Elon Musk’s Grandfather Was Arrested for an “Energy Currency.” Now Musk Is Talking About One.

Musk once said that money might one day be measured in kilowatts, mass, and energy. Sitting across from him when he said it was Isaacson, who wrote his biography. In that biography, Isaacson mentions an episode involving Musk’s maternal grandfather, Joshua Haldeman. In the 1930s and 1940s, Haldeman was active in a Canadian organization called Technocracy Incorporated. The movement arose during the Great Depression. It argued that the price system had failed, that it should be abolished, and that social production and distribution should be accounted for in units of energy. Its members designed “energy certificates,” issued to each person according to the total energy available to society. The certificates were to be used up and then void. They could not be accumulated or transferred. The organization was at one point outlawed in Canada, and Haldeman was arrested because of it.

This lineage is interesting, but it does not need to be written as Musk inheriting his grandfather’s cause. A more accurate way to put it is that what a person hears at the dinner table as a child leaves traces in how he thinks decades later. The idea that energy is hard currency appeared in his family, and it failed when people tried to implement it.

The Technocracy movement did not start from a foolish premise. During the Great Depression, factories stood idle, goods piled up, workers were unemployed, and in the short run the price system could not connect productive capacity with demand. The movement proposed that since prices could not allocate resources effectively, society should simply account in energy units. In engineering terms, the scheme looked tough-minded. Energy is conserved. Joules are measurable. The ledger is clear. It seemed more rational than a monetary system full of speculation and fluctuation.

Economists have made several objections. First, energy cannot price demand. The same thousand joules used to bake bread and used to power a hospital show up on the ledger as 1000 J, but no one would think these two uses should receive the same resources. An energy ledger records inputs. It cannot see what people want, or what they will give up for one use over another. Second, Hayek and others argued that the function of prices is not bookkeeping. Prices transmit information dispersed in each person’s head. What a person is willing to pay reflects a combined judgment about his situation, substitutes, time preference, and opportunity cost. That information sits with countless individuals. No central energy ledger can replace it. Third, production at the time depended on people. A factory opening meant the labor time of thousands of workers, and the value of labor cannot be measured in joules. As long as labor remains a major factor of production, accounting in energy must miss the most important piece.

So the Technocracy movement failed on economic calculation, not just politics. It tried to replace prices with engineering units, but it could not replace the subjective valuation and information mechanism behind prices.

In this interview, Musk’s chain of reasoning runs like this. Robots and AI will take over all labor. Space development will multiply the scale of the economy. Producing goods and services will become a trivial matter. Set against the two reasons for failure above, his argument goes further.

If machines replace labor, then the cost of producing a thing comes down to raw materials and electricity. The robots consume electricity. The factories that build the robots consume electricity. The equipment that extracts raw materials consumes electricity. Break it down layer by layer, and what remains on the cost sheet is materials and electricity. The labor piece that was missing ninety years ago no longer exists in the future he imagines. This reasoning holds inside a closed engineering system. When labor costs approach zero, the hardest items on the cost sheet are materials, equipment depreciation, and energy.

His saying “mass and power,” not just “energy,” reflects a cross-planetary setting. In spacecraft design, the two most important spreadsheets are a mass budget and a power budget. Every gram of payload and every watt of power has to be allocated among teams, and they compete for it. At this stage money recedes. Once a project is approved, what constrains you is how many kilograms the rocket can lift and how many watts the solar panels can generate. What Musk manages every day is cost per kilogram to launch and Starship capacity. In his world, mass and power are units of account to begin with.

Mars is a different case. Suppose that in a few decades there is a settlement of a few thousand people on Mars. What use would dollars be in the hands of those residents? Goods from Earth have to wait for launch windows, roughly once every 26 months. How many tons can be shipped each time is fixed. What the settlement can do depends on how many kilowatts the reactors and photovoltaic arrays can provide in total. In that environment, how many tons of Earth supplies and how many kilowatts of power you are allocated is a person’s wealth, and almost the only wealth. This example is more persuasive than the Earth example because it is a closed system. Launch capacity is limited. Energy is limited. The resupply cycle is extremely long. Mass and power are hard constraints.

So the line works better as a transplant of SpaceX’s internal engineering language onto the scale of civilization than as a prophecy about the whole economy. Musk is not writing monetary theory. He is imagining the future with the budget spreadsheets he knows best.

Still, the prophecy is not pure fantasy. A couple of years ago I had dinner with a friend who works in data centers. He said that when talking to big customers now, the first question is not “What’s your budget?” It is “How many megawatts of power can you get?” After the phase when you could not buy GPUs even with money, the next phase is being able to buy the cards but not connect them to the grid. Now when leading AI companies announce new clusters, they directly cite gigawatts. When Musk’s own xAI built Colossus in Memphis, the most closely watched issue was also power supply. Generators, storage, and transformers each made the news. He himself has publicly said that the bottleneck for AI will shift from chips to power and transformers.

In some industries, money has already begun to lose its property of buying everything. You have money but cannot exchange it for power. At that point megawatts are the constraint. In the computing-power industry, performance per watt and tokens per watt are becoming more important metrics than performance per dollar. This trend will spread to robotics and manufacturing. If humanoid robots enter factories at scale someday, factory owners will care about how many kilowatt-hours each robot consumes per hour and how many units it produces. The labor-cost line will matter less and less.

Still, as for the endgame of this prophecy, I remain skeptical.

Money is money because it is scarce. Musk says on the one hand that production will become trivial, and on the other that energy will be the monetary unit of the future. Put together, these two statements fight each other. If energy becomes extremely abundant, energy will not be scarce, and something that is not scarce cannot serve as money. If it is scarce enough to serve as money, production will not be trivial. What can remain scarce in the end is land, location, attention, reputation, and things that are unique. None of these can be measured in joules. The price of a sea-view apartment has basically nothing to do with how many kilowatt-hours were used to build it. What is scarce is not concrete and glass but that sea, that location, that irreproducible view and property right. An energy ledger cannot express this kind of subjective valuation.

Accounting in energy still does not solve the information problem from ninety years ago. AI may be able to predict demand better than humans, but what it predicts is human preference, and preference is inherently subjective. As long as people still make choices because “I just like it,” there will need to be some tool for expressing subjective valuation, whether or not it is called money. The function of prices is not only to allocate energy. It is to let dispersed preferences, opportunity costs, and local knowledge appear in the form of compressed signals. An energy unit can tell you how much was spent. It cannot tell you whether it was worth it.

The Mars example has a similar problem. In the early days of a settlement, tonnage and kilowatts are hard constraints on wealth. But even there, oxygen, water, spare parts, maintenance capacity, social institutions, and relations of trust also matter. Mass and power are budget language, not a complete monetary theory. They explain what is blocking production, not why people are willing to live this way.

Coming from a man with enormous wealth, this is especially easy to question. Earlier this year, Nobel laureate in economics Acemoglu challenged him on X. Since you think money will not matter by 2036, why not commit to giving away all your wealth before then? Musk responded that he does indeed plan to do something like that. Later there was a promise to give SpaceX stock to every American. This challenge is rhetorical, not an economic proof. “If money doesn’t matter, why not give it away” catches an inconsistency between normative commitment and behavior. It cannot directly refute the judgment that energy is the hard constraint of the future. But it exposes one point. If a person still retains control over wealth, it shows that money still matters in his own decisions. At least during the transition, money remains a tool for mobilizing resources, organizing production, and allocating risk. Between prophesying the endgame and living in the present, there is a large problem of institutional continuity.

Compress this intellectual history into one sentence. Musk has transplanted SpaceX’s internal engineering language onto the scale of civilization. In a closed system it has explanatory power. In an open society it cannot replace prices and subjective preference.

The Technocracy movement failed ninety years ago because it tried to use energy units to replace the information function and the valuation function of prices, not because energy is unimportant. Musk today goes one step further than they did. He imagines labor replaced by machines, production becoming trivial, and the space economy expanding in scale. That removes labor from the cost sheet and pushes raw materials and electricity to the foreground. In closed or semi-closed systems such as spacecraft, Mars settlements, and AI data centers, this language works. Mass budgets, power budgets, megawatt interconnection, and tokens per watt are hard constraints.

But once we return to an open society, the problem returns. If energy is abundant enough not to be scarce, it is unsuitable as money. If energy is scarce enough to serve as money, production will not be trivial. What is scarce may be land, location, attention, reputation, and unique things, and these cannot be measured in joules. AI can predict demand better, but it cannot cancel the subjectivity of preference. As long as people still make choices because “I just like it,” there will need to be some tool for expressing subjective valuation.

So what deserves attention in Musk’s statement is not the endgame prophecy that money will become kilowatts. It is the medium-term trend it reveals. In computing power, robotics, manufacturing, and energy-intensive industries, money is increasingly not the only hard constraint. Power, transformers, grid interconnection, and performance per watt are becoming the bottlenecks. The engineer’s budget spreadsheet is entering economic discussion, and that is worth noticing. But a budget spreadsheet is not monetary theory. The unit of account of a closed system is not the price system of an open society. The information problem from ninety years ago will not disappear just because robots can turn screws.