Google’s drone company Wing unveils ‘aircraft library’

Sometimes you need a big drone. Sometimes a small one will do. Some deliveries are urgent. Others are far away. For each, a slightly different drone would be the optimal solution. That’s why Google sister company and Alphabet portfolio company Wing has built an ‘aircraft library’ … a vault of drone types it can pull out and get in the air relatively quickly.

In this Techfirst , we chat with Wing CEO Adam Woodworth about the company’s ‘aircraft library,’ about drone development, about what drones will do to the delivery ecosystem, and about the future of Wing.

Questions we address:

  • Why do we need drone delivery?
  • Take out your crystal ball: how big does drone delivery become? What percentage of last-mile deliveries?
  • You have unique aircraft: fixed-wing aircraft that are also drones. Why?
  • You’re releasing something called the aircraft library. what is it?
    • What will it allow you to do?
    • Where will that become useful?
  • What are the critical technologies you’ve had to master to date to get Wing to this point?
  • What remains to be solved?
  • When will you expand beyond your current locations?
  • When will drone delivery be as normal as an Amazon truck rolling the neighborhood?

Subscribe to the audio podcast

TechFirst is available on virtually all major podcasting platforms …

Full transcript: chatting with Wing CEO Adam Woodworth about Wing’s new ‘aircraft library’

(This transcript has been edited for length and clarity.)

John Koetsier: What is the future of drone delivery? Drones are at the epicenter of three or four tech megatrends that I follow … smart matter, automation, AI and decentralization. Alphabet portfolio company and Google sister company, Wing, is at the forefront of drone delivery. Wing has made over 250,000 deliveries in the US, Finland, and Australia. They have a unique take on how to develop the business and where it fits. To chat, we’ve got Wing CEO and former CTO, Adam Woodworth. Welcome, Adam. 

Adam Woodworth: Hi. 

John Koetsier: Hey, it’s super great to have you here. Let’s start big picture. Why do we need drone delivery? Why is it important?

Adam Woodworth: Yeah, and I think that at the core, the mission of Wing has remained the same since its inception. You know, Wing is … Wing, and drone delivery more broadly, is really about sort of removing the elements of friction from the current transportation ecosystem. So when you think about it today, like people order lots of stuff every single day. You know, every single day people are ordering things, getting delivered.

Right now, like you’re moving single pound packages in multi-thousand pound vehicles, and there are just more efficient, lower impact ways to do that. And Wing is fundamentally about eliminating that friction and sort of bringing that vision to reality.

John Koetsier: I gotta say, I’ve got a neighbor and there’s an Amazon truck there every single day, so I know exactly what you’re talking about. Crystal ball, really briefly… how big does drone delivery become? You know, what percentage of last-mile deliveries do you think it could eventually take up? 

Adam Woodworth: Yeah. And I think that one of the central truths to sort of this whole endeavor and this industry largely, is that I… you know, I firmly believe that delivery will always be multimodal. Like you’re always going to have multiple different ways to go and move boxes. If you take the example and extrapolate to a car, like, if I need to move around a couple tons of gravel, I’m going to rent a dump truck. I’m not going to put it in the back of a four-door sedan. In the same way, there are a lot of goods that fit really well into sort of the capabilities that we’ve developed, and there are some goods that don’t. Like we’re not going to be delivering refrigerators anytime soon.

But I think that the vast majority of goods that folks order on demand today can very easily be served by a drone delivery capability.

John Koetsier: So you’ve got a little bit of unique aircraft. You’ve got sort of fixed-wing aircraft that are also drones — maybe I’m using the wrong terminology here — why are you building your drones that way? 

Adam Woodworth: Yeah, so I brought a little visual aid here to help, but, you know, at the core of it, drone delivery poses the challenge of you have to go and bring the goods, right, to the customer. And vertical takeoff and landing aircraft are very much the easiest way to do that, right? So you have the ability to sort of slow down to a complete stop, you know, hover down over the front yard or the backyard and lower the package sort of smoothly and safely to the customer.

Hovering vehicles are typically not very efficient. So, the trade that you make for the ability to hover and maneuver like that is that you’re limited in how far you can fly.

And so Wing’s approach has been to develop what are called these sort of separate lift and thrust vertical takeoff and landing airplanes. So, rather than having tilt rotors or having the whole airplane tilt over, or having like a tilt wing or other different moving parts of the airplane, there’s basically sort of drone components mounted onto more conventional aircraft. So, if you think about it, the airplane will hover up to altitude and then in the same orientation sort of accelerate until it’s cruising like an airplane, and you effectively have like an infinite runway. You know, most airplane wings are sized around like sort of how you can take off and land on a relatively tight space. By eliminating that, by sort of being able to just like already be lifting the airplane, accelerate up until the wing is providing all the lift, you can have an airplane that’s just optimized for cruising.

So, if you look at airplanes, like the wings are relatively small, because they’re designed just for that cruise phase. The one other benefit of this that’s relevant is that you’re effectively carrying around like a redundant airplane with you.

So, you know, I’m flying around like a plane. If anything breaks on the airplane side, I’m carrying a helicopter with me. So I can sort of revert to helicopter flight.

Our system has the ability to sort of spill over control. So, if you know this aileron over here isn’t generating enough roll, these motors can help it. But essentially, if you think about it like robotics, you know, and you have six degrees of freedom and then you have a bunch of different forces you can apply. We’re a super over-actuated system. Like we have a huge number of different actuators that we can use to control the aircraft. 

John Koetsier: That is super smart. I mean, I didn’t even think about that in terms of your wing surface area that a typical plane has to have enough wing surface area to generate enough lift at very low speed. You don’t have that problem, and you’ve got multiple backups and safeties in case your main propeller dies or something like that.

Now, you’re releasing something that you’re calling the Aircraft Library. What is that? 

Adam Woodworth: Yeah, so, in short, we’ve been operating commercially for several years. As you said, we’ve accumulated several hundred thousand commercial delivery cycles. And that has all been on sort of the core of our aircraft fleet called the Hummingbird aircraft.

So, like this airplane, most of the airplanes that you’ve seen from Wing look pretty much like this, you know, maybe some colors have changed and the wings have gotten a little bit bigger, but largely look like this. And we’ve heavily invested in sort of the core architectural elements of this. So, like the avionics unit, the sort of brain that lives on the back of the airplane that sort of tells the airplane where to go and sort of how to control it; the propulsion system elements for hover that hold the airplane up in the air, like and we need a higher level of reliability on; the sensing to tell how fast the airplane’s going; the battery and the battery management system.

All these different elements, we’ve accumulated an immense amount of operational flight time on, we’ve accumulated an immense amount of test time on, and they’re really robust, really capable sort of like building blocks of the system. The airplane is sort of a wrapper around that.

Like, if you think about all those elements as like construction bricks, we’ve happened to build them in this shape of a plane, but we could disassemble that and put them all back together and make it look like a rocket ship or, you know, make it look like a steamboat.

The sort of core element here is that a lot of the elements that are sort of core to the functionality of the system, core to the reliability and performance of the system, those can be adapted to different use cases. So, what we’ve done while these aircrafts have been going and delivering packages to people every day for the last several years, our R & D team has been sort of like spinning up the flywheel of, okay, using some of these common architecture components, what else can we make? Like, what if this airplane was highly optimized around carrying a two-and-a-half-ish pound box?

And you know, if you go to an airport and you look out on the runway, like all the airplanes look different. You got big airplanes, you got little airplanes, you got airplanes with four engines and two engines. And that’s because they’re all designed for very specific functions.

So, in this case, this airplane was designed to carry about two-and-a-half pounds of stuff and carry it about six miles. What if we wanted to carry twice as much stuff? Well, you probably need a different looking airplane. What if you wanted to carry a third as much stuff? Well, then you probably have too much airplane here. And so what we’ve done is started to spin up that development flywheel of building out different aircraft configurations that we bring to sort of like an advanced prototype level. We solve a lot of the sort of hard engineering problems associated with the new configurations, but we don’t have to build a new avionics. We don’t have to build new motor controllers. We don’t have to build like new battery management systems. We can leverage the hundreds and hundreds of thousands of flights and all the testing hours and all of the confidence we’ve built in those systems. And that provides a level of sort of future flexibility that’s really interesting in this space. 

Right now, today, we focus mostly on the sort of stuff that people are ordering through on-demand delivery services. So, you know, a lot of prepared food, a lot of small household items, but by going bigger, you can … what if I want a pair of sneakers? What if I want a cordless drill? Like you’re going to need to build a different capability to do that.

And then, similarly, what if you want to go smaller? What if the genesis for the smaller airplane… and we’re sort of leaning into the Library analogy here, so the small planes called ‘Article One,” you know, it’s like the sort of first volume in the Library, big planes “Article Two,” we’ve got a number of different other Articles… but on the small aircraft side, what if you’re just delivering the bottle of prescription medication? What if you’re at the job site and you open up the toolbox and you are missing one hex key? Like what’s the… these sort of like truly on-demand, necessary items that may not be that big. This airplane is optimized for carrying bigger stuff and you can sort of reoptimize it for carrying smaller stuff.

So those are the threads that we’ve pulled on, and that’s just sort of the items that we’ve shown most recently, represent the sort of payload space there. So starting from the middle, here, and then going sort of bigger to one side and smaller to the other. There’s other dimensions that we’re also exploring, you know, what if you wanted this plane to go five or 10 times as far? What if you were only operating inside of a city where you don’t need to cruise, you’re just hovering all the time? What are the different design decisions you make? And we’re sort of building up the team and the process and the methods to go and rapidly churn through these configurations. 

John Koetsier: It’s fascinating. Kind of reminds me a little bit of software development where you’ve got modules or you call in an API or some libraries of code or something like that, and that’s known good. And you can call that in and use that, you componentize, and then you put together the components as you need and add a little custom flavor… and there you go, you’ve got something new. That makes a ton of sense and it gives you the capability, like you said, you know, maybe you want a rural one that’s going to fly, what, 20, 30, 40, 50 kilometers or something like that, right? But in the city, you don’t need to go that far. How is this going to help you? What does a Wing… I don’t know if I call it an airport or a Wing facility. 

Adam Woodworth: We call them nests. 

John Koetsier: A nest? Okay. Okay. I like it… nests. Excellent. How does a Wing Nest look like, three, four years from now? Do you have four or five different styles of airframe there?

Adam Woodworth: Yeah, I think part of the reason for this exercise is that no one element of the system is really the answer. I think for drone delivery to really work it’s about all of these pieces coming together and sort of seamlessly integrating into the businesses that we serve and the customers that we serve. You know, in the same way that we’ve looked at the aircraft development in terms of what are the sort of core elements that, like we invest a lot of time and effort in sort of getting highly optimized, but a focus on being able to sort of rearrange them into different configurations.

 

We’ve taken the same approach on the infrastructure. So my little visual aid here does not have the charging pins on the bottom, but on the bottom of our aircraft there’s four sort of spring pins and the airplanes land on big giant circuit pads, the circuit board. So, it’s like a three- foot square big printed circuit board that’s got positive and negative lines. And the foot on the bottom of the airplane is designed so that the plane can land in sort of any orientation on that and charge.

And so somebody doesn’t need to go out and put the plane in the right orientation. Somebody doesn’t need to go out and swap batteries on the airplane. You know, somebody doesn’t need to go out and like, sort of tell the airplane, “Okay. It’s time to go take off and do a mission.”

The infrastructure has to be sort of similarly simple and similarly modular to enable these sorts of futures. If you look at some of our more recent deployments in Australia, we put our nests on the roof of a shopping center. And at our most recent deployment in Dallas, the nest is actually in the Walgreens parking lot. And it’s sort of as simple as put down the pads… like all you need is enough space to put down pads, and a wall outlet to plug the pads into, and then the airplanes connect to the internet on their own. And that’s the sort of level of setup that I think the future necessitates. 

And so, to your question of what does the future of a nest look like?

I think that you’ll see lots and lots of these collections of pads sort of like distributed around cities and communities. And very likely there’ll be planes that look just like this one on the pads, and there’ll probably be some plane that looks different from this one on the pad. And there may be sort of a different mix of those, you might have a lot of these planes and just one of the long-range plane, or just one of the big, heavy plane.

But I think that like back to the thesis of delivery is always going to have the need for solutions optimized for the things you’re carrying around. And I would fully expect that, you know, for that big nest that has most of these planes and maybe some other planes, there will probably still be a delivery truck that shows up at the store and picks up the refrigerator, because that’s what the delivery truck is best at. And for the bottle of medicine that needs to get delivered, or for the cordless drill that needs to get delivered, it’ll go on something like this.

John Koetsier: Super fascinating. And I assume in the future there will also be algorithms that optimize where, what planes go to which nests in a city, you know, certain ones are busy overnight for the midnight snack, right? Certain ones need more during the day as people go into their workday and, as you mentioned, you’re on the job site, you need that hex key and you can’t complete your million dollar job without that one hex key, and other ones are needed more at lunchtime and suppertime. Super, super interesting.

What are the critical technologies that you’ve had to master to get Wing to where you are today?

Adam Woodworth: Yeah, I think sort of at the core, the sort of avionics unit, the sort of brain of the system that lives on the back is the, like that’s the core of all of this. So, the flight controls that adapt to sort of changes in performance to the system, the sensing systems that as you go and do the delivery make sure the plane’s in the right spot even if you don’t have GPS, even if it’s really windy, even if it’s really rainy, you know, building up the sort of reliability and the functions of that unit and sort of all of the heuristics associated with it. Like every flight we do, every flight we do today, the system learns more about itself. So, right now, today, if a plane is flying and it comes back and it says like, “Hey, I was just carrying like some more roll control than I usually come back with,” it tells the maintenance staff to go look at this plane and check out the aileron.

And those are the sorts of capabilities that the system gets iteratively better the more and more you operate it, like the more you fly, the more automated it gets. The more automated it gets, the more you fly and like the more the system knows about the health of itself. And I think that as you see these operations expand, it becomes a positive feedback loop, and that’s the part that I think is really, really compelling about these sorts of use cases. So, the avionics unit on the back, and then I think a lot of the development around sort of the hover propulsion technology, you know, particularly some of the rotor design work we did to sort of mitigate noise generated by the rotors. And just generally building a really robust power system so that like you can always count on those pieces of the system. From a hardware perspective, those are the sort of really core elements. 

The sort of other one is like how you physically build these. So, a lot of airplanes are like fully molded out of composites and the cuff material is really high, and our plane, we’ve done a lot of work to really optimize around sort of how you build a lightweight high-performance, high-reliability system with sort of like truly off-the-shelf components, particularly on the sort of airframe side. So, you know, the carbon spars that are the sort of main elements on these booms and through the wings. Like, a lot of the design decisions … a lot of the reasons why this airplane looks the way it does is that intersection between sort of performance, cost and, you know, those two things working together to always sort of meet the safety and performance numbers. So, like really optimizing the airframe design and really doing it in a way where you can build a lot of them all cost effectively. Like that’s something that a lot of thought and time has gone into and that… [inaudible]

John Koetsier: That makes a ton of sense. I mean, if drone delivery gets as big as we think that it could, you will need millions of these. And you will need to design these, you know, and order 10,000 at a time for a new city you’re bringing online. That’s significant. The question that I’ve got is what remains to be solved? Is it largely regulatory problems and, I guess, the operational problems of getting a city or a region up and running — putting the nests in place, getting businesses aligned and saying, “Hey, I wanna try this out,” getting consumers used to it and knowing that it’s going to arrive this way — what remains to be solved?

Adam Woodworth: Yeah. And I think that’s the question I ask myself sort of every day is what are the things that are still to do? You know, one of the elements that Wing has sort of built most of our design and engineering philosophy around, our company philosophy around, is like building two performance-based standards, building two sort of outcomes, right. And I think that if you look at things from the regulatory perspective, the places where the most progress is being made is where that is the focus. So sort of building the rules, building the approaches, building the certification standards, like to be impedance match for the types of systems that you’re assessing. And so, matching the performance of the system to the types of frameworks that fits into.

And so I think that we’re seeing a fair amount of that sort of globally, you know, we’re seeing more countries adopt similar approval frameworks and sort of, it’s just like, you just need more of that more frequently. It’s like those are examples where things are working really well. 

And I think, to the other element of your question, you know, this has largely been an industry of what-ifs for a very long time. So, like people have been talking about what drone delivery might look like. People have been talking about all the different use cases for drones, but it’s only sort of relatively recently that there’s operations at a meaningful scale. There’s a handful of companies doing a meaningful number of flights every day. And, for a space of what-ifs, like, doing is the ultimate cure for what-ifs, right? I can’t even think of all of the conversations I’ve had over the last, you know, long time about what this might look like or what this experience might be like.

And there are now communities that this is very real for them, they see it like every day. You know, in some of the communities we operate in Australia, it’s like, every day people order things. For that region, it looks like scaled out drone delivery. And a lot of those what-ifs have been sort of proven true or false, a lot of new hypotheses have been formed. 

But I think the next real step there is sort of what are the unanswered questions on that front that we will only experience through doing? And I see that, as you know, right now today, the industry is still largely talking about sort of the next neighborhood or the next city. What about the next state? What about the next country? What about, like we’re seeing sort of the expansion of flywheel start to spin up, and I would expect to see each one of those increments becomes bigger. And as you do that expansion, you’ll learn more about how a more distributed system functions at a scale like that. You’ll learn more about how you’re sort of like really deeply integrating with partners around the world that are providing the goods that you’re delivering with these aircraft. I think that that’s sort of… the next unknown is the most exciting piece, which is like, as you get to this next level of scale, what are all the things that you thought that were going to work that do work?

And then what are all of these new opportunities, new learnings that sort of unlock more opportunity in this space? 

John Koetsier: Absolutely. Lots of tough questions remain and, honestly, discussion, at least for me and I know many others. Not everybody’s crazy about lots of drones flying around their neighborhood. But for many, I see what’s going on in… you’re in, I think a couple markets in Australia, you’re in Finland. I think you’re now in two markets in the U.S.

And we look at that and say, “Okay, when’s it coming here? When can I order a hot coffee and, boom, there it is,” you know, two minutes later type of thing. Of course that remains to be seen and there’s complex answers to that. Anything you can tell us about expansion plans?

Adam Woodworth: Yeah. And I think more, most broadly, sort of the discussions have moved from the if to the when, and that’s like a big deal, right? I think that we’re at this inflection point as an industry where it … like we’ve reached sort of critical speed on the flywheel and it will keep moving, like you will eventually get the service. You know, to some of the points I talked about before, as we continue to grow, sort of starting to look at regional expansion versus like city to city, sort of statewide expansion looking at how you go across a much larger region than we operate in today. I think that those will be the sort of next next steps.

And, ultimately, you get to these frameworks where rather than looking at like, okay, this is the one outline on the map, like, how are we going to go open up a service there, you start looking at the characteristics of the map. And it’s like, okay, like, these population densities, this type of terrain, like you start looking at it as this type of airspace you’re operating in. You look at it as the characteristics of the environment, and if one looks like the other, then that expansion happens more organically.

John Koetsier: I think so as well. And one thing I look forward to is kind of what I call the “drone pony express” moment. And I know that this will never really happen as a matter of course, but I kind of wonder about that transhipment, you know, something’s coming from, let’s say, L.A., and it’s going to New York or something like that, and it’s going entirely by drone and it’s going from one to the next, to the next, to the next… that’ll be interesting to see.

I have no idea when that’ll be, if that’ll be 10 years or 20, or maybe never, I’m not sure, there’s some vast distances across. In any case, super interesting. Thanks for sharing about the drone library and some of your plans. Appreciate your time.

Adam Woodworth: Awesome.

TechFirst is about smart matter … drones, AI, robots, and other cutting-edge tech

Made it all the way down here? Wow!

The TechFirst with John Koetsier podcast is about tech that is changing the world, including wearable tech, and innovators who are shaping the future. Guests include former Apple CEO John Scully. The head of Facebook gaming. Amazon’s head of robotics. GitHub’s CTO. Twitter’s chief information security officer, and much more. Scientists inventing smart contact lenses. Startup entrepreneurs. Google executives. Former Microsoft CTO Nathan Myhrvold. And much, much more.

Consider supporting TechFirst by becoming a $SMRT stakeholder, connect to my YouTube channel, and subscribe on your podcast platform of choice: