9781847927125

THE AGE OF EXTRACTION

ALSO BY TIM WU

The Master Switch

The Attention Merchants

The Curse of Bigness

THE AGE OF EXTRACTION

HOW TECH PLATFORMS

CONQUERED THE ECONOMY AND THREATEN OUR FUTURE PROSPERITY TIM WU

ALFRED A. KNOPF NEW YORK 2025

THE BODLEY HEAD LO NDON

The Bodley Head, an imprint of Vintage, is part of the Penguin Random House group of companies

Vintage, Penguin Random House UK , One Embassy Gardens, 8 Viaduct Gardens, London sw 11 7bw

penguin.co.uk/vintage global.penguinrandomhouse.com

First published in Great Britain by The Bodley Head in 2025

First published in the United States of America by Alfred A. Knopf in 2025

Copyright © Tim Wu 2025

The moral right of the author has been asserted

Penguin Random House values and supports copyright. Copyright fuels creativity, encourages diverse voices, promotes freedom of expression and supports a vibrant culture. Thank you for purchasing an authorised edition of this book and for respecting intellectual property laws by not reproducing, scanning or distributing any part of it by any means without permission. You are supporting authors and enabling Penguin Random House to continue to publish books for everyone. No part of this book may be used or reproduced in any manner for the purpose of training artificial intelligence technologies or systems. In accordance with Article 4(3) of the DSM Directive 2019/790, Penguin Random House expressly reserves this work from the text and data mining exception.

Printed and bound by Clays Ltd, Elcograf S.p.A

The authorised representative in the EEA is Penguin Random House Ireland, Morrison Chambers, 32 Nassau Street, Dublin d02 yh 68

A CIP catalogue record for this book is available from the British Library

hb isbn 9781847927118

tpb isbn 9781847927125

Penguin Random House is committed to a sustainable future for our business, our readers and our planet. This book is made from Forest Stewardship Council® certified paper.

For Essie

CHAPTER 1 The Genius of the Ancient City Square 13

CHAPTER 2 Platformization 22

CHAPTER 3 The Golden Age of Tech Optimism 41

CHAPTER 4 From Enablement to Extraction— the Story of the Amazon Marketplace 48

CHAPTER 5 Scale as a Weapon 58

CHAPTER 6 The Great Harvest 66 PART II

THE BUSINESS OF HERDING

CHAPTER 7 A Long Slow Bet on Laziness 73

CHAPTER 8 Big Data, Knowing the Future, and Controlling the Future 82

CHAPTER 9 Artificial Intelligence and the Calculus of Human Dependence 89

CHAPTER 10 Platform Power Beyond Tech 103 PART III

THE DANGERS OF CENTRALIZED ECONOMIC POWER

CHAPTER 11 Economic Mania 119

CHAPTER 12 Some Solutions 125

CHAPTER 13 The Persistent Dream of the Self- Correcting Economy 135

CHAPTER 14 Artificial Intelligence and Crypto: The Technological Answers to Economic Inequality 142

CHAPTER 15 Mere Redistribution 152 PART IV

AN ARCHITECTURE OF EQUALITY

CHAPTER 16 Platforms and the Architecture of Equality 159 Epilogue 173

177

179

195

THE AGE OF EXTRACTION

INTRODUCTION

In the spring of 2024, Apple briefly ran an extraordinary advertisement. It depicted various instruments of cultural achievement— books, a record player, an upright piano—all being crushed by a giant press. A trumpet, on its end, was twisted and bent; a piano collapsed, and paint cans exploded. By the end, everything had been magically reduced into a single device—an iPad—which was described as “the thinnest Apple product yet.”1

“The only thing that could have made it more dystopian,” wrote one commentator, “would have been if actual human beings had been playing the instruments, reading the books, or wielding the paintbrushes.”2 Apple quickly announced that the advertisement was not intended to send any larger message. But as Freud once said about jokes, advertisements often reveal uncomfortable truths. And the truth referred to in that advertisement is one we all know: that we live in a time where human experience has been transformed. The spectacle of human achievement crushed into a tiny device, possessed of its own intelligence, resonates with something we are all experiencing—a sense that as we augment humanity, we may, at the same time, have come to marginalize actual humans.

It was at roughly the same time that the artificial intelligence firm OpenAI debuted a new voice that sounded strikingly similar to actress Scarlett Johansson.3 Johansson had previously refused offers to license her voice, but apparently the firm went ahead and created a replica anyway. She is in a better situation than the golfer Jack Nicklaus, however, who sold the rights to his name and likeness and has been unable to control the conduct of an AI-replica named “Digital Jack,” operated by a firm named Soul Machines.4 It speaks to the apparent replaceability of even the most famous of humans.

The fears of our diminished importance arrive at a time when many have come to feel a parallel sense of economic marginalization. Over the last quarter century, even when the economy has been healthy by traditional indicators, many people nonetheless remain doubtful about their economic future. There is an entire generation who worked hard, and still cannot afford decent housing. They fear being worse off than their parents, something once unheard of in this century. The contrast with corporate citizens is stark; many corporations seem to have escaped any usual orbit and blasted off into an outer space of perpetual profit, bringing their executives along for the ride.

We must see that the problems of technological and economic marginalization are entwined. Technology has never been neutral, but rather reflects ideology and what it is designed to do. Today’s great tech platforms are impressive, entertaining, and convenient, but also designed to be some of history’s most advanced tools for extracting wealth and resources from the broader economy. Consequently, as they become essential to everything, we are at risk of building an economy that is perpetually unfair for much of humanity.

The phrase “wealth extraction” is the key to this book. It refers to the ability to take money from everyone else and is born of being essential and unavoidable. As the tech platforms have grown and evolved over the last decade, they have focused their attention on refining their methods of extraction. In return for an undeniable and unescapable utility, they are fine-tuned to take as much as possible—data, attention, profit margins—from everyone else.

We remain in the early days of platform capitalism and commercially relevant artificial intelligence. But we risk falling into a two-class age, where many industries become divided into two groups: the extractors and their agents on the one side; dependent businesses, consumers, and employees on the other. There is every reason to fear living in a future with a technologically armored wall between the haves and have-nots. We should also fear a future in which the private power and wealth aggregated in the tech platforms comes to influence and combine with the public powers of government.

The main goal of this book, then, is to help readers understand the emergent form of economic power in our time—the artificially intelligent tech platform. You may already work in an industry deeply influenced by platforms or have had your life affected by them in other ways. If not, you should know: they are probably coming for you.

This book also aims to answer a question: Just what happened to the broad spread of prosperity and democracy many expected to follow the Internet revolution? Back in the 1990s and 2000s, many believed that the popular Internet would make everyone better off in an evenhanded manner while spreading democracy around the world. That was wrong: a handful of platforms and owners have taken the lion’s share of the new cash, and it is autocracy, not democracy, that is on the rise. If we are to imagine a better future, we need to understand what happened and why those prophecies did not come to pass.

To that end—in order to understand our past and our future— this book begins by explaining the origins of platform power. The power of the tech platforms relies on ancient economics: there have been essential platforms in every civilization. Platform power also bears some similarity to other historic forms of economic power, such as land ownership and the industrial power of manufacturing. But platform power is distinct from both of these, as its power does not lie in production but rather in catalysis. It resides in the hosting of economic activity and the extraction of value, including the harvesting of specialized assets like data and human attention. As we shall explain, the tech platforms have combined old principles and new technologies to become the ascendant

economic powers of our time. That’s why few can ignore what it will mean for the ongoing transformation of the world’s economy.

IS THIS A PROBLEM THAT SOLVES ITSELF?

Some may find the preceding paragraphs much too gloomy and pessimistic. Technological advance and prosperity have always gone hand in hand, after all. Surely we should expect advances in platform technology and artificial intelligence to make humanity wealthier and happier. Even if some have ended up with a little more right now, we ought to expect everyone to benefit over the long run. This particular brand of tech optimism has well-known adherents. Jeff Bezos, Amazon’s founder, has repeatedly suggested that we live in the best of times. Marc Andreessen, a venture capitalist well known in Silicon Valley, writes that “there is no material problem—whether created by nature or by technology—that cannot be solved with more technology.”5

Sam Altman, the CEO of OpenAI, is among those who believe that technological advances are our best solution to the problems of uneven wealth. To his credit, he takes seriously the fact that humanity is suffering from a spread of inequality. He is the cofounder of a group known as “Tools for Humanity,” which identifies humanity’s “grand challenges” as “global income inequality, governance of existential risks, and distinguishing humans from artificial intelligence.”6

Altman argues that improving artificial intelligence is the clearest way to solve economic inequalities and the problem of poverty. In a manifesto written in September 2024, Andreessen wrote that once we achieve better AI, “we can have shared prosperity to a degree that seems unimaginable today.”7 In an interview with author Nate Silver, Altman said, “If you have something like an AGI [artificial general intelligence], I think poverty really does just end.”8

I take Altman as well-meaning and there is an admirable optimism in this declaration. But it is, at best, a statement of faith. The development of super-intelligence becomes our salvation. History, unfortunately, gives us a more realistic picture of the

impact of new technologies on the distribution of wealth in society. It is unquestionably true that technological change creates wealth. It is the division of those spoils that has always been the tricky part. And too often, technological advances have been used to widen, not narrow, economic divides.

There have certainly been technologies, such as the agricultural plow, that have spread new productivity and wealth to property owners around the world. But there have been other technologies, such as the cotton gin, that have taken a bad situation and made it worse. In the case of the cotton gin—which perpetuated plantation slavery—the problem wasn’t the technology itself but the economic structure of the American South.

The nations that have actually succeeded in sustaining longterm growth and equality have not, historically, taken a trickledown approach. Wealthy but equal nations like Denmark, Japan or Taiwan aggressively divided productive assets. During its best and most forward-thinking years, the United States actually sought to balance economic power, whether through the broad distribution of land and other productive assets, the breaking up of monopolies, protecting worker organizing, or building a social safety net. But too often, the United States accepted a centralization of economic power, as in slave states before the Civil War, or the tolerance of monopolies in the 1890s and 1920s. The latter approach has not gone well, in the United States or elsewhere.

It is unscientific to take technology as some kind of omnibenevolent godhead that will act by itself to solve humanity’s problems. What technology will do for us depends on how we design it and where economic power resides. That’s why the real challenge for humanity is to design an economic future that works for more people, rather than assuming that future will just arrive someday.

THE BETTER ALTERNATIVE: STRUCTURAL BALANCING

“Power ought to serve as a check to power,” wrote Montesquieu in 1748. Every successful political system includes checks on the abuse of power. But most of our political theory was invented

with the problem of the tyrannical state in mind. What we need to recognize, in this age, is that the same traditions are the key to our economic future. For power is power, whether private or public.

You may be skeptical that the concept of balance has any place in the world of commerce, given that capitalism is said to thrive on animal passions. You might think that balance would hinder growth or dampen incentives. But when done well, this is simply not true. Take as an example an institution that could not be more commercial, yet puts economic balancing at its core. That institution is the American National Football League.

It may be surprising to hear it, but the United States’ National Football League has long embraced the kind of balancing favored by figures like Montesquieu. It is a simple fact that a city like Los Angeles or New York has a larger local media audience than a city like Kansas City or Baltimore. Without balancing, the teams in large, rich cities would employ all the best players and dominate their rivals. It would be a two-class league: divided between perpetual winners and losers.

The NFL intervenes strongly to prevent that outcome. It wisely avoids interventions that might seem unfair, like giving smaller cities a ten- point lead. The interventions are, instead, structural, centered on salary caps, the draft, and schedule. The league also has a strong union that bargains collectively for the players, and the league divides advertising revenue evenly among teams. The result is that a city that is smaller and less wealthy, like Kansas City, can compete with giant metropolises— and win.

Some might consider the NFL’s rules to be heavy-handed. But does the NFL embrace fairness at the price of prosperity? The answer is no: the NFL is the most valuable sports league in the world. More generally, as The Wall Street Journal points out, American sports leagues are more valuable than the unbalanced European leagues, despite having smaller audiences.9 The point is that prosperity, fairness, and growth are not incompatible. And when it comes to nations, as the World Bank affirms, it is the

more equitable, economically balanced countries that tend to be wealthier on a per capita basis.

Structural balancing, done correctly, speaks to an enlightened self-interest that can yield greater wealth and prosperity for all. Tocqueville wrote extensively on this: “the principle of selfinterest rightly understood.”10 We would live in a much wealthier and broadly prosperous world if we could draw on the full potential of the entire population, and not just a few regions, classes, or companies.

Our extraordinary technological revolutions have given humanity great abundance. For the first time in history, we have the ability to create a sustainable and happy prosperity for all. Jeff Bezos was correct that we have more than ever before, with technologies capable of producing more than enough for everyone.

What we don’t have is the structure—an architecture of equality to match current technological realities. The program outlined in this book calls for strong anti-monopoly policies meant to curb obvious and illegal aggregation of power. This is a project that I’ve been personally involved in for more than a decade, and one that is underway in the 2020s with American and European lawsuits against Google, Amazon, Facebook, and others. But for the longer term, we need neutrality rules for platforms that both preserve the economic flourishing that platforms catalyze and also stop those same platforms from extracting too much from everyone else.

What this book will show you is how we can do it. That means a journey with several steps. First, it is essential to learn the basics of platform economics. The neutral platform is the foundation of any broadly prosperous society, being a structure that catalyzes economic activity. But it can also become an instrument of undue wealth extraction, as we learned in the age of the railroads, and as we are learning again today. The next step is to understand the specific history of today’s tech platforms: the first computing

and communications platforms run by IBM and AT&T, and the government interventions that forced them to serve the broader economy. From that point, we can understand the current generation of platforms—Google, Facebook, Amazon, Microsoft, and the rest—and their adoption of the business models we experience today. The second half of the book takes a step back to understand the dangers of centralized economic power over a broader historic context, examining what kind of interventions have worked and will work to restructure the economy.

A large and wealthy middle class is key to our future prosperity. It has been, at times, a trademark of democratic nations, and it should be our goal. If we do it right, tech platforms will actually play a major role in creating and sustaining a broadly wealthy country, and also in creating an economic model worth exporting to the rest of the world. But if we get things wrong, we risk a future in which our technologies actively worsen the division and resentment that are the curse of our age.

UNDERSTANDING PLATFORM POWER

Not so long ago, in the early 2000s, we lived in an age of extraordinary optimism about the Internet and what it would do for all of us. It would interconnect all humanity, give everyone a creative outlet, and make democracy blossom around the world. Economically and socially, it would empower the little guys—favor entrepreneurial individuals—at the expense of faceless corporations. As tech pundit Jeff Jarvis wrote in 2009, “small is the new big.”1 He wrote that “a tiny start-up can become a manufacturing company using somebody else’s factory and distribution while selling to a worldwide market. [. . .] The Lilliputians have triumphed. The economies of scale must now compete with the economies of small.”2

Jarvis was hardly alone in his optimism. Business writer Seth Godin titled an entire book Small Is the New Big (2006). Professor Yochai Benkler, then at Yale, argued that self-organizing production systems would compete with and possibly overcome managerial or price-driven systems.3 Richard Florida’s 2003 The Rise of the Creative Class assumed a world in which creators would easily become the agents of their own economic destiny. One way or

another, the message was the same: in the Internet age, the spoils would belong to everyone.

Not very often are so confident a set of predictions so wrong. The reason they were wrong is important to this book and essential to learn from. The key mistake was a failure to truly understand platforms and their unique brand of power. For these writings and projections came at a time when several open platforms—the Web, the Internet, a disciplined Microsoft—really did favor the little guy. But it was not to last, as we shall see.

What these writers did not predict is that it would not be the little guys who prospered, but their hosts—a new generation of platforms that would find ingenious ways to take the lion’s share of the Internet economy for themselves. To understand this point and how it happened, we must back up to introduce the very basics of platform economics.

THE GENIUS OF THE ANCIENT CITY SQUARE

Everything needs to happen somewhere. That is why every civilization has had specialized spaces that facilitate commerce, speech, and other activities. In ancient Greece, the town square, or “agora,” served not just for buying and selling stuff but also for religious festivals, entertainment, and government.1 The bazaar was invented in the Middle East and much of commerce in ancient China centered on the market-town, or 市. These are the ancestors of today’s tech platforms, and we need to understand what gave them their economic significance.

It might help to better define what we mean by a “platform.”

(The English word comes from the French platte fourme or “flat form.”) It can be described as any space or structure that in one way or another brings together two or more groups to transact or interact while reducing the costs of doing so. They can be buyers and sellers, but also readers and publishers, listeners and speakers. And as the French word suggests, a platform usually implied a certain evenhandedness.

This definition of a platform covers a lot of ground. It covers the most ancient form of transactional platform just described: the city marketplace. The term encompasses more, including