In 2013, Jeff Bezos introduced the world to a concept that promised to revolutionize delivery.

“[00:00:20] Charlie Rose: Bezos kept telling us that he had a big surprise, something he

wanted to unveil for the first time [...] Bezos: but there’s no reason they can’t be used

as delivery vehicles.”

Within a matter of years autonomous drones would engulf cities, sweeping across skies,

delivering packages to front yards just thirty minutes after their order.

The announcement floored Charlie Rose and America alike—grabbing headlines for weeks

and setting off fiery debates over just how disruptive the disruptive technology would

be.

Drones would offer urban and suburban consumers a clean, quick, convenient delivery option

for food, medicine, or whatever else five-pounds-or-less without burning fossil fuels, without getting

stuck in traffic, without making them wait.

The idea felt far-fetched, it was exciting, it was the future arriving in front of our

very eyes, and it signaled that the race to take delivery drones to market was on. 

By the time Amazon landed its first package, Zipline was already delivering medical supplies

in Rwanda, SkyDrop had flown a 7-Eleven slurpee and a Dominos pizza straight to consumers,

and Google’s Project Wing had air dropped burritos to hungry college kids.

A wave of startups sent their maiden drone deliveries skyward to much media fanfare while

major parcel couriers—DHL, UPS, FedEx—substantiated the hype by partnering with the budding tech

companies set to help solve their last mile problems.

Inventors, investors, eccentric billionaires, and the world’s biggest companies were all

pulling the same rope.

Anything, anywhere, anytime: the dark days of Doordash and two-day delivery were over;

the drone delivery era was coming… or so it seemed. 

It’s now 2022 and save for the smallest fraction of a percent of people, it’s not

automated drones dropping off your small packages and food orders.

Pizzas aren’t falling from the sky, they’re showing up in the hands of a highschooler.

Burritos aren’t delicately dropped on your lawn from above, they’re left on your doorstep

by a hustling gig worker.

Your Amazon order won’t show up in thirty minutes, it probably won’t even show up

same-day.

The world 60 Minutes introduced in 2013, the world that felt closer and closer to reality

with every inaugural delivery, just isn’t here. 

Fundamentally, the fast delivery niche still exists.

The last mile still accounts for around 40% of parcel shipping costs, roads are still

increasingly clogged with traffic, green shipping alternatives are still desperately needed,

and consumers still want products as cheap and as fast as possible.

Outside of a few specific locations, drone delivery has yet to take off—and in those

few specific locations, it's hardly more than a proof of concept.

Certainly, delays are understood—expected even—when it comes to the acceptance of

a disruptive technology.

Delayed acceptance though, is at very most only part of the story. 

In 2021, Amazon fired staff and closed its Prime Air offices in the UK.

From the former center of Amazon’s drone project emerged stories of mismanagement and

disarray: employees drank beers at their desks, managers were given no direction, executives

ignored the stalling division aside from the occasional pizza party.

While the company responded to these reports with a statement affirming its continued investment

in drone delivery, Amazon hasn’t released any promotional material for the project since

2016 and Prime Air’s website doesn’t seem to have been updated in years.

The most generous possible interpretation is that Amazon’s project is definitively

on the backburner.

Others aren’t even there.

While Amazon remains quiet on their future intentions, DHL announced in summer 2021 that

it was officially abandoning its parcelcopter project nearly eight years after it’s maiden

flight. 

So, two of the most important drone delivery companies put their programs on ice, few companies

are getting the investment they used to, and no company has yet realized the imminent future

of widespread operations laid out a decade ago.

So, what went so wrong with drone delivery? 

Well, this is Phoenix—a sprawling desert metropolis home to 5 million people.

On first glance, Phoenix seems the perfect candidate for a drone delivery service: its

year-round sunny, dry, still climate would make for easy, reliable flight conditions;

its autonomous innovation friendly city and state governments would welcome them with

open arms; and its sprawling, low-density neighborhoods would make for countless hungry

and impatient residents lacking walkable dining and shopping options.

Surely, this is the low-hanging fruit.

Surely, a drone delivery company could come in, connect any house with any product within

minutes, and demand would immediately outstrip supply, right? 

Well, perhaps not.

Connecting any house with a drone delivery provider doesn’t quite work because in the

center of the city, right here, is Phoenix Sky Harbor International Airport.

In order to assure the safety of arriving and departing aircraft at the busy hub, the

FAA restricts the use of drones within this area.

So, a Phoenix drone delivery service probably just couldn’t operate here… and here,

in the restricted airspace around Luke Air Force Base… and here, around Phoenix Goodyear

Airport… and here, and here, and here, here, and here.

It’s not entirely impossible to operate drones within restricted airspace, but, from

a legal perspective, it ranges from somewhat to extremely difficult—enough that it probably

wouldn’t be worth pursuing for a drone delivery company, at least at the start. 

The rest of Phoenix, though, is fair game… to an extent.

You see, drones need somewhere to deliver to, and it’s got to be safe.

When the concept was first introduced, the vision typically presented was of a drone

flying down, landing on one’s lawn, releasing its delivery, then taking off and flying away.

That didn’t work…

at least not in reality.

Drone delivery is a novel technology and, like any novel technology, the public views

it with an air of distrust—the worst thing the industry could do is prove that distrust

warranted with a series of high-profile accidents at launch.

The first instance of a delivery drone injuring a customer will inevitably ignite a media

firestorm, which could lead to a legislative clampdown, so manufacturers naturally must

strive for perfection.

Perfection is tough to scale, though.

Delivery drones must act autonomously to be cost competitive, and autonomous operations

require computer vision and artificial intelligence able to reliably identify a clear landing

zone.

Determining whether someone is behind or infront of a window, noticing when a dog is running

towards the drone, knowing what’s a pool and what’s dry ground—these are all challenging

for a computer to tackle on its own, and so attaining perfection proves rather difficult. 

Therefore, whereas the logistics field generally considers the last mile of delivery the most

difficult—once the economies of scale are gone—drone delivery is a last mile solution

with its own last foot problem.

It’s fairly straightforward to get a drone to a couple dozen feet above the ground—getting

a package safely to the ground has proven more challenging. 

Some solutions have emerged: Zipline, focusing on longer-distance delivery to a set number

of facilities with dedicated delivery zones, drops its payload in a packaging with an attached

parachute that carries it to the ground.

Matternet also uses dedicated zones for delivering to commercial facilities, while they’ve

developed a system of delivery stations for use by urban consumers.

Uber Eats, meanwhile, implemented a scrappy yet inefficient system where delivery drones

would land on the top of delivery drivers’ cars, then those delivery drivers would walk

the food to the customer’s door.

Most solutions for the last foot problem, however, have gravitated towards one method.

Wing, Skydrop, Flytrex, Wingcopter, and others have developed systems where their drones

hover above the destination at a safe height and lower their payloads to the ground using

cords—far less risky than landing a heavy drone propelled by fast-moving rotors.

What all these solutions have in common is that they require a roomy, controlled, obstruction-free

area to make their final deliveries.

However, in the places where people actually live, that’s hardly a given.

Yards are the best delivery zones that are widespread, but not everyone has a yard.

While it’s a safe bet for single-family homes in an area like Phoenix, it can be hit

or miss whether multi-family homes and apartment buildings have a big enough yard and, even

when they do, their communal nature means that the customer couldn’t necessarily guarantee

that the landing area would be free from obstructions as would be the case with their own, private

yard.

So, at least for an early drone delivery service, it probably wouldn't work in restricted airspace

and probably not for anything beyond single-family homes either. 

These and other legal, technological, and practical constraints combined mean that the

scope of what works in terms of drone delivery is narrow.

It’s pretty easy to start crossing off cities—Boston’s winter is too harsh, New York’s density

inhibits yards, DC’s airspace is too restricted, Pittsburgh's landscape is too hilly, this

could go on and on.

Even within the cities that might work, there are only so many areas that might work.

While it varies by company, most delivery drones tend to be able to fly to deliveries

as far as about six miles away.

So, assuming early operations would base out of a single location to capture economies

of scale, meaning their drones would have to return to said location to charge after

each delivery, that means a viable first delivery zone in Phoenix—optimizing for a large area,

free of airspace restrictions, centered on wealthier neighborhoods—would be this.

310,987 people live in this zone—a small chunk of the metro area’s 5 million.

However, in Phoenix, only 63.2% of housing units are single-family, which are likely

to have the private yard necessary for a delivery, and only 92% are occupied meaning, in this

prime zone, at least extrapolating using city-wide data, which is the most precise available,

there are only 180,820 possible users of a drone delivery service.

This is, clearly, an imprecise methodology, but it’s indicative of how the prospect

of drone delivery—the prospect of anything, anywhere, anytime—is getting diminished,

and diminished, and diminished down into a niche service for a lucky few.

A small system linking a strip mall to the neighborhood behind it, a fixed route flying

COVID vaccines from a distribution center to vaccination sites, six shops delivering

to a small part of a small town in Virginia—drone delivery has hardly moved beyond proof of

concept, and it’s not even clear that they’ve proved the concept. 

 In 2016, when asked about same-day delivery, 70% of respondents said they were content

with the cheapest option while just 23% of respondents said they’d pay more for same-day.

For drones to prove commercially viable they’d need to decisively corner that quarter of

more willing consumers, and to become ubiquitous, they’d likely need to operate at no extra

charge from ground delivery at all.

Most people, it turns out, are simply okay with waiting a day or two for their packages,

while all want them delivered as quickly and as cheaply as possible. 

When the drone delivery hype hit fever pitch, one bit of nuance went overlooked.

Consumers simply don’t care about how a package gets from b to c, so long as it’s

quick, cost-efficient, and reliable—they’d opt for a new technology once for the novelty,

but by the 100th time that wears off.

Eventually rationality will return.

In fact, when surveyed in 2020, consumers perceived drones to potentially threaten those

most important factors for delivery—they said they were uncertain about drones’ reliability,

cost, and were worried about the job loss they could incite.

Meanwhile, competitors have figured out a number of low-tech solutions that fulfill

these consumer desires: look no further than food delivery apps. 

Since 2017, the very moment when drone delivery hype hit fever pitch, the food delivery industry

has tripled in size, ballooning into a $150 billion sector globally.

In this, speed matters and consumers expect to pay for the delivery cost—facts that

seemingly pave a lane for drone delivery.

But between Uber Eats, Grubhub, and Doordash, the power players are already established

and the competition is already fierce.

These comparably low-tech companies don’t even tell the consumer whether to expect their

burger to come by car, moped, bike, or foot; they just prioritize getting food to doors

quickly, pleasing the consumer regardless of method and undercutting drone delivery

in the process.By-and-large, food delivery apps closely match the upside of drones within

urban and suburban areas without the hassle of complying with FAA guidelines and figuring

out the last foot problem.

Adding to the competitive problems facing drone startups, these companies and others

have since expanded into grocery, medicine, and goods deliveries.

Put simply, from the consumer perspective, the problem drone delivery was designed to

address has already been solved without building out a massively complicated aerial delivery

network. 

The current low-tech, gig economy model isn’t perfect, though.

For consumers and restaurants alike, the usage fees are expensive; for those delivering,

the pay is minimal; and for the big players, profit has proved elusive.

One partial solution is automation.

Here still, though, drones are likely to lose out.

Ground-based autonomous and semi-autonomous robotics have begun popping up in test markets

and partnering with the likes of Uber Eats and Grubhub to expand their reach.

While a recent partial ban on sidewalk-wandering robots in San Francisco points to the hurdles

the technology faces, these hurdles just won’t be as numerous as those facing drones.

Automation and technological advances may well help smooth out food and last-mile delivery.

In the near future, your prescription, your lunch order, or your afternoon coffee might

be showing up at your front door courtesy of an autonomous vehicle—you’ll just need

to reach down and grab it from a robot instead of unclipping it from a drone above.  

Now, many probably now look at drone delivery in retrospect and find it unsurprising that

the bombastic claims of the 2010s failed to pan out, but far fewer would have expressed

a dissenting opinion just five years ago.

That’s because this is a rather classic story: that of a hype cycle.

A new idea comes around, a few early players start development, then something—a launch,

a demo, or even just a domino effect—sets off a media firestorm painting a rosy picture

of a future revolutionized by this new technology.

This story is so archetypal in tech that there’s even a theoretical framework defining the

process: Gartner’s Hype Cycle.

According to it, after that media firestorm, that peak of inflated expectations, results

slow and sentiment starts shifting downward.

Investors complain and the public’s memory fades until the media begins coverage of the

purported failure.

The public grows disappointed, but then grows silent, and in the silence, first generations

are adapted into seconds, failures inform potential success, and slowly something meaningful—albeit

minor compared to the original vision—starts to work. 

We are here.

While what’s happening may be drowned out in the media by what’s not, there are applications

that are starting to work.

Three years ago, Zipline was a small Silicon Valley startup operating a few dozen drone

delivery flights per day in one region of Rwanda.

They relied on the principle that many medical products are crucially important when used,

but not used regularly, and often have short shelf lives, making them tough to economically

and efficiently stock at smaller clinic operations.

In less developed regions, poor road infrastructure makes many remote clinics many hours or days

away from a distribution center, despite relative proximity as the crow flies.

Therefore, Zipline’s drones acted as a quick, low-cost distribution system for necessary

medical products to remote areas dotting Rwanda’s rolling hills.

Far more places than Rwanda’s Southern Province fit this description: nowadays, Zipline operates

similar systems in the country’s Eastern Province, four regions in Ghana, the US, and

a number of other locations are in active development.

Excitingly, the news of Zipline’s impending expansion to the Ivory Coast hardly made news:

it wasn’t written about in Wired, TechCrunch didn’t publish an article about it—just

a simple press release and some industry and regional coverage.

This is progress.

This shows that Zipline’s deployments aren’t proofs of concepts, aren’t publicity stunts,

they’re actual, real, commercial implementations.

Crucially, Zipline didn’t find a use case that drones could fulfill—they found a use-case

that only drones could fulfill.

They found the healthcare use-case—they found the low-hanging fruit, and other companies

are noticing.

Matternet and Wingcopter are now placing heavy emphasis on their medical potential as well. 

As the early use-case matures, cost will come down, acceptance will rise, and innovators

will find more uses that only the novel technology can fulfill.

Once one becomes clear, more must be possible.

Eventually, everything will creep closer to that idealistic vision first presented at

peak of the hype and then, just slightly delayed behind expectations, the new technology will

finally have actually changed the world. 

Unlike delivery drones, what I hope does live up to expectations is our brand-new, feature-length

documentary, which came out today after almost a year of work.

It’s called The Colorado Problem: A River in the Red and rather than trying to describe

it, you’re about to become one of the first people to see its trailer, but before I need

to very quickly remind you that if you’re not already a subscriber to CuriosityStream

or Nebula, the two sites where you can watch this doc, you subscribe to them both for less

than $15 a year at CuriosityStream.com/Wendover—lower than the monthly fee for that big, red streaming

site that just jacked up its price.

It’s these exact subscriptions that enable us to do these big, ambitious documentary

projects, so a massive thank you to those of you already subscribed.

But now

to the point: the trailer.

As a reminder, to watch this doc, click the button on-screen or head to CuriosityStream.com/Wendover

and you’ll get access to both CuriosityStream and Nebula for less than $15 a year.