FROM CAPITAL TO COMPUTE TO CURRENCY: How the Engine of Capitalism Changed, and What We Must Build Next
Isaac Megbolugbe
September 2026
An Introduction to a Three-Part Inquiry
There is a straight line through the history of modern wealth, but we rarely see it because we live inside it.
It begins with a simple question that Adam Smith first asked in 1776 and that we have been answering ever since: What makes an economy grow? His answer — accumulated stock, or capital, put to work to make labor more productive — became the foundation of capitalism. Capital ceased to be just money. It became the engine. It was the loom, the steam mill, the factory.
For two hundred years, that engine ran. It built empires and industries. But engines, as Marx warned, do not just produce. They begin to rule. Capital became not just a means, but a master — money that must become more money — and eventually the name of the entire system.
This inquiry begins there, with that first transformation: Who made the role of capital central to capitalism? We trace how capital moved from Smith’s tool to Marx’s protagonist to Sombart’s label for modernity itself.
Then a second transformation occurred, one most economic histories still miss. In the 1940s, we invented a new kind of capital that did not just make muscle stronger but made calculation cheaper. For seventy years, computing moved from the back office to the product to the market to the means of production itself. We did not just produce computers; we began to live inside them. And when a computer learns to synthesize knowledge — not just process data — computing stops being an industry inside capitalism and becomes the engine of capitalism itself. That is the subject of our second essay: *When and how computing becomes the engine of capitalism*, culminating in AI.
But that brings us to a third, more urgent transformation that is happening right now. If computing is the engine, it runs on two fuels that are now inseparable: computing power and energy. You cannot have one without the other. A chip without gigawatts is sand. A gigawatt without chips is just heat. Together, they have become the primary currencies of global progress — the new gold standard upon which wealth, industrial power, and national sovereignty are pegged.
This structural rewiring — from processing data to synthesizing knowledge — rewrites the rules of everything. Wealth is no longer accumulated, it is synthesized. Supply chains become inference chains. Nations that cannot make, power, and control their own intelligence stack become digital colonies, exporting raw data and importing finished intelligence.
This is where economics becomes existential. Without a deliberate framework, this concentration leads to only one outcome: chaos. A world of extreme cognitive inequality, energy wars for gigawatts, and a feudalism of intelligence where two or three stacks own the future.
Hence our third essay: A strategic framework for the emerging economy where computing power and energy act as primary currencies. What should we build to ensure this new engine serves both human flourishing and geopolitical rebalancing?
Woven through all three essays is a profound irony we must confront: We dreamed of dematerializing the economy into the cloud, into weightless software and ethereal intelligence. Yet to produce the most abstract thing humanity has ever created — artificial thought — we need the most concrete things on Earth: more copper, more concrete, more water, more reliable gigawatts than any industry before it. The tool we built to decentralize power has become history’s greatest centralizer. And the intelligence we need to make abundant for flourishing, we must also make scarce for stability.
This is the story of how capital became central, how compute became capital, and how compute and energy together became currency. It is the story of how we got here, and the blueprint for what we must do next before the engine we built outruns the humanity it was meant to serve.
WHO MADE CAPITAL THE CENTER OF CAPITALISM?
How a Tool Became a Master and Then a Name
Capital did not start at the center. For most of human history, wealth was land, labor was people, and money was just a measure. To make capital central took three intellectual revolutions.
I. Before Capital: When Land Was Wealth
For Aristotle and the feudal world, the economy was a household — _oikonomia_. Wealth was what the earth yielded. Merchants who made money from money were suspect. Even up to the Physiocrats in France in the 1750s, led by Quesnay, only agriculture was truly productive. All else merely transformed it.
Capital existed — ships, tools, coin — but it was not central. It was auxiliary.
II. Adam Smith — Capital Becomes the Means (1776)
Adam Smith’s move in _The Wealth of Nations_ was quiet but total.
He redefined wealth. Wealth is not gold in the king’s treasury. Wealth is the annual productive power of a nation.
What makes productive power grow? Two things: division of labor and the stock that makes division possible.
That stock is capital.
Smith wrote: to divide labor you must first accumulate. A weaver cannot specialize without a loom. A loom must be saved for from past revenue. Therefore, frugality and saving — accumulation of capital — is the private virtue that creates public growth.
For Smith, capital is central in a *technical sense*. It is the engine that lets labor become more productive. It is tools, machines, inventories, and the wage fund that sustains workers while they work.
He did not yet call the system capitalism. He called it the system of natural liberty. But he made it impossible to explain growth without capital at the center.
III. David Ricardo and the Classicals — Capital Becomes a Class (1817-1848)
Ricardo, Mill and the classical economists hardened Smith’s insight into a social structure.
Society now has three classes corresponding to three incomes:
– Landlords get rent
– Capitalists get profit
– Workers get wages
Capital is no longer just a loom. It is the social power of the class that owns the loom. Profit is the reward for advancing capital and taking risk.
This class analysis set the stage for the next man, who would take it to its logical extreme.
IV. Karl Marx — Capital Becomes the Master (1867)
If Smith made capital central as a means, Marx made it central as a master.
In Das Kapital, Marx’s definition is famous: Capital is not a thing. It is a social relation in motion — M-C-M’. Money buys Commodities (including labor power) to make More money.
The circuit must expand. A capitalist who does not reinvest surplus value ceases to be a capitalist.
Marx therefore flipped Smith:
For Smith: Capitalists use capital to employ labor.
For Marx: Capital uses capitalists to employ labor to expand itself.
This is why Volume 1 is called Capital. It is the protagonist. It has its own laws: accumulation, concentration, centralization, crisis. The system should be named after its protagonist. He called that system the capitalist mode of production.
Marx is the reason we today say the system is capitalism and debate whether capital rules us. He made capital the subject of history, not just of economics.
V. Werner Sombart and Max Weber — Capital Becomes a Spirit and a Name (1902-1905)
Ironically, the word “capitalism” was rarely used by Smith or Marx in English in the 19th century. The man who put it in the dictionary was the German economist, Werner Sombart.
In Der moderne Kapitalismus (1902), Sombart argued capitalism is not just an economy but a mentality — the spirit of calculation, rationality, enterprise, endless accumulation. He traced it from double-entry bookkeeping to the stock exchange.
His contemporary Max Weber, in The Protestant Ethic (1905), added that this spirit needed a cultural ethic to become central — a religious calling to treat profit as a sign of grace.
Sombart gave capitalism its name. Weber gave it its psychology.
VI. Afterward: From Center to Everything
Once named, capital became central in three competing 20th century stories:
1. Neoclassical economics— Capital is one factor of production alongside labor and land, with a price (interest) set by supply and demand.
2. Keynesian economics — Capital is central but unstable. Its accumulation depends on uncertain expectations. The state must manage it.
3. Financial capitalism — From the 1970s onward, capital becomes self-referential. Not machines, but financial claims on future income.
Conclusion: Three Makings
So who made capital central?
– Adam Smith made it central to how wealth is made.
– Karl Marx made it central to how power works.
– Werner Sombart made it central to what we call the system.
Smith gave capital its function. Marx gave it its dominance. Sombart gave it its name.
We live in the convergence of all three — in a system that cannot produce without accumulation, cannot explain power without capital, and cannot name itself without it.
WHEN COMPUTING BECAME THE ENGINE OF CAPITALISM
From the Engine of Production to the Engine of Prediction
If capital was the engine of industrial capitalism, computing is the engine that now drives capital itself.
Capital, as Smith defined it, was saved-up labor embodied in a loom, a steam engine, a factory. Its job was to make labor more productive.
Computing did not replace that job. It absorbed it, accelerated it, and then changed what capital wants to accumulate.
That shift happened in five stages.
I. 1945-1970: Computing as Capital’s Calculator
Early computing — ENIAC, IBM 701, Sycor — was not an engine. It was a super abacus.
Corporations like General Electric, U.S. Steel, and insurance companies bought mainframes to do what clerks did: payroll, inventory, ballistics, census.
In this phase, computing _was_ capital in Smith’s old sense. An expensive machine you accumulated to make existing production more efficient. It lowered the cost of calculation, not the logic of capitalism.
The metaphor was still factory: The computer was in the back room.
II. 1971-1990: Computing Becomes a Product — The Microprocessor
Two inventions in 1971 changed the engine:
1. Intel 4004 — the microprocessor put the computer on a chip.
2. Floppy disk and later hard disk — memory became storable and sellable.
Suddenly computing itself could be mass-produced and accumulated like cars. The Apple II (1977), IBM PC (1981), Microsoft DOS (1981) created a new industry where the product was computing.
Capitalism’s engine was still production, but now it was producing computers. The capital accumulation loop was:
Money -> Chip Factory -> More Chips -> More Money.
This is what Moore’s Law (1965) made economically explosive: every 18 months you get twice the power for the same capital. No steam engine ever did that.
Computing became the most profitable place to put capital, because its productivity doubled while its cost fell.
III. 1991-2007: Computing Becomes the Market — The Network
The internet turned the computer from a product into a market.
When Tim Berners-Lee added the World Wide Web (1991) and Netscape commercialized the browser (1994), the value was no longer in the box but in the connection between boxes.
Three new forms of capital emerged:
1. Network capital: Metcalfe’s Law — value grows by the square of users. A telephone with one user is worthless. With a billion users, it is priceless.
2. Platform capital: eBay (1995), Amazon (1994), Google (1998) discovered you do not need to own inventory. You own the market where inventory meets buyer. The platform is the loom, but it weaves other people’s labor.
3. Data capital: Every click, search, purchase leaves a residue. In industrial capitalism residue was waste. In network capitalism residue is data — the new raw material.
Capital was no longer just accumulating machines. It was accumulating users and their traces.
IV. 2008-2022: Computing Becomes the Means of Production — The Cloud and the App
The iPhone (2007) and Amazon Web Services (2006) completed the inversion.
Before: You bought a computer to run your business.
After: You rent computing from someone else’s computer to run your business.
AWS, Azure, and Google Cloud meant that computing became what economists call general-purpose capital — like electricity in 1900. Startups did not need to save to buy servers. They rented cycles by the second.
Simultaneously, the app economy turned every human activity — hailing a taxi (Uber 2009), staying somewhere (Airbnb 2008), listening to music (Spotify 2008) — into a computable transaction.
This is the key shift: In Smith’s world, capital made labor productive. In this world, *computing makes consumption productive.* Your leisure, your ride home, your conversation, produces data that can be capitalized.
Shoshana Zuboff calls this surveillance capitalism. Economists call it rentier platform capitalism. Both agree: The engine is no longer the factory producing goods. It is the platform producing predictions about you.
V. 2023-: Computing Becomes Capital Itself — AI
AI, especially Large Language Models, is the final stage. Why is it different from all previous computing?
Because all previous computing needed you to tell it what to do. AI does what capital always wanted to do: produce surplus without adding labor.
Four properties make AI the engine of contemporary capitalism:
1. AI collapses the cost of cognition. Industrial capital mechanized muscle. AI capital mechanizes pattern-recognition, language, code, diagnosis — what used to be called skilled labor.
2. AI turns data into capital that makes more capital. In classic capitalism: M-C-M’ — Money buys a Commodity to make More money. In AI capitalism: D-M-D’ — Data trains a Model to make More Data. The model is both product and means of production. It appreciates the more you use it.
3. AI creates increasing returns, not diminishing returns. A steel mill gets less efficient after a point. A foundation model gets _more_ efficient after a point — more data, more users, better model, more users. This is why capital is flooding into it. It is the first capital in 250 years that violates the law of diminishing returns at scale.
4. AI makes prediction the product. Industrial capitalism sold goods. Platform capitalism sold ads. AI capitalism sells predictions — what you will write, buy, want, do next. The most valuable commodity is no longer the present good but the future behavior.
In this sense, computing has come full circle. It started in 1945 as a tool to calculate payroll. It is now a system that calculates you, and then sells that calculation.
Conclusion: From Loom to Model
Smith’s capital was embodied labor: a loom that holds yarn.
Marx’s capital was a social relation: money that must become more money.
Computing capital is embodied knowledge: a model that holds human language, images, behavior — and then weaves them into new yarn without a weaver.
That is why computing became the engine. Not because we have more computers, but because capitalism discovered that the most profitable thing to accumulate is no longer the machine that makes things, but the machine that makes predictions about people.
The engine changed from production to prediction. And AI is its most efficient fuel yet.
THE COMPUTE-ENERGY STANDARD: A Strategic Framework for an Economy Where Processing Power and Power Itself Are Currency
The Rewiring
For seventy years we lived through a slow inversion.
1945-2000: Computing was a tool inside the economy. A faster way to do accounting, logistics, payroll. The economy produced computers.
2000-2023: Computing became the market around the economy. Platforms mediated exchange. The economy lived inside computers.
2023-: Computing became the economy When you synthesize knowledge — not just process data — the ability to synthesize is wealth generation, industrial organization, and sovereignty at once.
We are shifting from the Bretton Woods system where currencies were pegged to gold and then to dollars, to an implicit new standard where national power is pegged to two inseparable commodities:
1. Compute — measured in FLOPS, in trained models, in inference per second.
2. Energy — measured in reliable gigawatts, not just total terawatt-hours.
If you have one without the other, you have nothing. Compute without energy is an idle chip. Energy without compute is heat. Together, they are the primary currencies of global progress.
This rewiring changes the rules.
How The Rules Are Being Rewritten
1. Wealth Generation: From Ownership to Synthesis
Industrial wealth = Owning means of production
Digital wealth = Owning means of distribution (platforms)
Cognitive wealth = Owning means of synthesis (foundation models + compute + data)
A country that can synthesize a new drug, a new material, a new supply chain in 72 hours is wealthier than a country that owns ten factories that make last year’s drug.
Wealth is no longer accumulated. It is synthesized on demand.
2. Industrial Organization: From Supply Chains to Inference Chains
The 20th century supply chain was: Mine -> Make -> Move -> Sell.
The emerging inference chain is: Sense -> Model -> Simulate -> Synthesize -> Deploy.
TSMC, NVIDIA, ASML are not tech companies. They are the new oil majors, steel mills, and shipyards combined. The fab is the refinery.
3. National Sovereignty: From Borders to Stacks
Sovereignty used to be defined by territory, then by currency reserves. Now it is defined by your place in the stack:
– Bottom of stack: Do you have sovereign energy and sovereign silicon? Can you make and power your own chips?
– Middle of stack: Do you have sovereign models? Or are you renting intelligence from 3 companies in 2 countries?
– Top of stack: Do you have sovereign data and sovereign applications in your language, your law, your values?
A nation with no sovereign compute is a digital colony. It exports raw data and imports finished intelligence — the exact structure of 19th century colonialism, only the raw material is human behavior.
What We Must Do: A Five-Pillar Framework
Without deliberate design, this system leads to chaos: extreme concentration, energy wars, and a world where 80% of humanity are consumers of intelligence they cannot audit.
To serve both human flourishing and geopolitical rebalancing, we need a strategy built on five pillars.
Pillar 1: Treat Energy and Compute as Strategic Public Goods, Not Just Private Commodities
Gold standards failed when private actors could not provide public trust. The same is happening now.
– Nations must build National Compute Reserves — like petroleum reserves — of assured training and inference capacity for universities, hospitals, and small business.
– Grid policy must shift from “green at any cost” to “reliable gigawatts for intelligence.” A fab requires 100MW 24/7, not intermittent. A model training run cannot pause because the wind stopped. This means nuclear, geothermal, gas with CCS, and hydro must be re-classified as AI infrastructure.
– Pricing: Move from pricing kilowatt-hours to pricing reliable teraflops-hours.
Pillar 2: Decouple Intelligence from Monopoly — The Open Stack Doctrine
Human flourishing requires intelligence to be auditable and forkable.
– Mandate open weights for models above a certain capability threshold that were trained with public data, just as we mandate generic drugs.
– Fund Public AI Labs — CERN for AI — where frontier models are trained on public corpora, for public problems: materials science, protein folding, climate modeling. Private labs chase ads and chatbots. Public labs must chase the bottlenecks of flourishing.
– Prevent vertical integration of Chip + Cloud + Model + App in a single unregulable entity. We broke up Standard Oil. We must not allow a Standard Intelligence.
Pillar 3: A New Development Model — Leapfrog to Synthesis
For the Global South, the old advice was: industrialize, then digitize.
The new path is: electrify reliably, then synthesize.
Kenya does not need to build Detroit. It needs 500MW of reliable power, 10,000 H100 equivalents, and sovereign models trained onSwahili agricultural data. Then it can synthesize better seeds, logistics, and education without importing the entire industrial chain.
Geopolitical rebalancing will not come from aid. It will come from *Compute Transfer Agreements* — the equivalent of the Green Revolution, but for FLOPS.
Pillar 4: Redefine Work — From Labor as Production to Labor as Direction
If AI synthesizes, humans direct.
The education system must shift:
– From memorizing answers (which models do) to framing questions, verifying synthesis, and applying wisdom.
– Tax systems must shift from taxing labor (payroll tax) to taxing compute rents and unproductive inference (e.g., massive ad optimization) while subsidizing human-directed synthesis.
A human flourishing economy does not protect jobs that models do better. It protects and pays for the distinctly human tasks: care, judgment, stewardship, meaning-making.
Pillar 5: A Compute-Energy Non-Proliferation and Stability Pact
Just as nuclear power needed IAEA, cognitive power needs an IEA for Intelligence.
– A treaty on no first use of frontier AI for cyberattacks on energy grids and financial systems.
– Transparency on total national compute — like nuclear inspections — to avoid a blind arms race.
– Shared standards for watermarking and provenance of synthetic knowledge, so global trade in ideas does not collapse under deepfakes.
The Embedded Irony
And here is the profound irony embedded in all of this:
We spent seventy years dreaming of dematerialization. Software would eat the world. The cloud would be weightless. Intelligence would be in the ether.
The more dematerialized intelligence becomes, the more brutally material its foundations are.
To produce the most abstract thing we have ever made — artificial thought — we need the most concrete things on Earth:
– More copper than any previous industry
– More concrete for fabs and data centers than for highways
– More gigawatts than entire nations used ten years ago
– More water to cool the machines that think
We wanted to transcend energy and matter. Instead, we made energy and matter the currency of transcendence.
The second irony: We built computing to decentralize power — personal computers, the internet, permissionless networks. But cognitive computing, by its physics of scaling laws, centralizes like nothing before it. The bigger the model, the better it is. The better it is, the more users. The more users, the more capital to make it bigger. Decentralization’s tool became history’s greatest centralizer.
The third irony: Human flourishing depends on making intelligence abundant and cheap, but geopolitical stability depends on making the most powerful intelligence scarce and controlled. We must make it both abundant and scarce at the same time.
That is the paradox we must manage. If we only make it abundant, we get concentration and chaos. If we only make it scarce, we get a new feudalism where two or three sovereign stacks rule the world.
The framework above is an attempt to thread that needle — to build a compute-energy standard that is reliable like electricity, open like science, and governed like nuclear power — so that synthesis serves humans, not just capital.
Because capital learned to accumulate machines. Now machines have learned to accumulate knowledge. The question of our generation is whether knowledge will accumulate for a few, or for human flourishing itself.
Isaac Megbolugbe, Senior Advisor and Managing Principal at GIVA International. He is a recipient of Albert Nelson Marquis Lifetime Achievement Award in business and academia in the United States of America. Formerly at Fannie Mae as vice president and at PricewaterhouseCoopers as a global practice leader. He is retired professor at Johns Hopkins University and a Fellow of the Royal Institution of Chartered Surveyors. He is resident in the United States of America.
