by Arjun Jayadev and J. W. Mason
Last year marked a watershed: For the first time in history, growth in renewable generation exceeded global growth in electricity output, causing fossil fuel power generation to decline for the first time on record. But 2025 also saw the beginning of a historic surge in energy demand from AI. Just four companies — Amazon, Google, Microsoft, and FaceBook – are expected to spend over $1.5 trillion, one quarter of all U.S. fixed investment, on new datacenters in the coming year.
These twin shifts, and the balance between them, will be central to the future of both carbon emissions and energy prices. But they also tell us something fundamental about how the economy works. Because while both these investment booms take place through markets, in the sense that their inputs are purchased and their output are sold, neither is a response to market demand. These two engines of global growth are being directed by something very different from what we traditionally think of as market signals.
The reaction from economists and business commentators to the surge of investment into these sectors has been a mix of exhilaration and fear. On the one hand, these investments promise genuinely revolutionary change, but on the other they violate deep-seated ideas of how the economy is supposed to work.
The expansion of solar-panel production has pushed far past the limits of profitability. Yet despite a drumbeat of warnings of overcapacity and collapsing prices, production keeps growing.
AI investment is, if anything, even more unmoored from profitability. Unit costs for computation have been estimated to be as much as eight times greater than what is being charged for them. Add to this the immense capital expenditures required, and it is far from clear that there is any viable business model for the AI services currently flooding the market. Anthropic’s recent decision to throttle back its less-expensive subscription tiers for Claude code users was, arguably, a recognition of this reality.
Economics teaches us that markets guide resources to their most profitable uses. Resources are allocated based on the relationship between prices and costs. Where the sales price of something is above the cost of production, businesses will add capacity to produce more of it; where the price is below production costs, businesses will scale back or exit. But in both green energy and AI, as in many historical investment booms, that is not what is happening. The resources move first, and the profits come later – or perhaps not at all.
Some experts fear we are fearing another bout of irrational exuberance — a misallocation of resources on a vast scale, this time colored with an appealing green or silicon hue. But from our point of view, the most important question about these investment booms is not whether they are irrational. It is what they reveal about how capitalism has always worked. Decisions about what industries will be built up and which will be abandoned are guided not by the invisible hand of the market, but by the hidden planners of finance.
True, the US AI boom is largely private, while the Chinese solar boom is driven by the country’s public sector. But this is less of a difference than it first appears. Investment in both cases reflects conscious decisions of a small group of people—venture capitalists in one case, government officials in the other. Indeed, given the Chinese model of devolved industrial policy with intense competition between local governments, decisions about investment may be more centralized in the US version.
When investors pour funds into ventures that promise profits years or decades from now, they are not responding to the market. They are making a conscious choice – or a plan – to reorganize the economy. The AI and green-energy industries illustrate this dynamic in especially dramatic form. But any business that operates at a loss (as almost all do in their early years) is doing so in defiance of market signals. How long a business can operate at a loss, and how high profits must be to justify expanding or even remaining in existence, are fundamentally questions about financing.
Finance exists precisely to allow production to depart from market signals. Money organized through finance is a bet on, and a catalyst for, a particular vision of the future.
The extravagant promises and, often, utopian visions, that accompany great investment booms function as coordination mechanisms. They align expectations, justify losses, and stabilize beliefs so that resources can be committed to loss-making industries until they become profitable.
In the language of John Maynard Keynes, such moments are driven less by calculable returns than by “animal spirits”—the fragile, shared confidence that induces investors to act in the face of a fundamentally unknowable future. “Enterprise,” as he wrote, “only pretends to itself to be mainly actuated by the statements in its own prospectus… Only a little more than an expedition to the South Pole, is it based on an exact calculation of benefits to come.” The combination of animal spirits and organized finance is what allows investment to move in advance of profit.
In this environment, money does not operate as a neutral means of exchange. In great waves of investment like we are seeing today, organized money is the instrument through which production is redirected, futures are selected, and value is constructed. All of this happens not in response to the market, but in defiance of it.
Among leading economists, Joseph Schumpeter was among the few to focus on the transformative role of finance. Credit’s essential role in innovation, in his view, has more in common with central planning than with traditional markets. Loans to entrepreneurs, he wrote, are “what corresponds in capitalist society to the order issued by the central bureau in the socialist state.” A loan is, in effect, an order saying: This business has the authority to take whatever labour and resources its project requires up to some certain amount. The hundreds of billions flowing into AI compute and solar manufacturing are precisely the result of such orders. Far from passive reactions to known future profits, they are reshaping the terrain on which future profitability will be assessed.
Finance is planning. To organize money is to select some futures and foreclose others, permitting certain transformations while making others impossible. In green technology, that planning reflects the priorities of the Chinese state, which has channeled both public and private investment into solar manufacturing, battery technology, and electric vehicles. It harnesses markets to do so, but the shift is taking place at a scale and speed that markets alone could never achieve. A similar story is unfolding in AI: The investment boom reflects the convictions of a small number of tech CEOs and venture capital principals. Their commitment to AI reflects their vision of the long-term future of humanity as much as it does expectations of profit.
Once we clear away the idea of monetary neutrality and the role of markets, we can see finance for what it is – what science fiction writer Kim Stanley Robinson called a “Ministry of the Future.” Markets are not an alternative to planning, but the medium through which planning takes place. Which leaves us with the real question: Are the futures being planned for us the ones that we want?
*
Arjun and I wrote this op-ed back in May, when Against Money was published. We didn’t manage to do anything with it then, but it feels just as relevant now.
Also, a note: I am going to try to start posting more regularly on this blog. For the rest of 2026, I am going to aim for one post per week. There’s nothing in particular that you need to do with this information, I am just putting it down here as a marker.