The Age of aCommerce

…in which a startup finds the next trillion, the climate is cured, capitalism fails in the face of Fully Automated Luxury Communism, and machines get legal personhood.

By Cecilia MoSze Tham and Mark Bünger

If a stranger gave you a trillion dollars, you’d be pretty damn happy. But what if the money was delivered in pennies — scattered around the stranger’s rocky field, thousands of miles on each side? You might as well be broke. The futility of walking around, searching for, and picking up so many pennies is obvious. When it costs more than a penny to pick up a penny, you’re better off leaving the pennies alone.

However, what if instead of using your bare hands, you invent a machine that you can use to find and pick up a penny every second? That means you’d rake in — literally — about $30 million per year. Great! But it would still take three million years to get your money. If you spend your earnings on building a thousand machines and hiring a thousand people to run them, you could get it down to 3,000 years. Still a long time, and pretty expensive. But if you also invest in research to develop penny-picking technology that works a thousand times faster (by autonomously spotting coins, for example, or optimizing raking routes through the field), you’d have your trillion in just three years. Now we have a business plan.

What would compel someone to make this offer? Perhaps it’s a farmer’s field, and the coins are keeping plants from growing. One person’s trash can be another person’s treasure, and making the conversion can be a win-win transaction (in the real world, scrap metal markets work this way). Or, it could be that the owner lacks the workforce or technology to collect the coins, but will let someone else do it for a share of the transaction profits. A real-world example would be an oilfield; the owners would be countries like Saudi Arabia once lacked the technology to drill and extract the oil. Decades ago, international oil companies like Shell and Chevron could control 85% of the world’s petroleum, because the countries where oil was located didn’t have access to technology. Today, oil-producing countries have invested heavily in gaining this technology (with advanced R&D centers like Saudi Aramco’s EXPEC ARC), and now they are the ones controlling 85% of the oil.

In fact, the economy is full of trillion-dollar penny fields — large but distributed markets we haven’t been able to access until now. The cost of picking up each penny is called transaction friction. And a growing number of organizations are investing in the fundamental technology that will vacuum up those little coins frictionlessly — at near-zero cost, and very very fast. They’re scaling up those machines to take over bigger markets. And they’re giving them autonomy, so the machines don’t need human intervention to organize and optimize how those markets will operate in the near future.

This emerging space is the machine economy, and we call frictionless machine-to-machine transactions autonomous commerce (aCommerce).

This article will explain what aCommerce is, what companies and technologies are creating it, and which industries will feel its impact first. Beyond industry transformation, aCommerce will restructure the economy, address the threat of climate change, and lead to algorithms getting legal rights — possibly even personhood.

What aCommerce is

One field full of pennies is energy. The world generates about $2 trillion of electricity each year, for lighting homes, baking cookies, and charging up Teslas. But the losses in energy are massive — of each megawatt generated, as little as 40% actually gets put to productive use. This isn’t stupid waste, it’s inevitable, because even with smart meters, we have no economically feasible way to coordinate the behavior of all the millions of individual users on a typical energy grid on the scale of less than hours (among many other reasons). Because of transaction friction, it costs more money to save that energy than the cost to produce it, so we let it go to waste.

Sources of transaction friction

Impacts of transaction friction

The problem is, smart meters and connected thermostats aren’t all that smart. They can’t act fast enough, they don’t take into account nonfinancial values (like cutting carbon emissions), and people are too confused and busy (not too lazy!) to adjust them constantly. For example, you might set your thermostat to be off when you are out of the house for work. But you probably don’t bother to turn it off on nights when you go out to eat, or when you take a quick trip to the store. You certainly don’t adjust it based on whether you’re wearing a sweater or not. You don’t switch off lights every time you leave the room, you don’t unplug your refrigerator when you are going to run the clothes dryer, and you don’t do a million other little things you could, to save energy. If you do any of these things, you probably aren’t just doing it to save money, but to save the planet. In fact, you might be willing to pay extra for renewable energy (more than 48% of Americans, and 56% of Germans are), but there’s no way to buy the power generated by your neighbor’s solar panels or make your teenager offset their hour-long hot shower with a carbon credit. All of these limitations — saving money not being worth the time it takes, while being unable to spend money to support important goals — are the result of transaction friction. And these consumer examples are trivial compared to the vast amounts of energy wasted in industrial processes.

So imagine now that all the devices in your home, and all the homes in your neighborhood, and all the neighborhoods in your city, could negotiate every aspect of the energy they use. Your car’s battery-management software coordinates with the gym appointment in your calendar and the solar charging station at the gym — and millions of other people’s devices do the same — to help the entire city plan the morning’s collective power production and distribution, and react perfectly to even minute deviations in real time. They could trade not only kilowatt-hours for money, but money for time, time for convenience, convenience for environmental benefit. Because they transact autonomously, they can make decisions that are better optimized globally and still in line with their owners’ values, without the owners needing to intervene on a minute-by-minute basis. Not just because we’re dumb or lazy, but because the autonomous devices pick up more pennies, better and faster than we can.

The technologies making aCommerce happen

Some aspects of the scenario above might sound a little familiar. That’s because developers of smart homes, smart meters, the Internet of Things, and digital assistants all have been describing parts of this future — but launched too early to realize the vision. Take Nest for example, the home automation system that Google bought for $3.2 billion in 2014, or Predix, GE’s industrial IoT network. These proprietary ecosystems are like the Compuserve network — which had electronic newspapers, social media, entertainment, and even shopping — but couldn’t compete with the free and open internet.

Take the example of music, which in 20 years has seen four major models for how we conduct transactions — at each step, losing friction and gaining more choice — so that what was once a major personal expenditure for a tightly limited selection and narrow use case is now effectively ubiquitous, unlimited, and free.

In traditional commerce, you bought a vinyl LP at your local Virgin Records store and played it at home; with eCommerce, you could pick from any CD in the world from Amazon and rip it to your computer; in mCommerce you could buy an individual song in the iTunes store and take it along on your iPhone; today streaming services like Spotify automatically and endlessly create personalized playlists with thousands artists you’ve never heard of yet love, and the cost per song is effectively zero (aCommerce).

The same dematerialization is happening with cars, although it’s not quite as far along: In traditional commerce, you bought a car at a nearby dealership or from someone you found in a classified ad in the newspaper; in eCommerce you could shop online, seeing inventory and prices from hundreds of sellers. Today, with mCommerce, you don’t need the car — you can order one on Lyft or borrow one from Getaround and just pay for the ride. aCommerce will arrive with autonomous vehicles of course, where your ride might be nearly free since it’s shared with other riders or even a package traveling the same route.

Traditional, electronic, mobile, and autonomous commerce

In each of these waves, the financial and operational cost of making a transaction has fallen, thanks to new technology. Using our penny-picking metaphor, the trillion dollars in traditional commerce was spread around in $10 bills (pretty easy to justify picking those up). But the cost of the transaction itself was still high; you would seldom write a check or swipe a credit card (remember the paper slips?) for $1.23, or drive to a store to buy something that cost 15 cents.

Building on the metaphor, eCommerce transactions would be $0.25 quarters, and mobile commerce $0.05 nickels. Despite the convenience of online shopping, the cost and insecurity of credit card payments over the web held back eCommerce until e-wallets like Paypal took those frictions away. Today, mobile payment services like Venmo, Zelle, Apple Pay, Amazon Pay, Visa Checkout, Google Wallet (and local services like WeChat or Alipay in China, Swish in Sweden, and iDeal in Holland) are growing in popularity, and global mCommerce will surpass eCommerce this year. At each step, new technologies have addressed the speed, security, geographic reach, and other aspects of the transaction, taking those frictions down by factors of 10, 100, or 1000.

So why is aCommerce happening now? What are the technologies enabling it, and what impact will they have? A few of the missing ingredients are starting to appear, specifically:

• An open, better network protocol. Protocols are standard languages for network features like communication, payments, and security (such as HTTP, 3G, and Ethereum). In aCommerce, IOTA is a leading example of an open-source, distributed ledger protocol that combines feeless exchanges of value (payments) with secure encryption, anonymity, and scalability. It’s less resource-intensive than blockchains, and more amenable to autonomous financial transactions and smart contracts than industrial M2M protocols such as MQTT and CoAP. Other examples include Skynet and the Computes mesh computing platform — novel swarm network architectures that could make today’s server-centric structure obsolete.
• Faster, more powerful computing. Artificial intelligence on the device. Better sensors, embedded vision. Swarm networks. It remains to be seen how these M2M protocols will utilize — or bypass — protocols that depend on the hardware and infrastructure layers of the stack, like Bluetooth or 5G. Typically, machines would transmit data over these networks to a central cloud-based server for analysis and processing, and receive instructions back. These protocols offer the benefit of global reach and interoperability, but they have an inherent drawback for the machine economy — power consumption. While the power drain of BLE might seem insignificant, it’s still big enough to add a cost to M2M transactions — and thereby limit the lower floor of the transactions it can handle. Both Skynet’s SkyChip and IOTA’s JINN are chipsets optimized to run neural network algorithms locally, and transmit data only when necessary.
• Social readiness is rising. While we used to trust only human bank tellers and doctors, our society of information overload has repeatedly shown its readiness to sidestep real-world logistics and financial negotiation for a more sustainable and convenient lifestyle. In previous waves of commerce technology, immaterial goods — in finance, energy, and travel — have led the way to online and then mobile adoption, before physical goods followed. Today, we already see AI agents taking over in these industries, with consumers more than willing to hand over their stock trades, energy monitoring, and airline tickets to an agent that represents their interests. The synergy of distributed protocols, computing, and society will make aCommerce grow.

The industries that aCommerce will impact first

Usually, an industry looks at a shiny new technology as a way to help solve its own problems — lowering cost to serve customers, for example, or adding revenue from a new market (like when iPods brought new revenue to Apple, which only made computers at the time). The bottom line might improve by 10% to 20% — certainly big enough competitive advantage to get excited about, and maybe even change the game for some time.

What most industries miss, however, is when those technologies give them access to another industry — and vice versa. Before iTunes, you couldn’t profitably sell music for $0.99 per song because that price was less than the cost of the transaction itself. The iTunes revolution brought music industry revenue to Apple, as it destroyed the factories, stores, and jobs associated with the physical media model. Now, iTunes is failing in the face of even lower-friction streaming; and freely-generated music is nigh.

Similarly, aCommerce will erase industry boundaries like the eCommerce and mCommerce waves before it, opening new markets for the first movers, and forcing consolidation among the losers. aCommerce’s early adopters will make money in their existing markets, then jump over industry boundaries to adjacent spaces where historical constraints have kept transaction costs high. Among the industries we see aCommerce hitting first and most forcefully are:

• Energy. Energy is one of the industries that will see the earliest and biggest impact of aCommerce, but that’s because it already has a network controlling what it sells (energy) and the data about it. For example, aCommerce could make viable the decentralized, renewable utility grids that Stanford researchers have been exploring as neighborhood transaction networks — and reduce the total cost of electricity by about 50%. The shift isn’t hypothetical: Innogy, a €37 billion German renewable utility company, has been looking at the machine economy inside and beyond energy.
• Transportation. As Innogy has found, one of the most obvious examples is the intersection between electric utilities and electric vehicles: EVs are like massive appliances that could easily double the demand for grid power, but they’re also energy storage devices that could even out the problematic differences between peaks and valleys in electricity demand, especially when renewable sources are added. aCommerce will facilitate the “buying and selling” of electric power between these millions of machines, but the industry (or startup) that sets the terms could devastate the other. While not as advanced as energy, the auto industry is waking up to this scenario: Tesla was the first to make cars, powerwalls, and solar panels — merging auto and utility in one. BMW and Volkswagen followed suit, launching ChargeForward and Elli respectively to provide “all in one” (i.e. frictionless) negotiated power services for homes, offices, and cars. This isn’t just a green embrace of electric vehicles by automakers. They’re reacting to the threat of ridesharing and soon, autonomous vehicles, which — without the “friction” of needing a driver — will be vastly better-utilized, meaning that we’ll actually need fewer cars to fill our growing transportation needs.
• Food. As vehicles become autonomous, they’ll not only negotiate power distribution, but also distribution of food and other retail goods. That’s critical, because of the $3 trillion in food we produce each year, more than 30% — $1 trillion — is unprofitable to harvest, expires before it is sold, or gets thrown away uneaten. The problem is the massive scale of industrial centralized farming and raising livestock; transporting it on huge ships, trains and trucks; and selling it at big-box grocery stores. Similar stories explain the waste in apparel, electronics, and other globalized industries that are the result of an earlier revolution in transaction friction — intermodal containerization. Toyota is planning an autonomous vehicle, the ePallete, that could play precisely this role by switching seamlessly from delivery of people to packages (with Amazon, DiDi, Mazda, Pizza Hut and Uber as partners). What if the same frictionless demand-responsiveness we see in energy and ridesharing could be applied to distribution of consumer goods? Instead of being thrown out at the grocery store, a box of soon-to-be overripe apples could be delivered from an orchard straight to the restaurant offering an apple pie special tomorrow — lowering the cost to the grower, the chef, and the environment. aCommerce would make it possible for the orchard, the delivery service, and the restaurant to close a deal for a few dollars — too small for any of them to care about individually, but in aggregate across the economy, a trillion-dollar boon.

aCommerce converges adjacent industries into competition

Assessing aCommerce’s long-term impact

We’ve been introducing automation to do these things forever — we just keep getting better and better at it. In fact, we are so ridiculously good at it, we have trouble envisioning even the near future impact of new platforms — of which aCommerce is certainly one. It took less than ten years for Amazon to ruin retail, Google to nix newspapers, AirBNB to displace hotels and gentrify cities, Facebook to go from connecting friends to throwing elections, and Uber to upend both taxis and public transit. aCommerce won’t be like those transitions, it will combine them — and do to the tech giants what they did to the industries they engulfed. What will the next ten years look like?

• More sharing, circular, and service economy, less manufacturing and energy. Frictionless transaction is what’s behind the sharing economy, since information makes it easier to find and get paid for sharing idle assets like tools, cars, and homes. It accelerates the circular economy, too, as people find sustainable products as cheaply and easily as conventional ones, and self-sort recycling becomes as easy as tossing away the trash. To imagine vast quantities of idle assets being better utilized, see cities with 20%, 50%, or 80% fewer vehicles and wasted transportation fuel — but more people and goods reaching their destination more quickly and conveniently. The assets that do survive will be optimally maintained, repaired, and remanufactured in a perfectly-efficient flow. Every kilogram of food is produced to order, healthy, and enjoyed — with any residual waste composted and packaging recycled. With less need for raw materials or transforming them into stuff we don’t need, manufacturing sectors will transform into service economies even faster than they are today.
• Tweaking capitalism fails in the face of Fully Automated Luxury Communism. There’s already been a slew of analysis on whether and when artificial intelligence will destroy (or create) jobs, and aCommerce explains why and how AI will do it. In preparation for a future economy where hundreds of millions of humans may be completely sidelined (and $16 trillion in wages lost), innovative economists and tech billionaires like Elon Musk have been testing out ideas like Universal Basic Income (UBI) or taxing companies that build job-taking AIs. But while US politicians threaten to break up its Big Tech competitors Facebook, Amazon, Netflix, and Google (FANGs) and wastes time on self-defeating tariffs against China, that country is quickly developing both the technology and the society that will make aCommerce revolutionary. Long before the US government gets its hands on the FANGs, competitors like Baidu, Alibaba, Tencent (the BATs), Lenovo, and Huawei could adapt their autonomous commerce platforms (where mCommerce already reaches three times as many users as the US) to local markets in the Americas, Europe, and Africa. If the FANGs are too crippled to compete in aCommerce, there won’t be anything left of them to break up. On the bright side — for the jobless, but not fans of capitalism — we might achieve Fully Automated Luxury Communism.
• Machines get legal personhood. In order to be able to (trans)act on our behalf, machines will need to take both responsibility and credit for their actions. If an algorithm hurts a human (from simply ruining a special dinner with a misplaced order, to injuring a trolley driver in a collision) it will need to pay for the damages. But to have something to pay with, the algorithm will need to own assets — presumably ones it earns while it works for its owner. While this may sound outlandish, it’s literally what we do for corporations today, and lawmakers in Europe and elsewhere are already debating it. The centuries-old notion of corporate personhood is so applicable to machines, that law professor Shawn Bayern proposed it in 2015. In addition to property ownership, aCommerce opens the door to other AI rights like free speech, privacy, and possibly even citizenship. When machines are transacting with machines on our behalf as well as their own, trillions of times a second, a new age will truly be upon us.