I remember being in Rwanda early days and going out and meeting with some of the doctors and lab techs that we were serving and asking for them like, you know, how's it going? What do you think? What's your feedback? Here I am kind of up-and-coming learning engineer thinking they're going to say something about the drone, but the main piece of feedback that I received was people get sick 24/7. Why are you guys only open 12 hours a day?
Right. Especially when you're delivering life-saving blood.
Yeah, exactly. And so that was a really key insight for me where it's like, man, we've found product market fit in a market where our product wasn't great yet, but it was solving a real need. Having that beachhead market where there's a real problem being solved, and when your customer is telling you that their main feedback is they want more of your service, that's a good sign.
Welcome Keller and Eric to the show. You guys have been working at Zipline for a long period of time. Keller is the co-founder and Eric, you are in charge of systems engineering and safety. We got lots of things to talk about in this whole world of drones, drone systems, and how you guys started in this hardware space before LLM even started. So, we have lots of questions.
Thanks for having us. But you don't like Zipline being described as a drone company, even though you're probably the largest autonomous drone company in the world right now.
I mean, we've always wanted to be an extremely customer obsessed company. The reality is none of our customers care at all about drones. Our goal was always to build an automated logistic system for Earth and to approximate teleportation. All the customers living on Zipline today don't care about the technology behind the curtain. What they care about is their ability to download an app, see a huge number of different brands and restaurants, click a button, and have it delivered to them five minutes later. We've always tried to focus on the experience rather than the specific technology.
Well, this show is about technology. We're excited about that.
This podcast is about technology. What is the underlying technology behind Zipline? You started
In 2011. Uh, you pivoted in 2014. This is way before anything related to AI, robotics, foundation models, anything related to that.
But you've kind of you were before all of that and you're riding the wave of all of the things that have come after as well.
Yeah, this was when like starting a robotics company was the dumbest thing you could possibly do. You know, talking to an investor at that time, it wasn't easy. It was particularly hard because so many of those conversations, I mean I was 23, 24, Eric joined the company around that time, and we were starting to describe this vision of an autonomous logistics system for Earth that would be 10 times as fast, half the cost, zero emission. One of the biggest problems was investors would say, 'Isn't this illegal in the US?' In fact, I think that's a question you asked me when we started talking about this.
We weren't allowed to fly beyond visual line of sight.
I mean, we weren't allowed to fly at all really. The answer was yes, it's illegal. Most investors would say, 'We don't invest in illegal things.' But for weird reasons, this is what basically took Zipline down this path of well, if it's illegal in the US, then we can launch in other parts of the world where the value of the service would be extremely high. Zipline decided to launch in Rwanda in 2016 delivering blood transfusions directly to hospitals and primary care facilities. This enabled us to have a use case that was so powerful that a government would work very closely with us to make it happen.
And to make it legal, or at least make an exemption to their existing regulatory framework. When we launched in 2016, we had this really cool drone, but it was a total disaster because we learned that the cool vehicle was only 15% of the solution. We had signed a contract to serve 21 hospitals and we served one hospital for the first nine months.
Spent a lot of time in Rwanda and sleep a lot.
And then when you're in the US, we'd get woken up at like midnight because that's when the distribution center was turning on. Everything would be broken, nothing was working. It was totally desperate, constant all-nighters and working through the weekends because we made this big error of thinking that the cool vehicle was the majority of the solution.
What we learned during that first is that the drone is 15% of the complexity of the solution.
The physical drone, the hardware of it is only 15%.
We had to build so many auxiliary software systems, maintenance systems, how do we hold the inventory and do inventory management? How do we integrate with a national civil aviation authority? How do we integrate with a national health care system? How do we do ordering and demand management? We had to build all of these other parts of the overall logistics system. So the aircraft is like 15% of the solution required to build something that feels like magical reliable teleportation 24/7 365 to the hundreds of millions of people who depend on the service.
Speaking of 24/7, I remember being in Rwanda early days and going out and meeting with some of the doctors and lab techs that we were serving and asking for them like, how's it going? What do you think? What's your feedback? Just really being customer obsessed and wanting to optimize the product. And I'm thinking they're going to say something about the drone, but the main piece of feedback was people get sick 24/7. Why are you guys only open 12 hours a day?
Right. And so, especially when you're delivering life-saving blood.
Yeah. Exactly. And so that was our, you know, you got to start somewhere. So we started being open 12 hours a day and trying to expand and grow from there. That was a really key insight: we found product market fit in a market where our product wasn't great yet but it was solving a real need. When your customer is telling you that their main feedback is they want more of your service, that's a good sign.
Yeah, we were 24/7 within the first year. So we went 24/7.
We're now 24/7 365. On Christmas day I usually call all of our distribution centers to thank them and check in. There is no day when these facilities don't depend on us. We went from serving one to 20 to 500 now to 5,000 hospitals and health facilities across the world, across eight countries. It's become the largest commercial autonomous system on Earth.
Eric, can you size that for us? The largest system on Earth. Just crossed 140 million commercial autonomous miles. How many times is that like sun and back?
One of the things that I like is every road in the United States, a lot of roads in the United States, driving on every single road more than 30 times.
Wow, that's good set, it's a lot.
You know seeing the impact that system is now having across all these eight countries, University of Pennsylvania just published a study showing a 51% reduction in maternal mortality thanks to Zipline. So half as many moms losing their lives in childbirth. Across all the different use cases, some of our partners estimate that we're saving between 10 and 12,000 lives a year. And that impact is growing exponentially, especially as a result of this new partnership with the US State Department.
What is that partnership with State Department?
In December, we announced a $550 million partnership with the US State Department to expand the impact of Zipline's life-saving service across a lot of the countries where we're already operating. With US aid being shut down, the US was seeking new ways of engaging in these countries and helping save lives, but they wanted to do it in a way that would accelerate the economies and help the US economically. The new strategy they're calling commercial diplomacy: the idea is that the developing world should be built on top of US AI and robotics technology. When you talk to these countries, they are sick of low-quality aid provided by NGOs for free because these services create dependence and prevent economic growth. What they want is high-paying jobs, entrepreneurship, technology. So the US is going through a big strategic shift to incentivize these countries to adopt that kind of infrastructure, making sure that as they accelerate, they are using US robotics and AI technology. This saves lives, saves money, and secures the US lead in manufacturing and robotics over the decade to come.
I'm curious about, you know, you guys because you now run the largest autonomous system in the world and you launched it 10 years ago. You've been in production for 10 years. You've learned a lot of stuff that your average engineer sitting behind a computer screen has no idea they're going to run into when they try to deploy AI into the real world. What are some of those lessons learned? What popped up over the last 10 years that you never would have guessed you needed to be good at when you first started launching these systems in 2016?
Yeah. We started off delivering life-saving products. Our customers need life-saving products all the time in all weather conditions. One of the weirdest things is actually solar weather. Solar flares on the sun send radiation to the earth, mess with the ionosphere, and cause the RF signals from GPS satellites to be faster or slower than expected. That can lead to degradation and challenges in navigation systems. We didn't think this was something we would have to figure out, but we've gone pretty deep in this space. It's two things: designing our navigation system and GNSS systems to be robust to these conditions, and designing the system to have redundancy beyond GNSS so that if things get really bad, we can still safely operate.
Eric, you're in charge of safety. Tell us about what you've learned about safety today and specifically about the compute failover system that you have.
Yeah, absolutely. There are so many things we've learned over the last decade of operating the system in the real world. One of the things we're proud of is how we developed compute failover. There's a flight computer that flies the aircraft. Lots of sensors come into this computer, it does a lot of math, and sends commands to actuators. You need to assume that any part of the system can have a fault. So what do we do if the flight computer has a challenge? Could be software, connector, various things. We have two flight computers, both think they're flying the aircraft at any given time. They receive all sensor information and send commands to actuators. There's a third orbiter computer that monitors the health of those two and tells everyone who to listen to. What if the orbiter fails?
If the arbiter fails, then the primary computer that was flying just keeps flying. So one's in charge. If the thing monitoring its health fails, then we just keep flying on the thing that was good and continue the mission.
Two heads are better than one.
Yeah. Something we're really proud of. We had one of these events happen a couple weeks ago. After a delivery, we had a hiccup on the main flight computer and switched over to the backup. The aircraft flew itself home, landed, everything was fine. Designing the systems to be robust and reliable through and through is how you get to two and a half million deliveries and 140 million miles flown with no safety incidents.
And a lot of what Zipline's doing is not revolutionary. No one has ever thought about having a secondary flight computer. That's how a Boeing 777 works, but the cost of a flight computer on a Boeing 777 is in the millions of dollars. So Zipline has to take a lot of the best ideas from aerospace safety and figure out how to build that using components from the smartphone supply chain.
Best practices, best practices, and then you got to figure out how to build that using components coming out of the smartphone supply chain. So you can do it for tens of dollars or hundreds of dollars. You can achieve similar levels of safety to traditional aerospace, but you can move 100 times as fast at 1/100th of the cost.
Yeah. So you mentioned that the aircraft is only 15%. Describe the other 85% in like layers and maybe go down deep in some of your systems like the detect and avoid systems.
So how do we test? Why do we test? First of all, we're not a software company, we're a real world AI robotics company. There are electromechanical systems out in the real world. So there's hardware test aspects, software test aspects, and integrated system test aspects. We have a lot of different environments and approaches. On the hardware side, we do component level testing, highly accelerated lifetime testing, where we're taking components like a motor and putting them through hell: making it rain, hot, humid, corrosive, UV, all while exercising the component.
And just to give a context for scale, there are 700 unique components on the aircraft designed from scratch by Zipline. We are designing not just the flight computer from scratch, but the power distribution board, motor controllers, battery, battery management system, the pod which is the smaller robot that makes deliveries. There's an entire Nvidia GPU powered flight computer on the pod. We're building the electronics for the docking station. Even the electric motor is designed from scratch because we need a thrust-to-weight ratio not available in off-the-shelf motors. So 700 unique components, 43 major sub-assemblies, all coming together on the manufacturing line.
Yeah. Right. Going through this type of testing. Other industries, when I talk to people from automotive or aerospace, they often say, 'I ask the supplier what the reliability of the part is.' But we're the supplier. So we're that vertical integration. We have component testing on the ground, system testing where we take full aircraft and put them through vibration tables, wind tunnels, thermal chambers. We want to know how it's going to break, not just if it's good enough. We take it to failure to see where the limits are.
At 49° C, which is very hot. Hot. Down to -25°, which is very cold.
All all the things. Yeah. So, we don't fly anywhere at 49 degrees.
We do. You do? That's We wouldn't test at 49 if we're not wearing. Where are you flying at 49?
I think Phoenix during the summer.
Phoenix during the summer. Yeah. So, and then where's minus 25?
It's a northern parts. Northern parts of the United States. Actually, can I ask you guys how you think about that? Like, I could imagine a different version of the world where you guys are like, 'Hey, look, if it's too hot, we're just not gonna fly.' And if it's too cold, we're just not going to fly. And if it's raining too hard, we're just not going to fly. And like there are trade-offs to be made, and obviously your customers would prefer that you fly at all times, but how do you think about those trade-offs?
The easiest way to think about the trade-off was because of the use cases that Zipline started with. You should be able to count on us with your life and the lives of your loved ones as long as the sun is shining. It's just not; we developed the capability because you had to for the initial use case. Zipline would fly. In fact, for the first couple years we took a lot of risk. We would basically fly if it's a life-saving delivery. We had a civil aviation authority that was a great partner. We took a lot of risk, we learned a lot, and almost always it worked out. The worst thing that could happen was we had a parachute.
How often? It happened very often in the first few years, very rare today. To put it into perspective, our original goal was to be 10 times safer than cars. Alfred pushed in our last board meeting: 'That's a BS goal. We need to be two times safer than Waymo.' Eric reset the goal. Now Zipline's target is to be two times safer than Waymo. He's like, 'Cars, that's archaic technology.'
Waymo, I think, is about 10x, right? They're about 10x cars.
So our goal is to be 2x safer than Waymo. Yeah. But it's not the right comparison. You're flying. You have to be safe in the air.
I think it depends. We're substituting something that's typically going in cars. So it's debatable. But suffice it to say, we now have 140 million commercial autonomous miles and zero safety incidents.
If you were to drive 140 million miles, you would have 600 accidents, 100 injuries, and between two and six fatalities. This is why we pride ourselves on picking the right use cases. It's life-saving. We're going to be as safe as humanly possible. The outcome of all that testing is we have individual aircraft that have flown more than a million commercial autonomous miles. People think it's fragile, but these systems are already way more rugged and durable and robust than people necessarily think.
Can I ask you about the precision? The drone's 100 feet up and it drops the package to a circle with an 18-inch radius. How do you achieve such precision even when it's windy or raining?
First of all, the aircraft's about 100 meters up.
