one third of the land mass is forested still. And we want to keep it that way. But the technology that we've built is this network. Remember, I'm a telecom guy, so we're building telecommunications in the forest. And once you have a telecommunications network in the forest, you can do many other things than just detect fires. So we're working, started to work actively on a fuel moisture sensors to detect and actually to measure the fire risk by measuring the moisture of the fuel underground, soil moisture sensors. But we also want to go into tree growth to measure the growth of a tree that could help carbon sequestration projects that sell carbon credits to measure how much CO2 is sequestered in a tree and measure that rather than estimate it.

But we also want to do gunshot detection and chainsaw detection to prevent illegal logging and poaching. So you can do many other sensor applications that once you have a network. And so we're actively working on some of them already on our roadmap. And then there was one thing which I want to throw in because it's my project that now turned into reality. I've got laughed at that for quite a long time, but the European Union just funded it with 3.8 million euro that is. And that is we want to build an autonomous drone system that we will position permanently in the forest that will actually try to act on the fires that we are detecting. So that's the vision that we're going to start working on now. And we now have the funding to actually really embark on this idea. The vision here is to detect a fire as early as we do already, but then also to not just go there and fly, which we will do see videos, but also try to put it out. So that's the vision where we want to go.

Mike Toffel:

So this will be like scooping water from nearby water sources and then dumping it? Is that what we're talking about or something else?

Carsten Brinkschulte:

Water is not really effective unless you throw a lot of it. If you look at those air tankers that throw tens of thousands of liters, or I don't know what that is in gallons, a lot of stuff on those fires. I don't think that really works with drones, or at least we don't want to do that. So we want to build smaller things, smaller drones, and a lot of them, a swarm of them, and then use different ways of extinguishing the fire than water.

Mike Toffel:

What are those ways?

Carsten Brinkschulte:

Well, there are different ways of extinguishing fires, which sound a bit like sci-fi, but you could use sound to extinguish fires. And we want to try that, see if that works. So it's a very research-heavy project that we're embarking on. But we want to go different ways than the beaten path. And that's to not use water.

Mike Toffel:

Well, we'll look forward to following that along. That'll be very interesting to see. So you mentioned soil moisture monitoring earlier as one of the future applications of this network. That makes sense for me from a agriculturist perspective, right? So farmers monitor soil moisture so that they can decide how much irrigation to deploy. But when you're doing this for your clients who might be a timber company or power lines or railroads. How do you imagine them gaining value from knowing the soil moisture? What action would that prompt them to take?

Carsten Brinkschulte:

Yes, for example, afforestation projects or reforestation projects are a little bit like a farm in a way. So they're literally farming trees and it's entirely possible and some of them do bring water into the areas where the trees are grown, in particular the small ones when they're still very vulnerable to drought. So it is indeed the same concept as you would as you have today already in IoT and farming just applied to forestry and with the networks in the depth of the forest where there is no mobile network. It is basically a soil moisture sensor that you would normally see in a cocktail.

Mike Toffel:

Okay, so let me pivot a little bit and talk about your funders. So you have different types of funders who have different objectives in funding your organization. Can you talk a little bit about that?

Carsten Brinkschulte:

Sure. Now we have raised money from several VCs, nine by now. We have some strategic investors that have an interest in the product that we do because it relates to their own core activities like steel, the world's largest chainsaw manufacturer has invested in the company for obvious reason because we do digitalization in the forest. have Semtech, which is a chipset manufacturer that basically owns the LoRa standard, which we use for radio communications. But we also have Telus, a mobile operator out of Canada that has invested significantly and is also a reseller of Dryad, actively selling our solution. So that these are strategic investors. We also have your financial investors, deep tech investors like e-capital. And we have impact on investors like Toba Capital from Los Angeles that have invested in Dryad because if we're successful, we'll have a significant sizeable impact and not just profits that we're aiming to do. So we have a portfolio and a broad spectrum of investors.

Mike Toffel:

What you're calling the strategic investors, those who own the technology or either the sensor technology or the network technology that you deploy, I can easily imagine there might be some trade off there in a founder's mind. In the one hand, it provides you preferential access, perhaps, to their technologies and advanced notice and maybe preferential pricing as those technologies continue to evolve.

Carsten Brinkschulte:

Yes.

Mike Toffel:

That seems very valuable. And it certainly seems valuable to the funders to see how their technologies are performing in the wild, as it were, to make sure that that's aiding their R &D efforts as well. But does it also provide maybe a little bit of constraints on your side? Because if you want to change sensor technologies, or you wanted to change network technologies, you have funders who have a vested interest in you maintaining the course that you've chosen. Is that a trade-off that you have to confront?

Carsten Brinkschulte:

It can be, and you have to be careful when you're accepting strategic investors that they don't put constraints on you. We actually do not have any constraints because I had to accept any constraints because we have the strategic investor. So Steele does not constrain us in any way. Tellus doesn't do that in any way. So no, we did not have to accept any constraints.

But it is definitely something that you need to watch out for if you're accepting strategic investors, because of course it's tempting for them to make sure that you work within their strategy and not go to use a competing company's product.

Mike Toffel:

Right, right. So let's look ahead. What is next for Dryad? Are you looking to deploy these technologies that we talked about a moment ago, the soil moisture and other types of sensors? Are you mostly focused in the next few years on deploying the wildfire technologies throughout the world, as you've been talking about, or some combination?

Carsten Brinkschulte:

It is indeed a combination. So we're really at the point where we're scaling the company. That is obviously a key objective because for financial reasons, you're going to have lots of money and revenue. But you also want to deploy millions of those sensors to achieve the impact that we're aiming to achieve with the technology. scaling is a key task and objective for the company. But at the same time, we also want to broaden the product portfolio to improve our competitive positioning and add the ability to upsell and cross-sell different applications into the same customers, that increases the value that we can extract from customers and at the same time the impact. R &D will continue to do a lot, will continue to be R &D heavy, while at the same time scaling the sales and marketing supply chain logistics.

We have to scale, of course, production because we are a hardware company, and we need to be able to manufacture millions of those devices when we're actually selling them. So manufacturing and scaling the manufacturing is a key aspect at the same time.

Mike Toffel:

So let me ask you a question that I ask all of our guests. I wonder what advice you have for folks who are thinking, I want to work somewhere in this business and climate space. I want to use machine learning or AI technologies. How do I get started in understanding the landscape and thinking about? What type of career opportunities are out there. Do you have advice for folks as to what resources they should consult or how they should think about the problem?

Carsten Brinkschulte:

Well, I think the key thing is to find a problem and a solution that somebody is willing to pay for. Because there are many, many, many problems related to climate. And there many, many solutions, but not all solutions have an actual market. And I think that that's really difficult in the beginning when you're imagining a new product and a new idea to identify solutions that also have a real market where there is a willingness to pay. Because if you don't have that, it will be very difficult to commercialize and scale your solution. then it will be very difficult for the solution to have an environmental impact. So this formula of impact for profit, I think that that's very important.

But it's also very important to quantify your impact. I see a lot of green tech companies or companies they claim to be where I'm not so sure how much impact they will actually have because everybody's building a business plan showing how much revenue you will be able to make and how much profits and costs and whatnot will be involved. But I think we also need to quantify our impact and quantify the future impact because that to me has to be a function of the business plan. You need to have a calculation just like you calculate and forecast profits and revenues and cost, you need to also forecast how much impact you will have. Because after all, you want to have an impact, right? And you need to quantify it. If you don't quantify it, I don't think you're a serious climate change company.

Mike Toffel:

Yeah, great advice, very important. Well, Carsten has been a very interesting, wide-ranging conversation, and I just wanted to thank you for joining us and spending your time here on Climate Rising.

Carsten Brinkschulte:

Thank you, Mike, for this pleasure.