If Jeff Bezos has his way, U.S. skies will be filled with thousands of small, unmanned drones ferrying packages from warehouses to customers' homes in a half-hour or less.
The Amazon.com Inc. CEO's vision seems at a glance to be drawn from science fiction, and it faces many regulatory and logistical hurdles before it takes wing. The drones' viability as a delivery device also will depend on their cost and energy efficiency compared to traditional shipping.
Powering the drones would be easy. Even on Amazon's busiest delivery days, they would likely only increase U.S. electricity consumption by a hundredth of a percent, according to estimates prepared for Greenwire based on Amazon's delivery volume and a drone's energy needs. An executive for a drone company estimated that the energy used for flights contemplated by Bezos would cost less than 2 cents per delivery.
But it remains to be seen how that would compare to trucks full of packages snaking through city streets. Precise comparisons would require city-specific scenarios and detailed data that are not immediately available. But some rough estimates suggest drone deliveries could cost twice as much per package as similar deliveries via trucks.
"That's not a surprise at all, given that 1) the drones need to fly, so it needs to overcome both gravity and air resistance [and] 2) there is always energy savings when you deliver in bulk/groups of packages," said Shu Sun, an energy and technology analyst with Bloomberg New Energy Finance, in an email.
To be sure, Bezos' announcement Sunday on CBS's "60 Minutes" -- coinciding with the kickoff of the busy holiday shopping season -- could turn out to be little more than a well-timed publicity gambit. But in pursuing drone-based delivery, he is entering territory where other companies like Matternet already are trying to clear a path.
The vision of the future shared by Bezos and his fellow innovators is one featuring less traffic on city streets from large delivery trucks, meaning reduced congestion and less greenhouse gas emissions. Drone proponents also believe the technology would eventually use significantly less energy than existing shipping methods, although it's a comparison that remains difficult to make at the moment.
"We're working to figure this question out," said Andreas Raptopoulos, CEO of Matternet, a startup that aims to build drone delivery networks.
It's difficult, Raptopoulos said, to draw solid conclusions at this point about how much energy would be required for a network of drones and how those figures compare to traditional delivery means. But "I want to say [it would be] orders of magnitude less" expensive, he said.
The "last mile" of a product's journey is the most ripe for improvement, he said, because getting something from a distribution center outside a major city into someone's home is "the most inefficient part of this trip" and can be responsible for as much as half the overall transportation costs.
At the same time, there are indications that drone deliveries would use more energy on a per-package basis compared to traditional deliveries via companies like FedEx and UPS.
That's where ancillary benefits such as reduced road congestion become key selling points for drone proponents, but even the technology's biggest boosters acknowledge that much is unknown about its prospects. Furthermore, the environmental footprint of the comparable technologies would depend heavily on what fuel is used. Regardless of how much energy it takes, an electric drone whose battery is charged via solar panels or wind turbines will produce less greenhouse gas emissions than a truck engine running on diesel fuel.
The major delivery companies all are pursuing their own environmental measures, experimenting with lower- or no-emissions vehicles powered by biofuels, natural gas or electricity.
Calculating energy use
While the comparison between drones and trucks will surely provide ample fodder for researchers in the coming years, some back-of-the-envelope calculations provide a sense of the energy profile of drone delivery compared to more traditional means.
In response to a request from Greenwire, Shu estimated that a network of Amazon drones in the United States would use about 1,300 megawatt-hours of energy on its busiest day. That's based on Amazon data showing it shipped 26.5 million packages per day at its peak last year, about half of which were in the United States, and an assumption that about half of those packages would be eligible for drone delivery. Shu assumed Amazon would use drones similar to those now on the market, which would use a 200-watt-hour battery that would be recharged for each delivery.
In the grand scheme, that amount of energy is barely a drop in the bucket. U.S. electricity consumption was about 11 million megawatt-hours per day in 2012, so even peak demand from Amazon would only increase that by about 0.01 percent, according to Shu. Still, he says, the drones could have a significant localized effect, given that they would likely be concentrated around big cities.
If Amazon decides to pursue a goal of powering the drones with renewable energy, their impact on demand growth could be slightly more substantial, although still quite small. For example, last year, wind produced an average of 384,000 MWh per day, while solar generated 35,000 MWh per day, according to the "2012 Renewable Energy Data Book" prepared by the National Renewable Energy Laboratory.
Amazon didn't respond to a request for comment.
Comparing those projections to energy use by the largest delivery carriers becomes even more complex. A UPS spokesman pointed to the company's annual sustainability report, which showed that each ground delivery consumed about 0.113 gallon of fuel, including gasoline, diesel and compressed natural gas. Assuming that measurement was based on the energy equivalent of a gallon of gasoline, that means UPS required between 13,000 and 14,000 British thermal units of energy -- equivalent to about 4 kilowatt-hours per package it delivered.
That's substantially higher than the 0.2 kWh average per trip that would be required for a drone delivery, according to Shu's calculations, although the UPS number likely accounts for the entire journey of a package, rather than the "last mile" that would be covered by a drone.
A FedEx spokesperson didn't respond to a request for comment. However, data from its FedEx Ground service, which specializes in small package deliveries just in North America, may provide a useful comparison. The service delivered about 1.46 billion packages last year and spent about $17 million on fuel, according to data from the company's annual reports. Calculating energy used per trip, based on an assumption that the spending was mostly on diesel fuel, results in an estimated 0.1 kWh per package, about half the estimated energy used for a drone trip.
By no means are these rough estimates definitive, and the Amazon drone program -- if it ever comes to pass -- is at least several years away.
If nothing else, the company's entry into the space brings new expertise to bear on a range of problems that will be key to a range of industries, said Jim Greenberger, executive director of the National Alliance for Advanced Technology Batteries. Just like with electric vehicles, if the drone delivery experiment is to succeed, it will need better batteries that can store more energy at less cost.
"It is the threshold tech challenge of the 21st century," Greenberger said. "I can't imagine Jeff Bezos isn't thinking through those problems."