In June 2026 the Economist newspaper ran a piece reflecting on a new proposition circulating in tech circles — should tokens, or compute, become a form of money in their own right. Sam Altman, on the Lex Fridman podcast, argued that compute will be “the currency of the future” and that it should be used as a form of UBI. The ‘direct’ payment for things in GPU-hours is presented as an upgrade to fiat currency. More solid, more ‘real’. Perhaps Nigel Farage would call it ‘proper’.
In the Dogecoin world, the tangibility of a token sounds like a scarily reasonable argument. But it isn’t. I want to explain why it isn’t, and why the replacement of money with tokens would be a direct win for organised capital and rentiers over productive work.
Let us start with what money is, or pretends to be. School teaches us 3 functions: a medium of exchange, a store of value and a unit of account. Okay, but that list tells you what money does, not why it exists. So the question remains: a medium to exchange what? The answer in my textbook is goods. If you really want beans from the guy down the street, money solves the problem of trading chickens for shoes. Nice story, partially true, but misleading too. It lacks a crucial component. Anthropologists who have searched for barter economies have not found any, actually. Keynes knew this and that is why he referenced the island of Yap and its amazing Rai stones which were tokens of wealth in his 1930 Treatise on Money (Keynes got confused and referenced Rossel Island but I guess he didn’t have a fact checker). What anthropologists and historians find are systems of obligation — tallies, tabs, debts paid and repaid over time within a community that knows each other, trusts each other, and keeps count.
Of course, what people were really keeping track of within a community was not goods sitting in a barn. David Graeber convincingly argues it was debt. I like that argument but I’m not so sure in the way we think of “debt” today; i.e. a monetary contract to be settled at a certain price, in which many people are drowning. The debt, in pre-capitalist society, was much more to do with mutual social obligations, which is what they meant too. Dowries that thank parents for raising their daughter (still existing in many societies, still misunderstood by those outside them as “bride price”). A vow to bear arms after someone stood up to a bully for you. Help with the harvest or a cup of flour to borrow. And as societies changed, and not everyone knew their neighbour as kin exactly, money became more about the thing we are all born with, and which we could bring to bear for our neighbours: our labour. A day thatching a neighbour’s roof, a season’s help at harvest, a debt of work carried forward until it could be paid back in kind. Money, when it finally crystallises out of these systems, is what allows that reckoning to continue in our society of friendly strangers. It allows a tailor and a farmer and a mason to square up without trusting each other on a personal level, or determining how many shirts are worth a sack of grain. It is a common denominator for labour that cannot otherwise be directly compared, because an hour of tailoring is not the same thing as an hour of farming, and yet both cost the person doing it a single portion of their wild, beautiful and finite life.
And that’s the real purpose behind the textbook definitions, in my opinion (sorry!). Money is a device for making different kinds of human time comparable. All the rest: the coins, the notes, the entries in the ledger — that’s just scaffolding on the foundation of that one simple fact. Ah, but capital, you will say. We’ll get to that.
Adam Smith understood the central importance of labour behind value and money. The first line of The Wealth of Nations makes it clear that a nation’s wealth is based on its annual labour: "The annual labour of every nation is the fund which originally supplies it with all the necessaries and conveniences of life which it annually consumes." For Smith, wealth was never the pile of gold in the treasury. It was the population’s productive activity, year after year. All of this ties back to fiat money, warts and all. When a state issues currency and asks its citizens to pay tax, what the note is actually backed by isn’t the paper or some store of gold in the BIS in Basel, it’s the fact the economy behind it keeps producing things worth having, made by people who get up and work and pay tax. But the credibility of its money, however indirect and abstracted the connection has become, stems from the claim the state has on the fruits of that continuing labour. Obviously, none of these definitions are mutually exclusive. Trust (and obligation) is the social force that makes exchange possible; labour, once you get past small-scale kinship reckoning and into a market economy trading between strangers, is mainly what’s being counted, because that’s what every party to an exchange actually gives up to produce anything worth trading.
Gold is still considered by many to be the original money. The classical political economists were grappling with the relationship between gold and money and value. Gold had an arguably weirdly direct and also adjacent connection with labour. In this respect Ricardo and Marx (see below) directly tied the gold price and hence the value of the currency to the amount of labour necessary to extract the gold from the ground. That is, it was then connected to the economy’s productive capacity.
This, I believe, makes fiat currency more akin to a true claim on a nation’s productive capacity, and more in alignment with the views of the classical political economists, than something tied to a “proper” thing — be it compute or shiny gold metal. Adam Smith was not against paper money, and apparently passively accepted gold as a backing. But we should note the way in which he phrased this. He wrote that money is ‘the great wheel of circulation’, and that ‘the great wheel of circulation is altogether different from the goods which are circulated by means of it. The revenue of the society consists altogether in those goods, and not in the wheel which circulates them.’ Money circulates wealth; it does not produce it. He develops what is sometimes called his labour-command theory of value, according to which the real value of a sum of money is not its face value but the amount of other people’s labour it can command in exchange. Earlier in the same book he states it as plainly as he ever states anything: "Labour was the first price, the original purchase-money that was paid for all things." Of course he believed that a well-run banking system, substituting paper for gold and silver, turned a country’s "dead stock" of idle bullion into "active and productive stock," releasing real capital to be put to work. But he was clear that this is only as long as the paper remains tied to something concrete, and as long as banks distribute it prudently. Push him too far and you get his famous warning: an economy resting on unanchored paper is “suspended upon the Daedalian wings of paper money” rather than travelling on “the solid ground of gold and silver” — Icarus’s wings, beautiful right up until they aren’t.
David Ricardo went further than Smith. In the Bullionist Controversy, when the Bank of England was issuing paper that was no longer convertible into gold on demand, and this was thought by some to be causing inflation, Ricardo wrote that gold barely needed to touch the currency at all. His idea was notes redeemable in big gold bullion bars instead of coin, taking actual gold almost entirely out of everyday circulation. As he himself said, "it is not necessary that paper money should be payable in specie to secure its value; it is only necessary that its quantity should be regulated according to the value of the metal which is declared to be the standard." Thus, gold provides the standard to which the quantity of currency should be disciplined — it need not be the currency. That’s a very proto-fiat position, worked out two centuries before central banks stopped pretending otherwise.
Quesnay and the French Physiocrats had a harsher view, almost comic today but with a serious principle. Only agriculture produced a real surplus, the produit net, for them, whereas manufacturing, trade, and thus money and finance were "sterile," machinery making machinery. The principle they were getting at is that there is value only if the thing is useful to humans directly. The notion of sterile manufacture might be hard for us to swallow. Then they were thinking of food (and perhaps of clothing) as the useful products of the land. But used logically it would be a sharp tool to turn on the financialisation of many economies since the 1980s.
Marx continues the line of thought started by Ricardo. Gold becomes money, he argued, because it is the universal measure of value, “the equivalent commodity par excellence.” The reason why one commodity can take the place of the value of all the others is that its value is produced in the same way as theirs: through labour. Iron, wheat, coffee and gold are all, in his phrase, ‘materialisations of uniform labour’. They differ only in the amount of labour-time embodied in each. Money is what enables that mutual labour content to be compared and settled between strangers.
But Marx’s deeper concern is what happens when finance gives the illusion that money can make money directly. Returns on capital with no productive step in between. He thought that was always a delusion. Profit has to come from somewhere in real production and finance just takes a share of a surplus produced elsewhere. Of course, reading Marx’s words today can seem archaic. We need to bring this into the twenty-first century and not limit that definition to traditional wage-labour but include active entrepreneurs and those putting their own earned capital at risk in a new venture. This is active creation, not passive income like rent on land you did nothing to improve, or just sitting around waiting for interest to grow on your cash.
Underlying value seemed to fall by the wayside of most economic thought about 150 years ago. In the 1870s, the marginal revolution, led by economists such as Jevons, Menger and Walras, replaced the classical labour theory of value with the concept of subjective utility at the margin. To them, the value of a good is whatever the next buyer is willing to pay for it at the moment of consumption, regardless of the labour that went into it. This shift gave us a useful descriptive tool for short-term pricing, enabled by the possibility of more complicated calculations with calculus. But it carried an analytic cost in bypassing questions of the relationship between value and price, and severing the link between markets and human production. In this tradition, economists regarded money as "neutral" — a medium of exchange, but distinct from production.
This premise was inherited by modern monetary economics. The Quantity Theory makes the value of money a function solely of its supply in relation to the goods it chases. Central banking takes it as a matter of managing market expectations and credibility. The endogenous-money theorists point out that modern money is created as bank credit as soon as a loan is issued, and that there is no necessary relation to labour or output at all. This is not necessarily anti-labour and anti-productive activity. But it does look like it’s stopped considering the question interesting. The tech figures who advocate compute as money are not departing from mainstream economic theory, but taking the logical next step in the disconnect between money and productive activity.
As a side note, the cryptocurrency scene, until 2026, seemed to have a lot of fans that thought they were going to take over from fiat currencies. I am not being glib, I would say they have no connection with the basic three functions described earlier. More importantly, they have nothing to do with production or labour (although there is something nostalgically sweet, but also horribly expensive about bitcoin’s vast mining operations). Crypto has proved popular for a variety of other reasons including privacy, efficiency and other tangential factors including criminal activity and asset accumulation. But, generally, the value of a crypto is based on scarcity built into the protocol, and the expectation that tomorrow more people will want to come in than want to get out today. Early holders win as later holders buy; the whole structure depends on continuous new demand, not continuous new production. That is the structure of a Ponzi scheme, even if the people running it do not think they are running one. I do get that a lot of people in crypto really believe in what they are building, and the underlining cryptographic techniques are really clever. But cleverness in the plumbing does not alter what is in the pipes. It remains a currency that has value only because of the demand of perpetual new buyers, not the production of any actual value.
I added the crypto aside to make the point that compute-as-currency is not like other Ponzi-like money substitutes. But it is arguably a more sinister proposal because it does not even pretend to elegantly solve the scarcity problem. It does come in bite-sized units which have an underlying value and which are recognisably uniform. But trace a unit of compute back to its source and you won’t find an hour of anyone’s labour. You find a data centre sitting on land owned by someone, drawing power from a grid controlled by someone, running on chips fabricated in a small handful of plants owned by a small handful of companies. If you think fiat is not a hard currency, then name a money that is harder than fiat. You’ve given monetary authority to the people who already own the land, the energy, and the silicon — which today is a very short list of companies and states. Everyone else would be earning tokens just to spend them to rent access to machines they will never own, working to service infrastructure rather than trading their labour with other people’s labour. That’s the old feudal relationship in a new dress. The lord had not to buy the peasant’s grain at a fair market price; he owned the land on which the grain grew, and rent followed automatically from that fact. A compute-backed currency does the same to digital labour. It doesn’t have to own you through a transaction — the ownership is built into what the currency is made of before a transaction occurs. That might sound a little over the top on the feudal argument. But then again it might not. That is what Yanis Varoufakis argues very cogently. And on this one I am with him.
I see Ping An Bank has a credit card with compute as the rewards instead of air miles. Tokens of compute seem like the right space for this kind of air-mile or supermarket-discount reward. It recognises the rewards are tied to a monopoly issuer, which may give you some benefit on something you want but doesn’t hand over the entire economic system to the provider of those rewards.
Money, whatever form it takes, is worth defending only insofar as it stays anchored to what people actually do with their time. Fiat money has a weak, imperfect hold on that anchor in the form of the state’s reliance on a productive population. There might be better ways of money. Let’s see. But not token currencies with compute backing. They cut the anchor line altogether, and tether value instead to the ownership of fixed physical assets that most of us will never hold a share of. That’s not a technical improvement of currency. It’s a proposal to empower landlords, energy owners and hardware monopolists, dressed up in the language of decentralisation.