The Migration of Scarcity

——The Second Act of the Technological Revolution Trilogy

In 1971, Herbert Simon wrote a sentence that has been repeatedly validated over the past half-century: “The abundance of information means the scarcity of what it consumes—attention.” Replace “information” with any general-purpose technology, and the sentence still holds. Technological revolutions do not eliminate scarcity; they merely move it from one place to another. The losers hold onto the old scarcity, while the winners move ahead to new bottlenecks.

Rule: Where Abundance Appears, Value Migrates to Its Complement

Economics has a simple law: When something becomes abundant, value does not disappear but migrates to its complement. When flour becomes cheaper, what becomes valuable is the location of the bakery; when film becomes cheaper, what becomes valuable is the content; when transport capacity becomes cheaper, what becomes valuable is the road network and scheduling. Each general-purpose technology revolution fundamentally involves a large-scale “artificial creation of abundance”—and the outcome of commercial competition depends on who can first see where the scarcity has moved.

Five Waves, One Migration

Electricity: From “Whether there is electricity” to “Whether there is a reason to use it.” After power generation technology matured, power companies found themselves facing a demand scarcity rather than a supply scarcity—factories used electricity during the day, but generators were idle at night and in the early morning. A fascinating scene emerged in business history: power companies and GE desperately promoted electric irons, toasters, and washing machines to households, even extending credit sales. They sold appliances not to earn money from them but to create load. In Chicago, Insull invented time-of-use pricing to move demand to off-peak hours. Abundant parties would subsidize their complements to create demand for their own abundance—a model that was later repeatedly replicated.

Broadcasting: Sarnoff’s Two Moves. RCA’s David Sarnoff first foresaw that radios would become household appliances (the famous “wireless music box” memo), earning a fortune from hardware; then he realized that when every household had a receiver, scarcity would shift from “equipment” to “worthwhile content”—so he founded NBC in 1926. Note the cause-effect relationship: NBC was initially created to sell more radios by producing content. A generation later, the same pattern repeated with television: profits from selling televisions nurtured the television networks, which then became more valuable than the entire manufacturing companies. Hardware moved to abundance, content became the bottleneck, and ultimately, attention became the ultimate bottleneck—this migration route completed three steps in half a century.

Computers: IBM’s In and Out, a Textbook Control Experiment. The same company answered correctly once and incorrectly once. In 1969, under antitrust pressure, IBM separated software and hardware pricing, inadvertently creating an independent software industry—when hardware became abundant, scarcity shifted to software, and IBM caught up. But in 1981, when IBM made PCs, it judged that value was in “whole machine manufacturing,” outsourcing the operating system to Microsoft and the processor to Intel without exclusivity. Clones flocked, and the whole machine quickly became a commodity, while DOS/Windows and x86 chips became new bottlenecks. IBM handed the new scarcity to others, leaving itself the old link in the chain that was about to become abundant. For the next twenty years, most of the PC industry’s profits flowed into Microsoft and Intel’s pockets. The followers learned quickly: Microsoft, Google, and others skillfully used the strategy of “turning complements into commodities”—opening their complementary layers, giving away browsers and Android for free—essentially creating abundance elsewhere to strengthen their own bottlenecks.

Empire Edition of Migration: From Sails to Cables, Coal to Oil. This migration also ruled the fate of nations. After steamships made maritime transport capacity abundant, the British Empire was the first to see where the scarcity had moved: not on the ships, but in the ability to know prices and war conditions before the ships arrived. Britain wove an “All Red Line”—a global submarine cable network with all landing points on its territory—and by 1914, half of the world’s submarine cable mileage was in British hands. Transport capacity was affordable to everyone, but the information channel was unique—empire ruled for another half-century on this new bottleneck (see the chapter “The Morning the Cable Was Cut”). The same empire left a perfect “proactive move” answer: starting in 1911, the Navy Minister Churchill argued for the Royal Navy to switch from Welsh coal, which was secure at home, to Persian oil—using energy security to gain speed and endurance. He fully understood that he was trading an old abundance (coal, which Britain had plenty of) for a new scarcity (oil, which Britain had none of), so the accompanying actions followed closely: in the weeks before the 1914 war, the British government bought a majority stake in the Anglo-Persian Oil Company. Seeing where scarcity will move is passing; buying the new scarcity’s territory before it arrives is a perfect score.

The Internet: The Tragedy of the Pipes and the Coronation of Attention. Telecom operators once believed they held the ultimate scarcity—lines and bandwidth. Bandwidth was indeed scarce, but continuous expansion made it abundant, and value shifted to the complement of bandwidth: search, content, and social relationships. Operators finished the information superhighway, then became “dumb pipes”; Google and Facebook built tollgates on the new bottleneck—attention. Newspapers died from the same migration: the newspaper believed its scarce asset was the printing plant and distribution network, but the profit support was regional monopolies on classified ads; Craigslist took the classified information almost for free, turning the printing plant into a ruin overnight. And the 2000 AOL-Time Warner merger—an epic disaster with nearly a hundred billion dollars in write-downs—was an extreme example of betting on the wrong scarcity: they bet that “channel and content bundling” would be the future bottleneck, while the bottleneck was migrating to user relationships on the open internet.

Why People Hold onto Ruins: Three Cognitive Traps

Legacy-Asset Illusion. The assets accumulated from old scarcity—power plants, copper wires, printing factories, and data centers—shine on the balance sheets, leading to the mistaken belief that “sunk costs” are “moats.” The definition of a moat is something that others cannot obtain, but technological revolutions are precisely producing what everyone can get.
Linear Extrapolation. The dissolution of scarcity is an exponential process: the cost curves for bandwidth, storage, and compute are cliff-like drops. Management accustomed to a linear world will sleep soundly in “still expensive this year,” missing the argument that “it will be nearly free in five years.” To determine whether something is about to become abundant, look at the slope of the cost curve, not the current price.
Identity Binding. “We are a telecom company/newspaper company/hardware company”—organizational identity is bound to old scarcity, leading to the view of new bottlenecks as “side projects” or “off-task.” Those holding onto the ruins are often not unable to see the migration; they are unwilling to admit that their territory has become a ruin.

How to Track the Migration: Three Operational Criteria

First, list all the stages of your supply chain and mark the one that is experiencing exponential price drops. Then, immediately check: is your profit precisely built on that stage? If so, the value is on the move—wait for it at its complement.

Second, find the “queueing” spots. The manifestation of scarcity is always in queues and premiums: waiting in line for telephone installation, then for domain names, and later bidding on search keywords. Today, what is being queued and contested is the new bottleneck.

Third, observe what the abundant side is subsidizing. When giants start giving away something for free (browsers, operating systems, model APIs), it signals their judgment that it will soon be abundant and not worth holding, while their true bottlenecks lie elsewhere. Understanding the giants’ free list is to understand their scarcity map.

For the Age of Compute

Apply these criteria to today, and you can sketch a rough migration map:

Moving Toward Abundance: Model capabilities themselves. Models are rapidly converging, open-source efforts are catching up, and the cost curve for inference is dropping steeply, like bandwidth did. Those who bet their long-term strategy on “my model is smarter than yours” should compare with IBM in 1981.
Where Scarcity Is Migrating: Unique data and proprietary scenario feedback, distribution channels and user relationships, energy and physical infrastructure (this round sees a rare migration “upstream”), verifiable trust (when content can be infinitely generated, “trust” becomes a luxury), and the hardest to quantify—the taste for defining problems. When answers are nearly free, asking becomes the bottleneck.

A century ago, power companies invented the “appliances” category to sell electricity; today, model companies will also incubate new categories without names to sell tokens. The next stop for scarcity has already departed, and the question is: are you still holding onto the ruins at the old station, or have you already bought a ticket to the next stop?