The Lonely Future of Buying Stuff

It’s 2036. We’ve automated ourselves out of shopping and shipping. Here’s how it happened.

By Justin Bachman
September 13,

Illustrations by Keith Rankin

It’s a chilly October evening in Englewood, Colorado. The year is 2036. Caitlin, 26, gets home from work to find a box at her door containing a pair of sneakers she bought online three days ago. They were an inexpensive impulse purchase—only $29 with free shipping—great for running on the weekends. The shoes, one of the 3,812 pairs for sale in the women’s department of, were made in a Chinese factory back in July.

The sneakers, fuschia with bright yellow details, are unremarkable, but their journey to this suburban door is more interesting.

Over the past couple of decades, humans have been entirely removed from the logistics business, which is now coordinated by machines and software. Making shoes long ago became a robot’s craft, at least on the low-end side of the business. Creating, selling, transporting and buying consumer goods such as casual footwear now requires just one significant human—the consumer—plus an individual here and there to oversee assembly and repair robots. Many of the basics we buy are now constructed, bought, and shipped with no one besides the customer ever laying eyes on them.

The sneakers' trip from a southern Chinese footwear factory begins on July 5, 2036, and ends on Caitlin’s doorstep 18 weeks later, on Oct. 26. Caitlin is the first human who has ever seen or touched this particular pair of shoes. She’s pleased that they’re as comfortable as she had anticipated.

BACK TO 2017

This future journey is an attempt to show one how different technologies may reshape the retail world. In a not-too-distant future, current tools and ideas may have reached a point where fulfilling our online consumption no longer requires many—or any—humans at all.

Of course, there is no guarantee this particular scenario will come to pass. Some unknown hurdle could cripple regulators' appetite for self-driving vehicles. Low-margin shoe assembly might flee Asia if it's cheaper to "print" footwear in local markets using 3D additive manufacturing tools. But the shipping and logistics worlds are keenly invested in the potential of these technologies. Some version of this reality is close.

7/5/2036 0545

Begin cargo shipping.
Destination: California

From the factory in Dongguan, China, north of Hong Kong, robots load hundreds of the shoes into a metal container, the kind that are stacked by the thousands onto container ships to cross the Pacific Ocean. These 40-foot intermodal containers are still a mainstay of international trade. Ours, a dull red, is stacked onto a ship for a 26-day trek across the Pacific to Long Beach, California, via Vancouver. It sits 30 feet above the ship’s deck, surrounded by 17,442 nearly identical neighbors that are packed with such consumer goods as toys, dining chairs and bookshelf speakers. There’s not a soul aboard. Once the ship leaves the bustle of Shenzhen’s port complex, it’s a near-silent drone, chugging northeast. A tag on each container pings a satellite if it shifts more than 1.5 inches during the voyage. A computer in a remote marine piloting office in South Korea is monitoring the position of it and about 4.8 million other containers worldwide, just like the one with our shoes. Jolts greater than a quarter-inch are exceedingly rare. If the ship’s cameras detect anything outside normal parameters, the computer in South Korea will dispatch a small, aerial drone stored on the ship to take a closer look.

BACK TO 2017

Ninety percent of all international shipping occurs on the seas, employing some of the largest vessels ever built. Compared to other forms of transport, container shipping is inexpensive, owing to an ecosystem of U.S. retailers, Asian suppliers and steel boxes that have squeezed costs enormously and turned virtually any item into one that can be sent cheaply almost anywhere on the planet.

Ships are moving rapidly to a digital future, where many vessels will become operated remotely and others will become fully autonomous, says Oskar Levander, vice president of innovation at Rolls-Royce Marine Power Operations Ltd., which is researching various technologies to further this transition. Rolls-Royce and others that advocate this type of shipping argue that unmanned vessels will offer lower costs, will become far more optimized and provide greater safety in navigation.

Ships that are “local,” such as ferries and tugs, will be among the first to become autonomous, Rolls-Royce predicts, followed by coastal vessels and then by those that ply the high seas. Captains will relocate from ship to shore, using an array of sophisticated tools: radars, cameras and sensors to drive vessels, much as U.S. military pilots, based in America, fly large aerial drones around the world. In other cases, a human navigator will control the ship only at specific periods. “We’re going to have captains in the shore centers driving this ship when they’re approaching port, or when something special is going on,” Levander says, predicting that much of the shipping world will turn to a mix of remote navigation and autonomous operations.

The evolution to “smarter” ships is also likely to spur consolidation among shippers. The average fleet size is eight ships, even as some of the globe’s behemoths—AP Moller-Maersk AS, Mediterranean Shipping Co. SA, Hapag Lloyd AG—operate hundreds of ships worldwide. Those disparities suggest that as more ships are piloted remotely, the number of ship owners will shrink as economies of scale guide the industry to larger, more uniform fleets managed more easily from centralized navigation offices.

Rolls-Royce predicts the first remotely operated vessels will appear by 2020, and that these will become “a common sight on the high seas by 2030.” The pieces are already being deployed in the commercial shipping market. Communication links are being tested and strengthened to achieve the kind of reliability required for unmanned vessels. Rolls-Royce is studying land-based ship control centers in Finland. “If you go 20 years into the future … I am sure we will have fully unmanaged container vessels,” says Levander.

8/3/2036 0447

Port unload.
Destination: Road transport vehicle.

After 26 days at sea, the container ship carrying Caitlin’s shoes arrives at 4:47 a.m., Aug. 3, at the Port of Long Beach, California, on a muggy, gray morning. The ship eases alongside the terminal, communicating with an automated mooring system to position itself, a process that takes about 40 minutes. The ship and cargo undergo a quick digital “examination” from terminal sensors before unloading commences. The soaring gantry cranes overhead are programmed by computers that have analyzed and sorted each container, mapping its destination and devising the most efficient way to unload and load this ship. The cranes slide to and fro across the top of the behemoth vessel, moving each 40-foot box onto an automated rover that drives the container to its rendezvous spot with a rail car or truck. (In 43 hours, our shoes’ container is truck-bound, headed to a depot in San Bernardino County, east of Los Angeles, and then to another autonomous truck destined for Denver.) At the port, small crews of human technicians rotate on shifts to attend to repairs should any crane or rover glitch occur. They also change tires when a rover needs a replacement. On Aug. 3, the machines handling the container holding Caitlin's shoebox operate well: no pauses, no worn-down parts that need looking at. No human gets within 100 feet of the sneakers.

BACK TO 2017

European ports were pioneers in the evolution of automation, back in the 1990s, when shippers and port officials realized that the role of containers in global trade could be further automated with larger cranes and more sophisticated computer software. The container itself was already a half-century old and had caused an earlier revolution in terms of extracting costs from supply chains. Before that, in the age of the longshoremen, most cargo moved to and from ships with large hooks and muscle. The next step was in more efficiently manipulating these metal boxes. Asia followed Europe in advancing the automation of its ports.

For the refinements that would lead to a global network of human-free ports, the schedule is years, not decades, as more specialized technologies mature, such as more sophisticated cranes, containers, robots and automation software. The world’s biggest ports—Shenzhen, Shanghai, Singapore, Rotterdam, Los Angeles—are all heavily invested in automation and eager to expand its role, both for efficiency and to adopt a greener approach to environmental concerns. But port capital projects are enormous in cost and scope and tend to operate on multi-year cycles. Spending is also tied to global trade: Economic dips tend to slow the pace.

On a March morning at the Port of Long Beach, the OOCL Taipei was being unloaded at the Middle Harbor terminal, one of the most automated in the U.S. (Rotterdam is widely considered the world’s most advanced port, in terms of technical automation.) Massive gantry cranes loomed over the roughly 8,000 containers the Taipei had brought from Asia. The cranes have an operator to oversee the delicate aspects of gripping and releasing these 50,000-pound containers, but the cranes’ movements are ordered mainly by computers. These machines coordinate the ship’s loading and unloading, and they help sort out which containers are headed for trucks and which for rail cars.

8/5/2036 0135

Haul cargo via truck.
Destination: Colorado.

It’s 1:35 a.m. on Aug. 5, 2036, and our shoes’ container box is ready to leave the coast after 43 hours in Long Beach. An intermodal trucking company’s autonomous tractor trailer is ready, having an appointment to collect its box at 1:50 a.m. Miss your place in the queue and it’s at least 45 minutes before an autonomous truck will be summoned back to the loading spots. The truck pulls up on schedule. Two minutes later, a crane deposits the container. In five minutes, after additional digital checks, the truck drives out of the port, north into light I-710 traffic. (At this hour, nearly all of it is self-driving; human passengers are asleep.) The truck heads toward a distribution depot in Fontana, California, off I-10, near the Ontario airport. It pulls in 90 minutes later, at 3:26 a.m., and a crane transfers our red metal box off the chassis.

The following Wednesday, Aug. 11, 2036, is a typically hot, summer day in the drylands east of Los Angeles. Our container is being loaded onto a second truck for its trip to Colorado, where the customer wants this load of 500 pairs of shoes stocked as inventory within 72 hours. The trip begins at 6:03 p.m., headed on I-10 west to I-15 to cross California, Nevada and part of Utah, but the truck can detour for efficiencies as its software sees fit. High winds in the mountains near Colorado's Beaver Creek ski area slows our truck a bit, as does a traffic accident outside Las Vegas. Our shoes arrive at the customer’s behemoth fulfillment center in Centennial, Colorado, at 8:40 a.m. The trip is completed in less than 15 hours, the autonomous standard for the 990-mile route.

BACK TO 2017

As a research field, autonomous cars has seen explosive growth—nearly every automaker from Bayerische Motoren Werke AG to Tesla Inc. is aiming to field products and help shape the emerging regulatory landscape, while the Google folks at Alphabet Inc. are equally stoked. The makers of big rigs aren’t lagging behind. (Elon Musk has a deep trucking interest, too.)

At the technical level, the driverless truck isn’t a terribly daunting problem, although loading one with packages is a logistical nightmare at a place such as United Parcel Service Inc. or FedEx Corp. Humans still handle the task of filling trailers, which carry mostly loose packages, plus some cargo that is loaded onto pallets. The “loose goods” method is still the most efficient way to load a truck, in terms of costs, but it's the hairiest for a machine to master. To fully automate loading and unloading, truck interiors would need to become standardized across the industry, with “cages” or other partitions that would segment the space into a platform that a robotic system could address. This would likely increase shippers’ and consumers’ expense.

The big advantages: safety and savings. Under U.S. rules, truckers have an 11-hour daily drive limit, which curbs a truck’s productivity. The autonomous vehicle field is generally ranked on a scale of level zero to level four, with zero the current driving paradigm and four a vehicle that drives as well as a competent human. A fleet operating with a level three standard—a human driver who primarily oversees the automation—could boost earnings by as much as 8 percent, according to a February 2017 report from Cowen & Co. equity analysts. With such automation technology, trucking could regain some of its competitive position versus railroads for long-haul shipping. There is also the matter of performance: Autonomous trucks won’t require rest or medicine to treat chronic conditions. They don't dabble with recreational drink and drugs. They leave no families behind for days or weeks. They require no caffeine or food and bathroom breaks. They don’t get tired, lose focus on the highway or fiddle with radios. The world’s biggest mining companies already use driverless trucks at their larger job sites, which are closed to the public.

Semi-autonomous level 3 trucks are likely to become “ubiquitous” on U.S. highways by 2022-2027, according to Cowen’s analysis. Trucking at levels 3 and 4 is at least a decade hence, and probably more, due to the larger road transportation ecosystem not yet able to achieve “five nine” reliability, or 99.999 percent. This is the competence that support systems will require to make levels 3 and 4 automation feasible.

8/12/2036 0915

Unload inventory at warehouse.
Destination: Unknown.

The container has landed, and robots make quick work of unloading it and shuttling products to their respective storage areas. The shoes are in the northeast corner, a couple of football fields distant from their arrival spot, sitting in aisle 28C. It is Aug. 12, and the shoes await a buyer.

BACK TO 2017

In the age of, warehouses are online retailing’s new department store, with armadas of high-tech robots working as digital clerks. Warehouse construction has boomed in tandem with the growth of online shopping, and the retail giant is leading the way. Internet retail has given rise to the warehouse robot—Amazon now has 45,000 robot staffers—which help to minimize humans’ movement and injuries at these huge fulfillment centers. The robots fetch products faster than people do. Toss in Target Corp., Wal-Mart Stores Inc., Macy’s Inc., Staples Inc. and a few thousand disparate sellers: The needs for robots looks robust.

In 2012, Amazon purchased Kiva Systems Inc., a pioneer in the warehouse robot industry. Three years later, the Massachusetts-based company was renamed Amazon Robotics and began sending all of its 320-lb. orange robots to its parent. That corporate decision has led to a scramble among companies that hope to take Kiva’s place supplying robots to toil in warehouses, according to The Robot Report, a trade journal.

But warehouses and robots extend far beyond Amazon and retail. Most firms involved in logistics—which means nearly any company that is importing goods destined for a customer—are considering the potential role of a robot in their supply chain and fulfillment processes. “It is impossible to overestimate the power of convenience in the American economy,” Erik Nieves, founder and chief executive officer of PlusOne Robotics, said at a 2016 automation conference. In logistics, the best means to ponder the role of robotics is to “count the touches,” he said. “Is there sufficient labor to merit automation?”

Warehouse robots have been hard at work among us for a few years, but improvements are still needed. For one thing, logistics robots’ vision and perception are lacking. They read and drive with motion and other sensors, but they don’t actually “see” their environments. In the future, warehouse robots will become more capable via sensitive cameras and recognition software, called visual servoing, along with the ability to grasp and to address picking and placing actions. These machines will be able to load and unload trucks—a major hurdle to date—and to replace humans when it comes to sorting and filling boxes.

10/26/2036 0345

Transport order to customer.
Final destination: Englewood, Colorado, 80113

Summer’s over and our shoes have been sitting for more than two months. (They’re fuschia, remember, not the most popular color in this Age of Dust.) The retailer’s algorithms are ready to move them. A script in the retail website automatically reduces the price by 33 percent. Caitlin, unable to sleep on a Sunday night before a busy week, starts browsing shoes online around midnight and spies this pair. The price is right, and Caitlin wonders if some new gear will help her get motivated to work out more. She clicks. Her choice of free shipping means the shoes won’t arrive by 7 a.m., or even the next day, but will on Tuesday, Oct. 26. That afternoon, a blinking, spider-like chunk of hardware alights on a loading dock at the Centennial fulfillment center, where a sorting robot has removed our well-traveled sneakers from a stack in aisle 28C and lined them up alongside dozens of similar boxed pairs. The robot affixes an RFID-enabled tracking tag on the shoebox and leaves it waiting to be flown a few miles to its final destination: Caitlin’s bungalow on Englewood’s north side.

The drone reads the encrypted information on the tag and learns the shoe box’s delivery deadline: 5 p.m. local time, Oct. 26, 2036. That’s only 75 minutes from now, so the drone prioritizes this pair over others, feeds the address coordinates into GPS, and ascends to 300 feet. In 29 minutes, Caitlin’s next-door neighbor hears a slight whirring outside and the drone rests the box on Caitlin’s doormat. When she gets home from work and opens the box, Caitlin will be the first person to set eyes on this particular pair of shoes.

BACK TO 2017

The Federal Aviation Administration predicted recently that the U.S. will have more than 420,000 commercial drones flying by 2021, a 10-fold increase from the current number. (The agency’s high estimate is 1.6 million.) “The commercial drone sector is very dynamic and appears to be at an early stage of growth,” the FAA said in its forecast.

Amazon, Google, Flirtey Inc., UPS and others see autonomous drone fleets as the key to even faster access to customers. But as these companies investigate drones, municipalities nationwide are considering laws aimed at drones’ unique snooping abilities. This consumer concern could well affect the growth of commercial drone operations, at least in some areas.

U.S. regulators currently allow only line-of-sight drone activities. The FAA is still formulating rules to incorporate autonomous aircraft into the U.S. airspace. But at some point, the parties involved in U.S. airspace issues will become satisfied that autonomous drones can cope with the various safety and operational considerations. The day is coming when a small drone zipping to your home or office in the name of speedier online retail gratification will be a routine event.

Edited by Francesca Levy
Illustrations by Keith Rankin
Design by Steph Davidson