In 1942, Joseph Schumpeter argued that capitalism survives not by reaching equilibrium, but by destroying it. Every major technology — from the power loom to AI — follows the same pattern: devastation, then abundance.
Before Schumpeter, economists imagined the economy as a machine tending toward equilibrium. Firms produce familiar products using familiar methods. Prices settle. Profits tend toward zero. The system runs, but it does not change.
This is what Schumpeter called the circular flow — an economy on autopilot.
Then someone breaks the routine. An entrepreneur introduces a "new combination" — a new product, a new process, a new market. She does not invent; she innovates. She turns an idea into a commercial reality.
The circular flow is disrupted.
The innovator's product is better or cheaper. Customers switch. Competitors scramble. The old firms — the ones doing business the old way — begin to fail. Not because they did anything wrong, but because the world moved.
Schumpeter called this the "perennial gale of creative destruction."
The destruction is not a metaphor. Industries collapse. Workers are displaced. Communities built around a single factory are devastated. The handloom weaver of 1820 is not consoled by the knowledge that his grandchildren will live better.
The pain is concentrated. It arrives first.
The new industries do not merely replace the old at the same scale. They expand total output. The automobile destroyed the horse-and-buggy trade but created an ecosystem — roads, gas stations, suburbs, motels — that employed orders of magnitude more people.
More firms. More jobs. More wealth.
Schumpeter saw that innovations arrive in clusters. Textiles and iron in the 1780s. Railroads and steel in the 1840s. Electricity and chemicals in the 1890s. Electronics in the 1940s. Computing in the 1990s.
Each cluster drives a wave of growth lasting 40-60 years. We call them Kondratieff waves.
At the heart of each wave sits a general-purpose technology — one that transforms not just one industry but all of them. Steam power. Electricity. The computer. Each is pervasive, improvable, and spawns innovations in every domain it touches.
AI is the next one.
AI meets every criterion. It is pervasive — touching healthcare, science, manufacturing, finance, law. It is improvable — capabilities are advancing rapidly. It is innovation-spawning — enabling drug discovery, materials science, and applications that were previously impossible.
The sixth wave is beginning.
Every previous GPT produced initial disruption followed by massive expansion. The pessimists were right about the pain and wrong about the balance. US employment grew from 78M to 157M during the computer revolution.
The question is not whether AI will create. It is whether we will manage the transition.
The pessimistic case assumes a fixed amount of work to be done — the "lump of labor fallacy." But when the power loom made cloth cheap, people did not buy the same amount at a lower price. They bought more cloth, and the freed-up income created demand for entirely new goods and services.
Cheaper computation did not reduce demand for software. It created an industry that employs millions in jobs that did not exist in 1970.
Each technological revolution produces a surge of growth. The destruction is real. The net effect is overwhelmingly positive.
Stylized representation of Kondratieff long waves. Each wave clusters around a general-purpose technology: textiles/iron, rail/steel, electricity/chemicals, electronics/oil, computing/internet. The sixth wave — AI — is beginning.
Schumpeter was no optimist about capitalism's political survival. He feared the social disruption of creative destruction would turn the intellectual class against the system that produced prosperity. The displaced workers and threatened incumbents would form a coalition against innovation.
The solution is not to stop the gale. It is to build better shelter — education, retraining, social insurance. Societies that embraced change and invested in adaptation reaped the benefits. Those that resisted fell behind.
Jobs displaced vs. jobs created across four technological revolutions.
Illustrative comparison. Each revolution destroyed entire occupational categories while creating new ones that were larger, more diverse, and more productive. Source pattern: Acemoglu & Restrepo (2019), historical labor statistics.
The specific forms of the next expansion are unpredictable. The inventors of the railroad could not foresee the tourism industry. The inventors of the computer could not foresee social media management as a profession. The entrepreneurs of the AI wave — the ones who will find new combinations we cannot imagine — have not yet appeared, or are working in garages whose significance is not yet recognized.
What Schumpeter tells us is that they will appear. Wherever a new technology reduces costs and expands possibilities, entrepreneurs rush in to exploit the gap.
Watch firms be born and die. Adjust the rate of innovation and see how creative destruction produces — or fails to produce — net growth.