The Billion-Dollar AI Opportunity No One’s Talking About - With Asseti’s Aonghus Stevens

0:02 Georgie Healy: Founders scale faster on Deel. Set up payroll for any country in minutes, hire anyone anywhere, get visas handled fast, and get back to building. Visit deel.com/dayone. That's D-E-E-L dot com slash day one.

0:22 Aonghus Stevens: In the next 5, 10, 15 years, we will see a switch to airspace being primarily unmanned priority, including, you know, you know, aircraft that move people around the world. And at the point that that happens, whoever controls that, that is an enormous opportunity.

0:36 Georgie Healy: What makes an incredible VC? A true partner while you're building your company, or someone that gives you money and then leaves you alone, Angus?

0:44 Aonghus Stevens: People spend 30, 40% of their time just like manually doing things, manually writing work orders, talking to vendors, you know, pulling data out in CSV format to merge it. You know, it's just an absolute chaos of a process.

0:56 Georgie Healy: There's an Irish drone supply chain startup called Manor. They managed to raise $30 million delivering things like coffees to residential homes. Is this lame or lucrative, Angus?

1:07 Aonghus Stevens: I would say that's lame.

1:08 Georgie Healy: Profitable business at 15 and then VC-backed AI-enabled startup. What are some of the misconceptions that you have around drones now that you are at that stage? Hello and welcome to In the Blink of AI, where I talk to the brightest AI startups and innovators. Each week. I did that from memory this time, guys. 30 episodes and I can do it now. I'm Georgie Healy, and this week I am speaking to the incredible Angus Stevens from Asseti. I actually met him last year as part of the AI Accelerator at Google, and I've been impressed ever since. I knew he would be great on the show because he's really not afraid for a spicy take, and he's a deep thinker, so I could kind of hit him with any of the headline news and questions and knew he would dive right in. We spoke about Angus's very early start as a teenage entrepreneur. We asked the Drone King, as I've called him, whether headline drone companies were lame or lucrative and why unsexy, in inverted commas, AI-enabled industries such as asset management are actually pretty exciting. especially if you're an investor. As always, we share our top AI hacks of the week, and we have our spicy takes to finish. This was a pure joy. Thank you, Angus, for being on the show. Let's dive right in.

2:37 Aonghus Stevens: You're listening to a Day One FM show.

2:40 Georgie Healy: Hey, Angus, thank you for joining in The Blink of AI. Before we dive into what you're building now, you're one of the most interesting backgrounds, I will say. And so I do want to dive into that a little bit first. How does it feel to be 24 and interviewed by Forbes. You're not 24 now, but when you were 24.

3:01 Aonghus Stevens: No, I'm not 24 now and I'm a long way past that. But, uh, you know, you just got to get on with it. You just got to keep doing things. It, uh, it doesn't matter what sort of what happens, I guess. Customers expect things from you and they're the ones that ultimately judge. So yeah, just gotta get on with it. Don't think too much of it.

3:17 Georgie Healy: So you weren't like me and just sharing it with and dropping Forbes magazines into like people's like pathways as they were walking past. You weren't, your ego didn't inflate at that time.

3:30 Aonghus Stevens: Pushing them out of the back of a plane. No, none of that. None of that. We, you know, it's good for exposure, those sort of things, but it doesn't, it doesn't necessarily deliver on what, what we need to as a business. So gotta, gotta balance it.

3:44 Georgie Healy: Yeah. There's a very adorable baby Angus photo, still on the internet for Forbes 30 Under 30, everyone check it out. It really speaks to what I find fascinating about your background, which is this passion for drones. I don't know if you'd still call it a passion now, but tell us the story about how you first found drones compelling and yeah, where that came from.

4:09 Aonghus Stevens: This is actually, I've been doing drones half my life now, so I, I'm 32 and I, I first read about drones when I was 16 or 15, so It does consume a bit, but I was reading a history textbook about, you know, drones.

4:21 Georgie Healy: As one does.

4:22 Aonghus Stevens: In sort of, well, I was at school.

4:24 Georgie Healy: Okay.

4:25 Aonghus Stevens: But it was about how drones, you know, had been used through sort of through conflict from very sort of early stages. And I knew nothing about that. And so I did a bit of research and, you know, one thing led to another and I sort of had stitched up 15 or 16 drone distributor agreements from, from companies across the world. They were very different back then. You know, it's almost 20 years ago. It was a lot used in agriculture and, and sort of really large, uh, large fixed-wing drones. I could never see someone, you know, flying a drone for fun, basically, which I clearly misread that. But yeah, the industry was very different back then, but it's been, you know, it's been a journey. Do, you know, do I love drones these days? Oh, I don't know. I love what they can do. Um, you know, and I think more and more. It's about enablement and using technology to do things. And that's where, you know, now drones are a fraction of what the sort of product is in its totality in Asetti. And I think they're an incredible enablement tool for enterprise particularly, but there's so much more, you know, there as well.

5:26 Georgie Healy: You said fixed-wing drones. How many different types of drones are there and what's a fixed-wing drone?

5:33 Aonghus Stevens: The drones sort of largely similar to aircraft fall into two or three Uh, categories. So fixed wings, like airplane style. Rotary wing has like a rotor, like a helicopter basically. And then the multirotor.

5:45 Georgie Healy: Oh, I'm familiar with those ones.

5:48 Aonghus Stevens: Yeah. Yeah. Yeah. The drone you see people fly in a park, that's multirotor. And then there's some hybrid versions of them as well. So there's like VTOL, which might be fixed wing with sort of vertical takeoff capabilities and things. So they largely follow the classes that exist in, in manned aviation as well. So.

6:04 Georgie Healy: The listeners would be so mad if I didn't ask why you'd use one for one case and one for another. Is there any clear-cut reason why I'd use one and not the other?

6:14 Aonghus Stevens: It sort of all revolves around, you know, what the use case is. So fixed-wing, large drone, much like if you, you know, you fly to Europe, you're not gonna fly there on a helicopter.

6:23 Georgie Healy: True.

6:24 Aonghus Stevens: It's usually inefficient. But if you, you know, wanted to, you know, fly around a canyon, you'd do that on a helicopter, not on a Boeing. And so it's a similar thing, you know, agility. There's a trade-off between the things. Usually to do with scale, accuracy of data, um, and those sort of things. The, the most pervasive drone is definitely multirotor. Yeah, they're, they're tiny, cheap to make, very sort of easy to control, which means they, uh, yeah.

6:48 Georgie Healy: That's the ones I see at the park. Yeah. My son really wants one. I'm like, you're 5, calm down. But if he starts reading history textbooks, I'll know, I'll know that, uh, he could be Forbes 30 Under 30, so maybe I'll get him one. Look, before we dive into more of the, the AI, you know, of it all, you are my go-to person to ask about the Southern Highlands because you're, you're often there, right? What do tourists get wrong about the Southern Highlands? What should they do when they're there?

7:17 Aonghus Stevens: There's so much to the Southern Highlands. I think that's the greatest sort of thing about it. There's so many little villages and those sort of things. If anyone wants Southern Highlands tips, hit me up. As you know, I, I'm good for Southern Highlands tips, but I can tell you where, you know, good vanilla slice is, where to get the croissants, where to get the coffees, and, you know, probably where to go to buy chickens. But that, um—

7:39 Georgie Healy: Chickens to eat or chickens to have as pets? There's a distinction there.

7:43 Aonghus Stevens: Well, I can tell you both, but I would have actually— I was more for ownership of chickens.

7:48 Georgie Healy: You're like, lemon slice, coffee, chicken.

7:53 Aonghus Stevens: They all went pulling into the same category. I did actually mean live chickens, but I can probably You know, good, uh, I actually don't know a good chicken shop down there, but, um, I'm sure there is one. I'll, uh, I'll circle up for our next chat.

8:04 Georgie Healy: I can help you with Silky Bantams, but if you want to eat one, I got nothing. Amazing. Look, we've got this part of the show where we call it Hack of the Week. It's a lot harder for me because I have to come up with a new one of these every week, and the guest always looks incredible because you've just got one. Um, so we're going to do that again this week and you can blow me out of the water. Can you share an AI tool or use case tools in AI that you try, you've tried recently and you love, that the listeners might benefit from?

8:35 Aonghus Stevens: Or of a workflow that I like doing. So I'm big on data. Yeah, love data. Love, you know, we've got dashboards all around the office, like everything in our business, we've got. That also sits in a data warehouse, which inherently means you've got to use, you know, SQL or something to interact with it. And I'm absolutely woeful of that. I'm completely hopeless. But what I do in, you know, ChatGPT, Claude, any of the sort of interface, I just dump all the database structures and I tell it what I'm trying to achieve and it writes like thousands and thousands of lines of SQL and it's amazing. And, uh, yeah, it's, uh, it's been a running joke in the office. So.

9:09 Georgie Healy: That is awesome though.

9:11 Aonghus Stevens: I can write a weapon. Well, no, it might not even be a good query. Someone would probably look at it and be like, what is, what absolutely is this thing? But I can show you.

9:17 Georgie Healy: But you get the outcome?

9:18 Aonghus Stevens: Yeah.

9:19 Georgie Healy: Yeah.

9:19 Aonghus Stevens: Yeah.

9:20 Georgie Healy: Uh, we were forced to do SQL at uni for Master's of Commerce with data analytics in it, and it already felt so archaic 2 years ago. So I think that that's a great hack. Like, why would you be learning SQL at this point, right? It seems kind of a waste of a skill set.

9:38 Aonghus Stevens: I don't even read what the errors are. I just like paste in back what the errors are. I'm just so lazy with it.

9:44 Georgie Healy: If it can do it. Mine is kind of hacky as well. I've been using ChatGPT to come up with learning courses for me, and actually I've been using it as of the last 2 days to write gym workouts for me as well. And I sent it to a friend who has had like a million hours of personal training, and she's like, I, I actually back this. I wouldn't pay for a personal trainer to use it. So in terms of learning courses, I've been asking it for AI-specific courses. Yeah. You know, getting the difference between tokenization and embedding versus like when tokenization is it using a subword or the history of it where they used to use a whole word and all of that stuff I'm getting through asking ChatGPT to teach me a learning course.

10:32 Aonghus Stevens: As well as being a hack of the week suggestions.

10:35 Georgie Healy: Yeah, I should use it for hack of the weeks because these are weirdly specific. I want to have a gym workout routine that exclusively uses handheld weights. It's pretty cool for that.

10:48 Aonghus Stevens: For my legs.

10:49 Georgie Healy: For my ankles. All right. I've called this section Drone King. How do you feel about being the Drone King?

10:57 Aonghus Stevens: I probably deserve to be in a mausoleum or something. I'm more like the old grandpa. Maybe I'm like Dirty Grandpa. Dirty Drone Grandpa.

11:07 Georgie Healy: Yeah, I could have, I could have been way meaner.

11:10 Aonghus Stevens: Drone King. I've got so many gray hairs these days.

11:12 Georgie Healy: Well, it's 'cause you're a ripe old age of 32, but, but seriously, you have been in business for long enough that I'm not surprised you've got a few. Profitable business at 15, and then, you know, VC-backed, AI-enabled startup. What are some of the misconceptions that you have around drones now that you are at that stage though? You talked about, you know, hobby drones, you've talked about, you know, them used in your past, but now that you're, you're building a scaled VC-backed company, is it so much about drones or a bigger picture? And what are we getting wrong there?

11:51 Aonghus Stevens: Yeah, and answer the question in sort of both parts. Like it's, it's never in a commercial sense, it's never about the drones. So, you know, why does someone buy a drone? They need photography, they need to be able to do something with it, or they need delivery, or, you know, they need space in the drone to do something. But no one, unless you're a hobbyist, you don't buy a drone just for fun. And that will wear off very quickly. It can only do so many things, you know, it's up and down, left and right. It's like, oh, okay, well there's, there's not much else to it. And that's sort of, You know, the utilization of drones to achieve something is where, if you look at Asseti and what we're doing today, we're an asset management solution around, you know, sort of predicting, planning, and operating property assets at scale. And as much like drone delivery, you know, it has a space in a courier framework potentially, but realistically people are going to use trucks to do this sort of stuff as well. You know, no one's just going to use a drone to deliver everything in the world. And so for us, drones play a part in our product. And certainly, you know, across sort of Australia and the US, we've probably got 5 or 6,000 pilots in our network now, but it's a very hyper-localized model where this data goes out and gets collected and comes back in. And then it, you know, is a source for us to do a bunch of other stuff, um, in the product itself. And so that's really where the, I think once the, once the drone is flown, that's where everything starts rather than, you know, some people sort of view that as like job done, but that's, that's just an enablement of a much bigger picture.

13:08 Georgie Healy: And the bigger picture in asset management, so those listening at home, Asseti Asset Management, why was this, uh, an area that you thought was particularly compelling, especially after years working in drones and the space? Why were you like, that's where it needs to be disrupted?

13:26 Aonghus Stevens: In our drone business, and we still have a drone business in, in the Asseti business as well. It sort of sits in there and we were basically making PDF reports for people and, and they would get a PDF report. It was like digital PDF report is the easiest way to explain. And they'd sort of print that off and, and do their thing with it and they'd pay us a lot of money for it., but they never use it again. And so it was like, okay, well there's actually a lot of value to this data. You know, an asset lives and breathes in its environment. It's like a person, you know, if you put a person outside for 50 years and did nothing with them, they'd be looking pretty weathered as well.

13:53 Georgie Healy: Yeah.

13:54 Aonghus Stevens: Yeah. That's, that's the same with an asset. It's like, okay, each of these assets has its own independent, like little personality and how can, how can we use data to better understand these assets? You know how they are now and how they live and breathe in the environment they're in effectively. And that's really what Asseti is about. Delivering for customers at scale, the ability to manage asset networks. So it might be 10 or 20 warehouses here in Australia, or it could be a few thousand sites in the US.

14:18 Georgie Healy: Mm-hmm.

14:19 Aonghus Stevens: You know, across areas is a massive scope and each of them has the same challenge of, you know, their assets degrade, you know, and, and that's essentially what it is. It's like, how can we slow down asset degradation or how can we make it cheaper to manage degrading assets? And that's, that's really like the, the trade-off at the end of the day. And that's what we're trying to, trying to build with Asseti.

14:39 Georgie Healy: It's huge, right? Like, one of my first jobs out of uni— I had a chemical engineering degree, and the options were in very crumbling-down manufacturing sites everywhere, I can assure you. But it's true that that stuff matters. Like, if we're working with equipment and machinery that's been out in the Brisbane sun for 30 years, it just has efficiency issues. And I can see, like, firsthand how much that impacted the final product in terms of what we were doing in a, in a food manufacturing plant. You guys use high-quality imagery to help with that conversation with your clients, um, and machine learning. Can you tell me a little bit more about that conversation and, and how exactly you, you use that imagery to help them?

15:27 Aonghus Stevens: Yeah, and that's sort of where, um, I guess part of the drones come in. So in a drone sense, we would fly. A drone over somebody's site, collect a few thousand images that process, um, or those images are reconstructed to 3D through a process called photogrammetry. Uh, and then that generates a really high-resolution 3D model essentially. So off that, that's when we do a range of, uh, sort of AI layers around defects and componentization. So breaking the asset down into like how many anchor points or how many fans or, or these sort of things across the asset network. Which sounds super unsexy, but if you've got like 1,000 warehouses or something, you don't actually know, you know, what you've got. And Then from there, it's, there's a range of other sort of solutions in the product around using data, you know, to, to do various things. The platform's expanding a lot more as well to provide essentially like an end-to-end solution for, for people who own assets. We obviously put data into it at the moment, but more and more it's around the workflows of, you know, managing condition and degradation in these assets. Um, and so yeah, lots of exciting, exciting things that are happening there. We launched, an expansion of the product earlier this year, AssetOS, which much like it sounds, an operating system for asset owners. So it, you know, it just does more and more, I guess, of trying to own the workflow for the customer and help them, you know, help them own what they need to.

16:43 Georgie Healy: Yeah. And you say unsexy probably because you've done this so many times too, that it's, you know, another day in the life. But if I knew that there was a drone coming to my site and it was going to take all these pictures, I would be excited. I would be quite pumped. Do you have people like looking out the window at the drone in all the pictures?

17:01 Aonghus Stevens: There are, there's certainly some characters. And yeah, we've got hundreds of millions of images now. But yeah, we've been across all sorts of sizes, like, you know, really remote warehouses, prisons, you know, schools.

17:13 Georgie Healy: Wow.

17:14 Aonghus Stevens: There's everything, like anything you can think of has been, been photographed. And there'd probably be, I dunno, 150,000 buildings, 100, maybe 200,000 buildings in the system now that, um, or in the platform that You're actively managed. And so yeah, we've seen most of it. Even had things, yeah, people throw things at it and—

17:32 Georgie Healy: Really?

17:33 Aonghus Stevens: Yeah.

17:34 Georgie Healy: And you've got an AI engineer that goes really deep into the weeds, but just for the listeners, what is his role specifically once you've got those images?

17:43 Aonghus Stevens: Yeah, so the AI, we sort of think of AI in two ways. The system one's like helping us get to our data outputs faster, which is more like an internal enablement thing. And then the second one's around like workflow, more like agentic style. Uh, hey, I, but in the workflow or in the sort of background processes, yeah, we, we run sort of dozens of like microservices or models essentially that are generating things like what's the amount of membrane roofing on this site or how much Colorbond roofing, you know, is on this specific asset or how many skylights, these sort of things. So each of them is specific to each sort of region and use case. Um, but it's really about building up, yeah, an asset hierarchy on the site and a defect register. So telling people who've got 200 instances of corrosion on this asset that need to be fixed, and these are the priority of the top 5 of them. Or, you know, trim these trees because you've got guttering and, and it's, it's an evolving, you know, it's constantly evolving over time that, particularly because we're going to new regions and things. And so, yeah, a lot of, I guess, process and enablement in the backend.

18:37 Georgie Healy: And as a business owner and, and, you know, having an AI company, compute is probably a big part of what you, what your capital costs go towards. How do you, how do you as a founder consider where to focus on improvement and where you're happy to spend the money versus where you're like, that is just not helping our customers and our end product and we can cut that out.

19:03 Aonghus Stevens: Compute is, it's an expensive place, that's for sure. Particularly because in our, so in the drone imagery space and other aerial imagery, 3D reconstruction uses a similar amount of like heavy graphics compute. So we would run dozens and dozens of servers for that. Yeah, I guess it's a trade-off of like what directly impacts the customer and what also what's working. Like for us at our growth stage, yes, you know, there's some, definitely some more stuff that could be automated or processes put in differently, but we are focused on building a product for the end customer. And it's like, hey, we can let this burn for the next 12 months in this process, but we know it's going to be completely removed anyway, you know, in a future iteration of the product, we're going to change something. It's like, Okay, then we, we sort of deprioritize that and yeah, then just like, just a ton of sort of intensity around being able to like focus on, on what needs to get done. And generally that, we wouldn't work on like a novel use case, let's say. For us, where, where we sit in, in the asset management space, our solution, you know, is, is sort of incredibly unique, but I wouldn't say we're at absolutely the bleeding edge of AI, you know, innovation, which is Which is positive for us because we can sort of pick and choose and rarely does AI solve one problem for us. So like in our series of 10 or 15 things in a process, we might put some things in there with AI, might, you know, some other just general automation, some other computer, you know, just like general processing stuff. And so yeah, it's, it's, I'm sure it's something in time that we'll have a ton of, a ton of sort of legacy issues.

20:39 Georgie Healy: Like, like every company ever. I remember when I was working at Ford, they had like this mainframe of like 1980s, like black screen, green text stuff, just because it was a legacy product that we couldn't, we couldn't, you know, upgrade because it had too many moving parts in the supply chain. You have reminded me of a chat we had once before. You know, you guys are an AI company that leverages AI, but you said, you know, your, your customers Like, it's not a brand thing that you guys use AI, like some startups out there that are like, we're an AI company, because your customers, you told me that some of them aren't even on the internet. Like, they're like, it's, it's not a branding strategy to push that down people's throats. How do you talk to your customers when you are a very technical company and they probably don't really care whether you leverage AI or not?

21:35 Aonghus Stevens: Yeah, I think it's sort of, yeah, it's telling, depends who you sell to, but for us, like telling a customer you use AI is like telling them our, yeah, website's built on a specific language. Like really, like really, it's like, how do we—

21:48 Georgie Healy: It's JavaScript.

21:48 Aonghus Stevens: Yeah, it's like, how do we help them get to their output quicker? So like in the terms of asset management, as soon as that insight, you know, is, is detected but not presented, it starts to decay in its value. So if it takes you an hour, a day, a week, a month, whatever to get it out, then all of a sudden the value of that insight is significantly reduced. And so it's like, how, how can you provide a real-time, you know, sort of insight for the customer? That's really dry. Um, well, that drives a lot of our, our design. And then how can we help them complete a workflow? So AI, you know, is integral in all this sort of stuff, but it's not our AI-enabled XYZ because sometimes they don't even know what the XYZ is anyway. You know, they know they've got to do this, but they also only do that because. You know, that's how someone who sent them an Excel spreadsheet has done it in the past. And so that's for us is like, you know, how can we help them get to the insight quicker and how can we close the loop for them? And if we can do those two things, you know, in, in parallel, then we make their life better and faster for that, you know, make their job faster. Um, and that's core to our success, you know, whether we roll out 10 or 100 users of AI in our marketing over the next year, reality is no one's probably going to care. And if they do care, they're probably not the decision maker realistically.

23:00 Georgie Healy: I see. And, you know, whenever you say workflow for asset management, tumbleweed goes through my brain because I have no idea what that looks like. What is a workflow for an asset manager and where is a like common bottleneck in the workflow that you guys are like, we could sort that out?

23:19 Aonghus Stevens: So there's a ton of like just process-based stuff. So what we do in Asseti, you know, defect register as an example. Previously we were just part of the problem that we just generate these defect registers and everyone's like, great, what do I do with it? Like, this is amazing. Like now I just have a ton of defects I don't know what to do with. And so that would usually go into a third-party system, like a, an ERP or enterprise asset management system that they would—

23:38 Georgie Healy: Yep.

23:39 Aonghus Stevens: You know, create a work order and then they would assign it to a vendor and then they'd monitor that vendor to make sure the work got done. Then they'd close that out. For us, a workflow is that whole thing. So in terms of it's like, how can, you know, whether we create the defect or somebody else does, how can we then feed it into a, an ERP or a system like that? And how can we then take that data out? You know, assign the vendor, you know, assign any reworks on the work order that are required, close that out, update the asset register that's been replaced. So in 12 months when you do an asset audit or something like that, it's not, oh, we've now got 2 air conditioning units that we don't know which one's right. It's like, oh, well that one got replaced. This is the new one. And yeah, that workflow piece, like in asset management particularly and facilities management, people spend 30, 40% of their time just like manually doing things, manually writing work orders, talking to vendors. You know, pulling data out in CSV format to merge it. You know, it's, it's just an absolute chaos of a, of a process.

24:32 Georgie Healy: And I'm guessing not very, like, it would be all like work order numbers and equipment numbers and just like a sea of numbers that like, I guess it would be easy to have human error if you're doing it in that way, right?

24:45 Aonghus Stevens: Yeah, people just don't, you know, in, in a lot of these legacy systems like ERPs and stuff, they're really powerful systems. But the information that sits in them is typically woeful. And therefore everything that is in the system is usually woeful because it's just human entered data. That's just like, someone's like, oh, I know I've got to put this field in, but I skip it every time because this is how I skip it. You know, it's like, okay, well that's like a required field for a specific reason.

25:07 Georgie Healy: And then someone's running— Probably shouldn't.

25:09 Aonghus Stevens: Yeah. And so with a lot of these legacy systems and processes, it's rather than, you know, for us it's not about trying to replace any of that. It's how can we actually make those effectively modern day systems and this process of modern day workflow to actually make asset management operate in the way it should in, in 2025.

25:25 Georgie Healy: Yeah. Another thing that floats into my brain, the mining industry, is that a compelling customer for you? I know a lot of your customers are US-based, but I'm thinking of, you know, if you're an oil and gas company and your equipment is a safety risk, if it isn't maintained effectively, long way around saying Are there certain customers that are more compelling for you and it's an easy sell or not?

25:54 Aonghus Stevens: Yeah, so we wouldn't typically push into mining or oil and gas, or sorry, we wouldn't typically say push into oil and gas. Mining we do a bit in, but for them, they've got so much money to make sure something doesn't go wrong that, you know, of course they're budget constrained, but realistically there's a significant pool of money there.

26:07 Georgie Healy: They will allocate.

26:08 Aonghus Stevens: Yeah. Whereas for us, we're very much about distributed, you know, networks for a customer to sort of feel the pain that a SETI solves for them. It's best if they've got lots of locations, you know, or they've got either a massive site or they've got tons of locations. Um, and that's why large format real estate typically, um, is benefit. Um, or is a, is a sort of benefits from the system, industrial, retail, government property, these sort of things. So I could tell you what most warehouses look like, but I, uh, yeah, where all your sort of goods get shipped from in the US and Australia, I've, I've seen pretty much all of those, but. Not, uh, yeah, not, uh, not necessarily the houses or those sort of things. And it's like, you know, you got to feel the impact at scale. So if you've got a roof that, you know, might be 9 or 12 football fields, like that is, and you say, I'm going to send someone on the roof to look at it. If you send someone out to look over just one football field, they're not even going to find everything that's on there. You know, that is, is, is just not going to happen. And in the US particularly, been in Australia, but the US, it's really hot in summer, you know, and in Arizona, like these roofs. As an example, just like crazy hot. And so you can't even put people on them for parts of the year, or they don't really want to spend extended amounts of time up there when they do go.

27:21 Georgie Healy: Yeah, and the safety element and all of that. Um, are you gonna do all of OpenAI and Microsoft and whatever the hell's, uh, data centers that they're building? Is that, is that on your, your go-to list of outreach?

27:34 Aonghus Stevens: Data center is certainly an interesting, you know, space for us. Um, and a lot happening there. I think the funny, the funny part, I guess, seeing in these industries as they move and change, and we're involved in a lot of, um, probably the early enablement work for, for sites too, just generally. And so it's a big explosion of data centers that, that has occurred. This land was acquired years and years ago, you know, it's been, it's been in the pipeline for ages. They're investing in it now, but a lot of this.

28:02 Georgie Healy: Okay.

28:03 Aonghus Stevens: A lot of this has been in play for quite a significant amount of time. And so yeah, it's interesting seeing customers and they're expanding sort of networks and particularly in the US, you know, a lot of investment in data centers.

28:15 Georgie Healy: It's funny you say that. I remember when there was the announcement earlier this year about the OpenAI data centers and the, you know, the growth of that and how much money was being spent. But apparently it was like a Biden administration thing that was approved initially. Like, as you say, it's not like Trump was signing it off very publicly. Politically, but it wasn't actually about him or anything to do with recent political movement.

28:41 Aonghus Stevens: Yeah, we've, um, some of those sites that, you know, would sit in a city have been in there for years. Yeah. Really? Having prior monitoring done on them when they were greenfield sites or something like that. So yeah, we're across a range of customers who have exposure to data centers and industrial and, and those sort of things.

28:57 Georgie Healy: Are you a bit jealous of the founders that, um, when they go see customers, it's in like Paris and Rome and you're like deep Arizona in summer.

29:07 Aonghus Stevens: Nah, I see some, uh, I see some, I think probably the most authentic parts of the world. I, uh, yeah, we get, we get around. It was funny. Uh, we were in Dallas for a, for a trip. I'd go to Dallas a bit and we went to visit a customer, you know, a bit out of, bit out of town. And I didn't think we'd be getting out to, uh, get an Uber that day. So I hired a car and, uh, wasn't too sure what I hired. And this blacked out Lexus turns up. And when we turned up to the customer, we, it was like, it looked like we were repossessing the place, honestly. Like, We hopped out.

29:35 Georgie Healy: Were you wearing a suit?

29:36 Aonghus Stevens: I don't think I wore a suit back then.

29:38 Georgie Healy: Had a clipboard?

29:39 Aonghus Stevens: I honestly was like, who are these absolute tossers who just turned up?

29:45 Georgie Healy: Oh, now you got VC investment, you think you can get a limo?

29:49 Aonghus Stevens: It was cheaper than an Uber would have been, but it was funny once it turned up, I was like, oh, this is like—

29:54 Georgie Healy: Oh, they already hate us. Funny.

29:57 Aonghus Stevens: Like 2 or 3 hours into rural Texas from Dallas and, uh, And there we are. Just, yeah.

30:03 Georgie Healy: You got a printout of your Forbes 30 Under 30.

30:07 Aonghus Stevens: Was that, that was quite funny.

30:11 Georgie Healy: That's classic. Okay. I would love to have your expert opinion to help me unpack a few headlines. It's a section I've never done on the show before, but you're, you're a good sport, Angus. So I figured let's, let's give it a go. It's called Lame or Lucrative. I'm going to share some drone-related headlines. And I want your hot take. Are you excited?

30:35 Aonghus Stevens: Oh, absolutely. Nothing more excites me than this.

30:39 Georgie Healy: Your eyes have lit up.

30:41 Aonghus Stevens: Let's go. Let's see if I'm going to annoy the industry. Let's go.

30:44 Georgie Healy: You're like, they're all lame. Next. Okay. There's an Irish drone supply chain startup called Manor. They managed to raise $30 million delivering things like coffees to residential homes. Is this lame or lucrative, Angus?

30:59 Aonghus Stevens: I would say that's lame. I think the, you gotta deliver a lot of coffees to, to win there. And yeah, there's a lot happening. There's been a lot of promise, but I don't think the future in drone deliveries is as glorious as is potentially made out, at least in the near term.

31:20 Georgie Healy: Yeah. And, um, just, just because I'm curious, is the, like, it feels like the drone is so expensive, but the coffee is so cheap and like, is human labor just cheaper than a drone to operate and service and do stuff like that? Or—

31:35 Aonghus Stevens: You know, drone delivery, there's a big thing at the moment, which is like, how do you integrate drones into manned airspace? And that's actually where I think the game is won, is if you can control who flies, not who— if you can control who gets the coffee from A to B, but you're not the person who gets it from A to B, you win in the industry. And that's the bit that with drone deliveries is still fairly open right now. There's a single global provider for, you know, manned aircraft information, like flight aircraft information, essentially like air traffic control. And in drones, it's a similar thing, which, you know, Google, their investment in, in Wing, which I think, uh, we'd spoken about before. They, um, I think that's where Google wins. Google's not interested because they can make $2 on a coffee.

32:22 Georgie Healy: Yeah.

32:22 Aonghus Stevens: Google's interested because there's a billion-dollar data sort of play there. And the data plays around drones, you know, and the management of that drone fleet in inverted commas.

32:34 Georgie Healy: You did tell me about Wing because that was going to be my next lame or lucrative. So, so thanks, Angus, uh, ruined that one. Um, no, let's tell the listeners, um, it is an Alphabet or Google subsidiary that delivers groceries, food, and more, but it's the big tech incumbent. And you think this lucrative because they own, they own the airspace, like they own the ability to provide that service. Is that what you're saying?

33:06 Aonghus Stevens: Yeah. So I think Wing's drone delivery service is lame. I think Google's play at drone delivery is lucrative. So, you know, Wing, um, has a bunch of data products that they, uh, and platforms that they're building much like the Google, you know, solution generally is, uh, And they've done hundreds of thousands of flights and their IP that they're building is the ability to get a drone from A to B in congested manned airspace, not the ability to get your coffee from A to B. And that is really what the big learning is there. You know, they're not, um, truly my view, but I don't believe they're interested in it because they reckon they can monetize the—

33:45 Georgie Healy: They can upsell that oat versus skim milk, Angus.

33:50 Aonghus Stevens: And you see how they've changed. So like it started off down in Canberra. Then it went up to, to Queensland, to Logan from memory, and sort of gradually been increasing the, the density of where they're allowed to operate. Like when it first started in Australia, I was in, I think it was in like Bungendore or something down outside Canberra, which is like no one was getting anything delivered there, but it was a good use case. They're arbitral, they're regulator. Then they've gradually become more, yeah, sort of built up through, through more congested airspace. And the way manned airspace, or the way airspace is assessed at the moment, general and regulatory senses, like, manned airspace is the priority. But Boeing and Airbus have sort of, in their view, generated their last sort of edition or last version of manned air— of sort of piloted aircraft, you know, primarily piloted aircraft. And so in the next 5, 10, 15 years, we will see a switch to airspace being primarily unmanned priority, including, you know, aircraft that move people around the world. Um, and at the point that that happens, whoever controls That, that is an enormous opportunity.

34:50 Georgie Healy: Wow. Genuinely very fascinating.

34:53 Aonghus Stevens: That's a really long answer to lay my look with it.

34:56 Georgie Healy: No, I loved it. I loved it. Uh, my last one though is a bit left of field, maybe a little bit outside of your sphere of interest, but flying cars. It's still manned airspace, Angus. Um, there's this, there's this company, Klein's Vision Air Car prepares to take flight commercially in 2026. This revolutionary vehicle promises to transform our daily computes. So we're rising above traffic congestion, Angus. No longer do we have to get on the highways to go from Sydney to the Southern Highlands, and we're going to be in flying cars. Lame or lucrative?

35:33 Aonghus Stevens: I think that's lame.

35:36 Georgie Healy: Tell us what you really think. I think beep beep.

35:44 Aonghus Stevens: I don't think much is going to happen there, to be honest. We thought, you know, flying cars are going to be a fly— drones, manned drones are going to be a thing that existed 10 years ago, but I'd rather buy a manned helicopter or something. There's just something really satisfying around like the sound of an engine. Whereas, you know, I think in a drone, a flying drone, like if those engines were to stop, that would be a very quiet experience.

36:06 Georgie Healy: Creepy, right? Um, you don't see a world where we're living like the Jetsons then.

36:12 Aonghus Stevens: I think we're a long way off that, but I'm really happy to be proven wrong and that it's an extraordinarily amazing, you know, implementation and sort of industry expansion over the next few years. If that's what I get wrong, so be it. I'm more than happy to.

36:25 Georgie Healy: Let's stick to the Waymos for the short term. Hey, you have been such a great sport. We are going to finish the episode with our rapid fire round. So kind of spicier questions. You can handle it though.

36:39 Aonghus Stevens: How good.

36:40 Georgie Healy: We've warmed up, right? What makes an incredible VC? A true partner while you're building a company or someone that gives you money and then leaves you alone, Angus?

36:50 Aonghus Stevens: For our stage, partner. You know, we got an amazing VC in Tidal. Don't tell them that, but we love working with them. And we have very sort of strong execution skills, but some, you know, areas where we need sort of positive, significant contribution for our VC. Um, and so I'd say, yeah, partner for us at this stage.

37:11 Georgie Healy: That's awesome. Uh, anything in particular that you think you should shop around for VCs for, for any other founders listening?

37:18 Aonghus Stevens: Still a character reference, you know, get them checked, but also don't raise, like, don't go to market when you need the money. For example, Tidal, when, when they joined our cap table, I think we spoke for them for, you know, 9 or 10 months and by the time we sort of, you know, got through, got through to a term sheet and things, then we demonstrated some stuff. They demonstrated they could work with us and add value there as well. And yeah, it's been an amazing experience so far with them. Doesn't mean everyone agrees all the time. And in fact, you know, that'd be bad if they did. Um, but we are incredibly well supported and sort of on our way, which is a positive thing.

37:53 Georgie Healy: Yeah. So supportive in fact that Kieran from Tidal, I think has a big cardboard cutout of you in his house. So, um, Yeah, that's what you want to see.

38:03 Aonghus Stevens: It's on my desk now. I was on a meeting when, so when he first brought it over, it sits behind a cupboard actually, but it, um, I had it behind me and I didn't realize about halfway through I just saw me there. I was like, oh, I was with some people in the US. I was like, oh my God, I look like an absolute—

38:20 Georgie Healy: The biggest wanker.

38:21 Aonghus Stevens: Yeah, like don't, like let's put blur on.

38:27 Georgie Healy: Oh my gosh, they would have thought you got that printed out. For the listeners, that happened because Angus was part of the AI Accelerator at Google and it was printed out on demo day. But now it looks like you did that.

38:39 Aonghus Stevens: Particularly by the third or fourth meeting where I kept it there. But no, it was, it was probably put beside the cupboard then behind my desk.

38:47 Georgie Healy: Oh, that's the best thing I've ever heard. I remember seeing Kieran walk out of the building with it. It was like I got so many photos, nothing's been funnier.

38:55 Aonghus Stevens: I think he bought, so their office is a couple of Ks, a couple of Ks from us here. I think he just walked through the city CBD with that, um, and that up there, but he's never had so many compliments in his life. So, you know, clearly he peaked that day.

39:09 Georgie Healy: Yeah, yeah, yeah. You guys are a beautiful, beautiful partnership. Um, snog, marry, or avoid? No, snog, marry, avoid. You've got to pick one of each for Google, Microsoft, AWS.

39:23 Aonghus Stevens: Ooh, I will pass this question because we are multi-cloud and all three support us in various ways. Don't bite the hand that feeds you.

39:34 Georgie Healy: I marry them all. Look, fair enough. We don't want to get you canceled. How about I switch this one out? Snog, marry, avoid. Jeff Bezos. Sungshin Bichar. No, I'm kidding, I'm kidding, I'm kidding.

39:52 Aonghus Stevens: Snog them all.

39:53 Georgie Healy: Snog them all. Get cardboard cutouts of them all. Did you watch Google I/O this year? Yeah, can recommend.

40:02 Aonghus Stevens: It's kind of cool. I'll be sure to jump on and have a look.

40:05 Georgie Healy: Yeah, yeah, I'll send you a link, don't you worry. If you got Jeff Bezos level of wealth, would you quit tomorrow and chill out on an island somewhere, or would you keep grinding away?

40:16 Aonghus Stevens: I love working. I don't work for money. I think that's risky when people do, but I love working. I love, you know, solving the problems we do. Oh yeah, I'd probably buy more farmland or something, buy a bigger farm, or I don't know. That's, that'd be like, that'd be my, my sort of my thing, I imagine. But no, I wouldn't stop working.

40:37 Georgie Healy: I really like Yeah, I do get that sense that you genuinely love what you do. Any aspect of your life where you actually prefer form over function? Because you have told me in the past that you're very into function over form as a business standpoint. I am.

40:55 Aonghus Stevens: I'm big on utility, so if it doesn't provide utility, then—

40:58 Georgie Healy: Your poor wife, she must really struggle.

41:02 Aonghus Stevens: I do love a nice car. So that's, um, looking at them. I don't, I don't know. I won't buy one. Um, cars are lovely. No, I think like I'm all about utility. I, uh, if it doesn't provide, you know, sort of functional or utility for us in, in a business or, or sort of otherwise, then, uh, yeah. Or maybe, I don't know, maybe with, maybe with kids or something, but they're sort of isolated, they're separate.

41:29 Georgie Healy: But yeah.

41:31 Aonghus Stevens: Yeah.

41:32 Georgie Healy: Um, okay. And the last question, who do you admire most in the Australian tech/startup ecosystem? Who would you shout out to?

41:41 Aonghus Stevens: That is an interesting one. I don't know if there'd be a specific person per se, but I really appreciate the journey that companies in an enterprise or sort of like B2B sense have gone through in similarly unsexy spaces as us. So, you know, the, the sort of journey that a safety culture or employment hero or a company like that, um, has, uh, has been on where there's just a lot of grinding, but also very sort of specific domain understanding around what the solution is. You know, I, I, for us, we're not trying to build a solution that we don't know the industry. Like we have an intimate, intimate understanding of the industry, what we're trying to deliver. Yeah. It doesn't make it easy. Um, and so yeah, I appreciate any of those, like, I don't mean grinding it out in like a hustle sense, but I mean just like year after year of building this solution with a laser focus on where, um, where that sort of business needs to be.

42:39 Georgie Healy: Absolutely amazing, Angus. This was so fun. I learned so much genuinely about drones, which kind of VCs to like partner with. Um, what's lame and lucrative in the airspace category. You've been such a great guest. Thank you so much. Before we close out, uh, is there anything you'd like to shout out to listeners? How do they find out more about Asseti, more about you? Where do they subscribe? What do they do?

43:08 Aonghus Stevens: If you want to find out more about, uh, what we're doing in the city, it's just asseti.com. A-S-S-E-T-I dot com.

43:14 Georgie Healy: Hey, those YouTube watchers.

43:17 Aonghus Stevens: Uh, asseti.com. Yeah, we're sort of transforming the way people manage assets. And if you know anyone who has a distributed asset network or has challenges with their assets, hit us up. But yeah, sort of, we sit in a part of the world that you probably don't even think about usually, which is sometimes the most important part of the world. There you go.

43:38 Georgie Healy: And if anyone wants to see baby Angus on Forbes, go to Forbes 30 Under 30, Angus Stevens. It's an adorable photo. Thanks so much, Angus. Bye.

43:48 Aonghus Stevens: Thanks, Georgie. Bye.

43:51 Georgie Healy: Thank you for listening to In the Blink of AI. You can check out the show notes for anything discussed in this week's episode, and we will be back next week. This podcast was produced by Day One with music by Dan Hansen and visual artwork by Sophie Tyrell. If you loved the episode, please tell your mates, and I love AI news. Please share your thoughts and suggestions Questions to georginarosehealy@gmail.com.