The Flying Supply Chain, with Isaac Roberts (CSO of Skyways)

You identify a problem or a need, and then you go and you build to it, and it's going to take deep tech engineering with a lot of focus to make that thing happen, and then you just sell the living crap out of it. Skyways has been focused on the defense department and saying, how do we do something that's actually meaningful? And so it's always been that payload and range increasing over time. We're starting to see reindustrialization occur, but I think the drone space is having a lot of pressure to get mature really quickly. Okay, let's put a missile on it. Let's go from camera to now we'll shoot at the things we see on the camera kind of stuff. One-third of casualties in the Iraq and Afghanistan wars were in logistics. No one's been going and saying, 'How do I move an ammo can from the ship to the beachhead and do it in a way that's not costing anybody their life?'

So, we're at the Skyways office. Behind you, there's half a million dollars in tooling. That one piece right there, we just got in for V3 Block 2. So this is the production floor of the company. Here's where we make all the final assembly and get everything ready to go out.

Yeah, for sure. Now, I can't but not point out, like, we come to a lot of these types of factory floors and production lines, and we usually see a giant American flag, but there's three here. There's an American flag, a Texas flag, and a...

Yeah, so obviously we're in Texas, right? So, as you know, there's a lot of defense tech happening in the Austin area, but Charles, the founder, he's from France, which actually he's over there right now with family. But it's this, let's say, like homage to origins. He, funny enough, he actually grew up next to the Boeing, or sorry, the Airbus facility over there. So like aviation's been in his blood since he was really young, been building planes, and started working over at Google as a software engineer and decided to merge the two, and then that's kind of how Skyways came about. So the French flag...

Yeah, in some cases more airtime. So...

Yeah. Exactly.

We see logistics as fundamentally broken around the world, right? So this started, Charles experienced it when he was over in San Francisco working at Google. He's crossing the Bay Bridge, dealing with all the traffic, getting to the office, and he's just like, 'Look, this traffic sucks. How do I avoid that?' And then the idea of a self-flying car. He's got a pilot's license. He's been doing aviation for a long time, so he understands the problem pretty well. Was looking at Waymo and how that was all coming together. So, wanted to tackle it. And that was about 10 years ago now. For me, another anecdotal item is like I just flew in from Colorado, right? So, got in at 1:00 a.m. after getting on the road at, you know, 5:00 p.m., right? So, we get in our little boxes, drive to the next box, get on a train, go over to the airport, get in the right terminal, get into the plane, fly. I'd much rather just get into my backyard, get in a flying car, and go to sleep while it's taking me to Austin, right? And so, I think that we kind of take for granted the potential. But there's a lot of complexity involved, right? So, what we're trying to do is not necessarily solve the problem of, you know, container ships that are going transatlantic or something like that. It's like these big bulk cargo systems like trains would be able to accomplish, but rather the really high value, really high need, urgent items getting done in a bespoke manner.

Yeah. So, you can think of us as like Zipline for defense and B2B.

So, I didn't know this until maybe a year and a half ago. But one-third of casualties in the Iraq and Afghanistan wars were in logistics. So, you think about that, like one out of three casualties was in some kind of convoy trying to handle some kind of logistics problem. That's a major issue, right? I mean, those are lives lost. Diesel fuel transport was a huge issue and actually one of the probably leading causes of troop deaths during Afghanistan. And if you're running an insurgency, that's exactly what you do, right? You cut the supply chain. So I think that the idea is that you can automate that supply chain for high-value items, allowing for a smaller footprint and a larger reach.

No, I think that's right. So, there's this anecdotal item: Walmart, when a hurricane starts to come to Florida, they move more strawberry Pop-Tarts into the state. You know that?

There's this data analog with like, you know, world events, something happens, there's a storm coming or whatever. I grew up in Florida, and so it was like every time a hurricane came, it was always hurricane party. You go to Publix, you buy a cake, you hang out at home. The end, right? There are some things you can predict, you can see coming from a mile away because hurricanes only move at 25 miles an hour. You know, tomahawks move a little faster, right? So, when you end up in this world of hyper-reactivity, the model for Amazon or Walmart moving Pop-Tarts starts to break down. And it's like as the world moves faster, data models that are built up to support those infrastructure items fall apart. So what Amazon's doing is just as the case you mentioned, they're staging things at various warehouses and they're monitoring the web traffic based on the location that it's being delivered to, and then they're figuring out how to sort like, okay, well, these battery packs will go over to this facility in San Antonio. These other battery packs are going to go over to the one in Austin, and people that go to the website in Austin are served up one battery pack as opposed to the other one in San Antonio. And it happens for like broad-stroke stuff, but defense is far more, let's just say, agile, reactive. You know, there's like a lot of conversations around staging. How do these things start to mesh when you have response needs that are, you know, mission-critical, right, and very short-term? So we're trying to build up a network of capability that has a much higher level of flexibility for the things that are actually like super urgent, right?

I mean, Amazon... Yeah. What was Amazon's batteries or like Amazon's what's their...

Yeah, Amazon Essentials. Exactly. It's like, yeah, let's just copy that because people seem to like that backpack. We'll make it ourselves and stage it all over.

Yeah. So, when Charles started the company, I think before he even left YC, he had two units purchased by the Marine Corps, right? Based off of General...

Like 9 years ago now, I think.

Yeah. Since the inception. Yeah. So they got something that was the V1, and it was all-electric eVTOL, but still a fixed wing for forward flight, right? But the propulsion was all electric, and so that means you've got limited payload, limited range, and the batteries are really heavy for it, right? So it was able to do around maybe 5 lbs of useful load and go maybe 50 miles. They got it and they were saying, 'Oh, this is amazing. How do we go and get more payload, more range?' And the Marine Corps, not exactly the biggest budget in the DoD, they moved it over to the Navy, and then Navy took it on as a part of the Blue UAS program, which is run by Bill McChion, or at least he's been our liaison to the best of my knowledge. And so Bill took it on and said, 'Okay, well, the Marines need this, but also this is really, if it's the Marines needing it, it's a Navy problem, and the broader Navy has this issue because it's ship-to-ship logistics, too, right?' So ship-to-ship logistics means you're not talking 50 miles. You need an order of magnitude more. So Skyways moved over from all-electric to a heavy fuel engine, and that was the version two. And so the Navy contract was for multiple years, but it was clearly take the same capability, put a heavy fuel engine on it, hybrid system for the eVTOL to continue to be capable, and then that was going really well. Went out and did some events, beat companies like Boeing on it, and then moved over to an Air Force contract.

Yeah, Shield AI is an example of one company that we beat out. Boeing was another. There's been a lot of competition, but nobody's been focused on the single problem that we've been building on the entire time, which is logistics, right? So, I mean, if you look at the drone space, I don't think I need to tell you or the audience, but if we just break it out for anybody that doesn't know, drones have been monetizing ISR for a long time, right? So, everybody's been saying, 'Oh, well, there's a thing that flies itself. That's cool. What can I do with it? Let me put a camera on it and look, I can go see things.' That's cool, right? Real value, the ScanEagle. Then of course you get into MQ-9, you know, Predator, Reaper systems. Okay, let's put a missile on it. Let's go from camera to now we'll shoot at the things we see on the camera kind of stuff, right? No one's been going and saying, 'How do I move an ammo can from the ship to the beachhead, right? And do it in a way that's not costing anybody their life, right? Because it's one ammo can instead of hundreds.' And it's this idea of like taking an entire supply line and moving it over to a more bespoke manner. So there's not been an obvious need because it's like everybody's looking at what is going on as opposed to what's going to happen, right? And so when you look at the logistics need, like the VBAT, it's able to maybe take a couple packs of pencils, right? The V3, you look at it, it's 7 cubic feet of contiguous space, right? That's payload capability, whether it's a modular payload or a whole bunch of cargo, whatever that is. And it's just singular in its capability, right? It's also very difficult to build because the closest competitor I can think of is Zipline, and they've, I mean, they're doing great, but man, do they have a slog, right? And they were, you know, raising a lot of money, going out and servicing Rwanda, doing a whole bunch of things. And then, of course, they don't want to get into defense, but now they're, you know, delivering burritos and stuff, right? And so the easiest way to think of it is Zipline was doing the same thing but focused on the commercial space, focusing on how do I get that Amazon or, you know, that burrito delivery, coffee, whatever it is, whereas Skyways has been focused on the defense department and saying, okay, how do we do something that's actually meaningful? And so it's always been that payload and range increasing over time, where Zipline can go and say 5 miles is good enough if you can carry a burrito and a coffee, right?

So, it was definitely performance that we beat them on. Performance. The funny thing is though, it's like, you know, you look at a big defense prime and they've got a million and one things going on and all these contracts that are getting paid on and they're resource fighting, right? Like they're trying to fight for resources to get this project, that project done. It's hard to fight for a project that hasn't yet been birthed, right? So, there have been multiple companies that have gone out and tried to do similar stuff, but they've been operating inside of bigger companies and then they just don't get the resources, right? Skyways on the other hand has been just a team of like grinders trying to solve this problem because they believe that it's a real problem and they all want the flying car, right? And so they realize strategically, and this is like a philosophical difference, but like companies like Joby, Archer, Beta, they've been pushing towards the flying car in a way that's like, well, helicopters are gas powered, they should be electric, and then that's what drove them. Skyways kind of went the opposite direction, but it's this modality of the customer doesn't care about ESG. The customer is like, you need to be able to go 1,000 miles with like 80 pounds. Like, how do you do this, right? And so that spec is a big difference. And so I think that it comes down to resources and focus on the long-term objective.

Yeah. So, this has been my song and dance internally, advocating for it. But like, what's really interesting is the Department of Defense, they don't really care what the FAA says, right? They're kind of like, we take over the airspace because, I mean, if you were to, then you'd have to go through type certification, supplemental type certification, and like a lot of these companies like Joby and Archer have been tied up in that for many, many, many years and probably still have a number of years to go. And you see them transitioning over to our frame of thinking, right? Which is like, well, regulations are killing us. We got to get into defense. Let's go and build something. And now there's an Archer-Anduril partnership, right? I think that Beta is now pushing into cargo. So our path has been start small, get bigger, focus on cargo, transition to people. The V3 is 100 pounds, 1,000 nautical miles is the spec target broadly speaking. But when it comes to the V4, because there will be a V4, my advocacy has been for about 1,300 lb and about 400 lb of useful load. And the reason is because you're seeing Part 108 on the commercial side with the FAA. And to be clear, the US is not the leader in this space right now. It's very much a laggard. Like China is going out and just saying, 'Hey, you want to fly? Like, get after it, you know.' But the US is more concerned about, you know, crashes, and rightfully so, right? Nobody... the demand for not crashing on my house is bigger than my demand for having a flying car, and at the end of the day, it's the body politic that decides that, and government is a laggard. So with that being said, the V4 is targeted at this Part 108 as the, let's just say, like regulatory framework that we can operate within. 1,320 is the max, it's the MTOW, max gross takeoff weight, I believe. That's their target right now, and the general rule of thumb that we operate under is about a third of the airframe MTOW is capable of being useful load, fuel, or cargo. So if we make the V4 and it starts to take on the same kind of mission profile that the V3 does, which is, you know, moving cargo ship-to-ship, ship-to-shore, all that stuff, island chain logistics, etc., then we have a V4 that starts to look like, okay, you can carry 400 pounds of cargo, you get hundreds of millions of flight hours on that airframe, then all of a sudden it's like, oh, this might be safe enough to carry a person. And I think that's the way that the flying car starts to actually unlock.

If you take it down to first principles, I think you see if you build a flying car and you want to like get off the ground with, you know, pocket money, right? You identify the things that don't change. I really like this philosophy from Bezos. It's like you focus on the things that don't change. What are the things that don't change in a flying car? It's like you're going to move some kind of weight, right? And you're going to do it without a pilot, right? And you're going to ideally, I mean, this is the whole idea of a drone. The presupposition is that it's going to be in the air, right? So, it's going to be in the air moving some kind of weight and it's going to be autonomous. Then you can start with something small and scale your way up, right? And this is a pretty common practice in, you know, aerospace, right? It's the scale-up modality. Like, you build a subscale system, then you get bigger and bigger. You prove out the concept because it's either you're touching a little layer or you're touching a lot of air, but at the end of the day, the air is going to be the same across the entire airframe. So with that being said, I think it was really intelligent for Charles and the team early on to go and say, we're going to just focus on something small, get some utilization on it, and focus in an area where we can actually fly and get flight hours and provide some value, and then work our way into larger markets. And I think that what we're seeing is that V3 hits the sweet spot, right? 25 lbs is just, it's difficult to make it work from an economics perspective and overall value chain because you're just not carrying enough like packages, weight, whatever. But that V3, 100 lb useful load really starts to make a meaningful difference because, not surprising, I mean it's half of one of our body weight for example, right? Most humans around 200 lb carry half the body weight on a single flight. Like, yeah, that starts to add some real value, right?

Constant connectivity? No. But it's useful. It's like, you know, typical autonomy Kalman filter approach, right? You just have a bunch of sensor fusion. And so it's like constant connectivity maybe not necessary. Some kind of connectivity over some duration of time, super useful, right? There's drift and, you know, you have error propagation, and so it's like you don't want to end up in the recent events. You don't want to end up in Poland when you're trying to hit Ukraine, you know.

I have infamously said, I think that Ukraine, as much as I hate to say it, I think Ukraine was the best advertisement for the drone industry that it's ever seen. It's tragic what's happening, but I love to see, you know, democracy, capitalism, freedom kind of fighting back on that tyrannical front. So obviously I have my own personal biases on this whole thing. But when you start to see the Ukrainians have asymmetrical power and go and kick the living crap out of the Russians using autonomous systems, and the one weakness that the whole conflict has was, you know, Ukrainians are a smaller population than all of Russia plus, you know, a bunch of North Koreans and anybody else who wants to come and play on their side for money. You know, they need that. And so it's like this urgent need that ultimately drives all technological innovation. So I think what's really interesting is you see it on the surface and you say, okay, well, there's all of this power that's being created and accessed by Ukraine, and now even we are getting access to that expertise through these international partnerships. But what you start to see is the governments generally understand all of a sudden, oh, this is really important for warfighting, this technology. We need to adapt it into the commercial segment because that's how we won World War II, was you have F-150-type production lines turning into Sherman tank production lines, and it's quantity, right? And so right now China is absolutely crushing us in quantity production capability, and as a near-peer adversary, that's a concerning thought, right? So we're starting to see reindustrialization occur, but I think the drone space is having a lot of pressure to get mature really quickly, and the laggard is the FAA. And so it gets into this question of, okay, well, we're a dual-use capability. How do we take that dual use and push it into the commercial market? And we have commercial partners that are ready to go, but we're blocked by FAA regulations. So it's advantageous for us to push into defense, even though it's not as big a market as what the commercial space will be.

I think the future is going to be real weird. I think it's going to get real weird real quick. And I think that we're at a point now where it's increasingly clear to everybody that's in tech or looking at it that we've got AI breaking out of sandboxes, going and attacking, you know, adversaries that aren't adversaries, getting access to data. We're able to build autonomous systems much faster than we ever were before. You know, and it's the lightning in a bottle moment, right, for anybody building anything, right? Because you just have all this access to these data centers and what they can put out. The question starts to be how we control it as a species, right? And it's a species-level thing. It's not a country-by-country thing. And I think this is the problem, is that I think Himelez wrote this book I mentioned before, I think it's Hyperwar, but he's making the argument that I believe is very valid, saying if China adopts the ability to go full kinetic and have the kill chain completely automated, then it's the natural outcome that all other adversaries will do the same because they have to keep up or they're going to fall behind. And so if you have a human in the loop that costs you an extra two seconds to push the button, you've got a problem. And having that go towards nuclear silos is a super scary thought, right? And I don't know how that gets pulled back because of the geopolitical angles. I think it's very concerning. I don't know if that answers the question, but I could talk about it a lot.

Yeah. Yeah. Yeah.

Yeah. Well, I mean, it gets into the physics of it, right? So, I would imagine that the V4, even at 1,300 lb with, you know, 400 lb cargo capacity, I think that would probably still be too light for a person or two in that particular use case, right? But if somebody wanted to, you know, pick up some blood or something, right, and then have it go... or extract some kind of, you know, hard drives or something, right? Then yeah, for sure. It's interesting you bring that up because what we're seeing is that when you make a really capable system, then everybody wants to use it for their own case, right? So, what we've done is take the V3 and the 7 cubic feet of contiguous space, make it into modular payloads where it's just a plug-and-play modularity for various use cases, whether that's ISR or some flavor of it, or cargo, or just a comms relay, right? And then there's of course marsupial operations and these other modalities that could be interesting.

So, this idea that like a robot carries a robot. Yeah. So like you might, for example, V3 could go and, for defense purposes, deploy a bunch of FPV drones and act as the comms node back to operators, or it could just be the comms node to a larger neural net that then feeds down into a system that would then be able to navigate to its target better.

So right now, this is just a... so how to sequence that answer. Right now we have some demand in oil and gas, right? And what we're seeing is there's a lot of understanding around ISR capabilities. So, infrastructure requires inspections, right? And so, these inspections are monitoring pipelines, looking at oil rigs, checking for methane leaks, checking for oil spills. You know, it's all just flying around and looking at stuff, right? And so, that's simple ISR, but in order to get the legs that you need, they use Cessnas or they use like PHI uses helicopters, right? Or Bristow uses helicopters. And there's logistics needs along with these ISR missions, right? Just like taking critical parts to an oil rig, for example. And so there's a lot of this that's going on with these manned aviation systems, but an unmanned system is going to be cheaper, right? The Cessna is more expensive than the V3. And so what we're seeing is that there's a bunch of demand in oil and gas to go and take these modular payloads, shove them into the aircraft, and then do this mission. And it could be a payload for this customer, a payload for that customer, or a generic payload for all of them just doing the job. And we can't fly those missions because of FAA regulations around the unmanned systems not being approved. So it's a Part 135 and we have a 4407, and there's all the FAA jargon that goes along with it. But at the end of the day, it's like, how do you stand up scalable operations? It's not one flight at a time approvals that require, you know, six months of planning to do, right? You need to be able to say, 'I'm going to go and do this just like manned aviation does.' And there's a real concern there, right? Because it's hard to do detect and avoid when you're flying aircraft. There's, you know, big sky, little planes, people move really fast in the air and all that. And so there's a lot of work that needs to be done, but we haven't seen that work come to fruition yet. And so what we're seeing is there's a bunch of oil and gas companies like our partners over at Shell, some folks we're talking to over at Chevron, and they're just saying, 'Hey, look, we want this. We see the value to an extreme degree. How do we make it happen?' And then it's six months or seven months of planning before we can get anything off the ground. Having said that, now we're starting to work with the, I think it's the EIP and HOMA, which is they have these test areas basically to stand up these commercial operations in the US, like in COTUS, and we're looking at the Gulf of Mexico as being the first target for us because there's a ton of oil and gas infrastructure out there.

So on the defense side, we've done almost as much. So the company's raised 24 million to date, and I think we've done about that in revenue so far. All of it's pretty well — I want to say now it's around 70-30, 80-20 kind of split between defense, leaning defense on the heavy, and then commercial.

Okay. But the contracts we've done — phase one, phase two, phase three, IDIQ. Now we're on a STTY contract, and that's our main vehicle. But now we're looking at some operational contracts and trying to dial all that in. And so, yeah, I guess the traction is, you know, to Lucky Palmer's point, it's like, you know, you build it and they will come kind of framework. But you got to build it and you got to have it ready. And then the DoD doesn't like low TRL levels going into real operations, right? They like things that work.

There's a company called Terod Dynamics that has a smaller payload, smaller range. They're, you know, all-electric system. They've got some novel design work. They built this subscale system, and, you know, we've gone out and done some exercises with them. So I know some of their team a little well. And then the real, like, competitors are more not flying, you know, it's like the boats. Yeah. Right. Because at the end of the day, you're trying to solve a problem. People don't really care about the modality of solving the problem. They care about the problem being solved and the timeline and the cost and the pain in the ass and all the things, right? They just want to solve the problem. And so sometimes a boat that's just able to go and do a thing by itself is fine. And then it's lower energy cost and all that. But if you need it faster, then a boat might not be so good, especially if it's rough sea conditions.