We're actively telling people that it's going to kill them and kill their families, which is like, you know, interesting marketing strategy from some of the AI labs, I think. At the end of the day, like, I think we need to show that you can use advanced technology, you can use AI and robotics to save lives. You can focus on some of the most important problems that humanity has, and you can do it in a way that fundamentally transform everybody's lives. Save time, save money, save lives.
Hello everyone and welcome back to MTS. Today, I am joined by Keller Rinaudo Clifton, who is the co-founder and CEO of Zipline, the autonomous delivery company that began by delivering blood and medical supplies in Rwanda and now operates healthcare, food, and retail delivery networks in the United States. And earlier today, Zipline and Uber announced plans to bring autonomous drone deliveries to Uber Eats with an eventual goal of reaching 1 million deliveries every day. So, Keller, so excited to have you on MTS.
Well, great place to start would just be with today's announcement. Can you share a little bit about what was announced and why now?
We announced this new strategic partnership with Uber to get to a million deliveries a day. You know, important to realize that that's actually still a very small percentage of the total instant deliveries happening in the US today, let alone the fact that Dara and I both are confident that delivering in a way that's 10 times faster and less expensive is ultimately going to lead to a big expansion of this overall market. In fact, if you were to just look at the customer ordering behavior that Zipline is observing right now across the cities where we operate, if you were to extend that to the rest of the US, there would be 10 times as many instant deliveries. So, we expect to go from about 5 billion instant deliveries a year to about 50 billion instant deliveries a year in the US alone. So, there's an opportunity to grow the market by 10x.
Um you know, there's not only that, but I think when you look at instant delivery, you know, the reality is we're kind of using technology that's a years old, right? Which is like a human driving a car to solve a problem or a customer need that we've only really well understood over the last decade. And so we think that ultimately if you reason from first principles, it's obvious why if you're delivering something that weighs 5 lb or 10 lb, you probably want to do that with a vehicle that is electric and autonomous and a vehicle that weighs 50 lb rather than 4,000 lb.
Definitely. So, what does that distribution specifically unlock for Zipline? What's something that it unlocks that maybe you couldn't have done as quickly on your own?
Well, I mean a couple big things, right? Zipline, Uber obviously has tens of millions of loyal customers and this means that Zipline service, this kind of teleportation-like service, is now going to be rolled out to all of those customers so that they can experience it for themselves. Also, Uber works with I think 1.4 million merchants and small businesses, something on that order. And so, not only enabling for each of those businesses because Zipline can deliver out to, you know, the system, the technology that we build is designed to deliver out to a 10-mile range, which is way more than is economic to make deliveries using cars.
And so for this reason, from any given restaurant or merchant, you can reach about 10 times as many customers via instant delivery as you could before. So, it creates a lot of value for those merchants and also, you know, all of those merchants are now going to be available on the Zipline app as well. So, it's just huge, huge. I hopefully this has a big impact in helping Uber grow Uber Eats, which is actually the biggest portion of their business at this point by bookings and revenue. But also, this, you know, adding all of this selection and all these amazing brands are going to help Zipline grow our own app and provide amazing experiences for restaurants.
Yeah, I'm also curious because you guys have been working on autonomous delivery for years. And is there a specific technical problem that had to be solved before you could get all of this infrastructure out in, you know, the main public's adoption?
I think the hardest thing, I think it's probably a general lesson for robotics in general, but it's very easy early on when you're kind of imagining trying to build a new technology, it's very easy to get focused on like the core thing. And in this case it would have been like the aircraft. Like how do we build a really high performance aircraft? And the hardest lessons we learned in the first 5 years of operating this system at scale outside the US were that the aircraft is only 15% of the complexity of the overall solution.
And this is something by the way, you know, when I was spending time with Dara, I realized, okay, yeah, like for obvious reasons they just completely get this fundamental idea, which is actually the vast majority of the challenge is in integration in, you know, the integrating auxiliary software and metrics and operations, maintenance, manufacturing, supply chain, charging sites, all this stuff that we're having to do across many different metros in the US. There's just a huge amount of complexity that you don't necessarily think of when you just think of, oh, drone delivery. Okay, cool. I need a cool drone and then it's going to work. So, there's so much that goes into these integrating operational layers, and I think that, you know, it just took a long time.
I mean, that's the reality. I think, you know, people hear about the hype cycle curve, right? Where you have like the early excitement and then the full on hype and then you have like the crash and then the trough of disillusionment and then usually, you know, 8 to 10 years of hard work to actually build something of value. I think you saw this cycle play out very accurately for autonomous cars. And I think you also saw this cycle play out for drone delivery. And you know, the reality is well after the hype cycle came and went, you then had 10 years of like building all of these auxiliary layers and integrating software systems that were necessary to actually make this thing work in the real world in a way that's super reliable. Zipline now has 140 million commercial autonomous miles and zero safety incidents. And it was only, you know, awarded regulatory permission to fly in the US after 10 years of operating these systems at scale outside the US.
Before we dive in, I want to shout out our sponsor Lovable. You know that app you've been meaning to build, the internal tool, side project, or product you'd otherwise lose a weekend to? With Lovable, you get software you can ship today. That includes an editable codebase you can inspect and change, plus two-way GitHub sync. Lovable also handles the backend and infrastructure, including managed Postgres, auth, storage, hosting, payments, and more, so you don't have to wire it all up yourself. And through Lovable's MCP server, you can create and deploy your project directly from the agents you already use. Turn ideas into software people love with Lovable. lovable.dev.
Now, onto the episode. How were you able to get it so complacent? So, how were you able to make sure that you don't have none of those safety incidents? How did you ensure the reliability of these autonomous drones?
I think there are, you know, nothing is easy when it comes to hardware reliability. You know, Zipline has a huge hardware test and reliability team. We have a huge safety team. Then we also do a huge amount of work in terms of testing new software releases at the company. As part of this building where I'm standing right now, I'm in Zipline's autonomous aircraft factory. This building is capable of producing 20 to 35,000 of these aircraft a year, which is about enough to get us to like one, one and a half million deliveries a day.
So, you know, one big takeaway from this partnership with Uber is like we got to massively expand the factory. But, you know, in this building is our entire hardware testing lab, our vibe chamber, our wind tunnels. You know, there are 700 unique components on the aircraft, almost all of which are designed from scratch by Zipline. 43 major sub-assemblies on the aircraft, and that's just the aircraft, not to mention ground infrastructure, the towers, the drop boxes, which are sort of the integrating mailboxes that we set up with all these different businesses. The dock, there's a lot of different hardware products that go into this kind of ecosystem to enable autonomous logistics to work. All of those are manufactured here in the factory, and all of them are tested to failure through, you know, tens of thousands, hundreds of thousands of, you know, equivalent hours of testing to try to find failures here before we ship hardware into the real world.
What does it actually look like when it comes to regulation for autonomous drones? What permission do you actually have to get to operate it at scale?
