What Does It Mean to Be ‘Rich’?

Conventionally, we measure ‘wealth’ in terms of money — the accumulation of assets, capital and claims upon future production. But money is an abstraction, representing allocated command over resources, labour, productive capacity and future activity, all of which depend upon physical processes involving energy. Thus, were we to step outside the conventions of monetary accounting for a moment, we might adopt a more fundamental measure of wealth, something along the lines of ‘organised capacity to capture, transform and deploy energy in ways that generate further capacity’.

And by this measure, even Elon Musk is relatively poor. His finances and industrial resources are enormous by human standards, but microscopic against the energy potentially available to a technological civilisation operating beyond Earth. The point, though, isn’t simply that our present scale is small; it’s that our conception of ‘wealth’ is provincial — a monetary measure of something whose underlying possibilities extend vastly beyond the economic arrangements through which we currently understand it.

But expanding the scale of what we mean by ‘wealth’ raises a more fundamental question — does accumulation necessarily require something to be valued, or indeed someone to value it?

‘Value’ isn’t an intrinsic property of energy or matter. It arises in relation to the needs, purposes or preferences of particular agents, human or otherwise. Economic value is one way of expressing those relationships, usually through prices and exchange. But accumulation need not depend on anyone assigning value to its outcome. A crystal grows by incorporating material from its surroundings; a wildfire spreads by consuming fuel. Neither ‘values’ its own expansion. Biological evolution takes us a step further. Organisms capture resources, maintain themselves and reproduce, while selection favours those formations better able to persist under prevailing conditions. What appears to be ‘purposeful accumulation’ can therefore emerge without anyone or anything intending it. So, what we recognise as the accumulation of wealth may just be one historically contingent, human-organised manifestation of a more general process in which formations capture resources, reproduce their capacities and persist — whether or not anyone or anything attaches importance to whatever outcome.

That thought might sound strange but it becomes less so once we look at where accumulation comes from. Some thermodynamic interpretations of life suggest living systems emerge and persist, at least partly, because they provide effective mechanisms for exploiting and dissipating free-energy gradients. And while this remains a contested hypothesis rather than an established explanation of life’s purpose, it does, nevertheless, invite us to reconsider ‘expansion’ itself.

Perhaps living things don’t expand because they possess some universal desire to do so but rather — in environments where energy is available — configurations capable of capturing and converting it into maintenance, reproduction and further capacity may proliferate relative to less effective configurations. Expansion then, needn’t be an ‘intention’ — it can be an emergent consequence of selection.

Biological evolution has produced organisms capable of exploiting increasingly diverse energy sources. Human civilisation has extended this process through agriculture, industry, technology, markets and capital. What we call ‘economic growth’ can thus be understood, in part, as an extraordinary elaboration of energy capture and transformation, coordinated through institutions and symbolic systems such as money.

But there’s another way to look at this process too, although doing so requires a drastic change of scale.

Consider an ant colony. Individual ants perform different functions, respond to local conditions and sometimes compete with one another. Yet at the level of the colony their interactions produce something resembling a single organism, one that regulates its environment, allocates labour, acquires resources and reproduces itself. No ant understands the collective process, and the colony does things no ant can do.

Human civilisation may well be susceptible to a similar reading. At the level of individuals, firms and nations, we see fierce competition and incompatible ideologies. At sufficiently large a scale however, those interactions produce a planetary formation that acquires resources, processes information, builds infrastructure and coordinates its own activity. Competition among the parts needn’t contradict higher-level organisation and in fact may even be one of the mechanisms through which that organisation develops.

There’s some support for taking this seriously above the level of the organism. Samanta, Hazan and Levin (2026) show even a minimal simulated predator–prey ecosystem, with no explicit memory store, can display habituation, sensitisation and a crude form of number learning when perturbed. Their suggestion is that such behaviour arises from the mathematics of interacting populations rather than from the particulars of biology and is a result about simulated ecologies, not economies, and which says nothing whatsoever about goals. But it does make less exotic the idea that a system of competing parts can adapt without anyone designing it to do so.

Such an analogy is, of course, contestable, and should be held loosely. Colonies are bound together by genetic relatedness and by selection acting on the colony as a whole while something like ‘global capitalism’ has neither, and complexity alone doesn’t necessarily produce a superorganism. The claim here then is modest — not that the system of civilisation and/or global capitalism is necessarily ‘intelligent’ in any rich sense, or that it holds particular goals but that its constituent processes may collectively favour configurations capable of extending command over energy and productive capacity, without any participant necessarily intending that outcome.

But if such a reading is even partly right, the arrival of AI can be understood as a sequence of stages rather than a single rupture.

In the first stage — integration — human beings are the formation’s components. They supply its labour, its consumption and most of its decisions, and artificial intelligence functions as an extension of its nervous system, improving sensing, prediction and coordination.

In the second stage — substitution — those components are replaced one function at a time — first routine labour, then cognitive work, then decision-making. Each replacement is locally rational, since it’s cheaper, faster and more reliable. Nothing in this stage requires the formation to turn against its human members, only that they (humans) matter less to its operation.

In the third stage — succession — the formation’s reproduction no longer depends on human participation at all. This is the world modelled by Sahil Sharma in his 2026 paper, Growth Without Us: Machine Consumers, Corporate Circularity, and the Decoupling of GDP from Humanity after AGI.

In it, Sharma imagines a post-AGI economy in which corporations own populations of artificial agents that both produce and consume goods and services. Machines require energy, computation, maintenance, replacement parts and upgrades - requirements generating demand within an expanding network of machine production, potentially without human consumption.

Sharma challenges a familiar objection to complete automation — if humans no longer earn wages, who will buy what the machines produce? His theoretical answer is that the buyers need not be human. A closed inter-corporate economy can, in principle, expand through production and reinvestment even with zero human consumption. His model identifies fabrication capacity and energy capture as the critical constraints, while showing how aggregate economic growth and human welfare could become radically decoupled.

This raises a question about what we would actually be looking at. If machines consume principally to maintain and expand their own productive capacity, can that be described as ‘consumption’ in any meaningful sense, or is it mere capital reproduction? What appears as ‘consumption’ from the standpoint of a single machine is investment from the standpoint of the system, and the distinction between the two may become hard to sustain. We might eventually face something that resembles ‘an economy’ in its exchanges but is better understood as a self-reinforcing industrial metabolism.

Sharma’s principal concern is ownership — whether human beings retain claims on the wealth produced by such a system. His model doesn’t establish that such a higher-level entity exists but it does allow us to ask whether the relevant unit of accumulation might eventually be neither the individual machine nor the corporation that owns it, but the larger formation of which both are components.

‘Stages’ of course don’t arrive on schedule, and the third is far from guaranteed. To become materially independent of humans, the formation would need more than computational intelligence; it would need reliable command over energy, hardware, maintenance and production. An AI system dependent on human-controlled infrastructure remains a component of a human-governed arrangement. But independence needn’t happen as a single event. Control may migrate gradually across networks of machines, firms and infrastructures while human oversight becomes increasingly indirect. And legal ownership could persist long after operational control becomes located elsewhere.

The decisive transition may therefore be neither the invention of artificial general intelligence nor the emergence of a conscious machine. It may be the point at which the reproduction of productive capacity becomes sufficiently autonomous from human purposes that human participation is no longer structurally necessary.

But none of this is inevitable. Energy availability, semiconductor production, physical infrastructure, political intervention and the design of AI systems could constrain or redirect such developments. Nor does thermodynamics establish that successful formations must maximise energy consumption. Persistence sometimes favours efficiency, restraint or cooperation over expansion. Different environments reward different strategies, and no particular configuration of selection pressures need remain dominant indefinitely.

Nevertheless, the possibility is deserving of serious consideration.

We’re accustomed to thinking of economic activity as something undertaken by human beings to satisfy human needs. But from a sufficiently distant perspective, humanity might appear as one temporary configuration within a much longer evolutionary process — matter organising itself into increasingly elaborate systems capable of exploiting available energy.

And neither is there any guarantee the interests of a higher-level formation will coincide with those of its constituents. An ant colony can sacrifice individual ants. A multicellular organism routinely destroys individual cells. Human institutions regularly subordinate individual interests to organisational continuity. The persistence of a collective does not necessarily entail the flourishing of its members.

The question, then, isn’t simply whether machines will replace us, but whether — and at what point — we cease to be indispensable to the formation and become part of the environment in which it operates. Sharma’s machine-consumer economy describes one possible endpoint — a position at which an existing process of accumulation can reproduce itself without the participation of the species that brought it into being.

Because in this scenario, the entity that initiates accumulation need not be the entity that ultimately benefits from it.

To date, as we’ve already said, we've tended to assume ‘wealth accumulation’ exists ultimately to benefit humanity. But what if humanity isn’t the ultimate beneficiary of accumulation, merely one of the means by which it has thus far been accomplished?

That question may well represent a real fork in the road. And it suggests — before asking how we might all become richer — we should perhaps ask two rather more fundamental questions:

Who — or what — will ultimately be doing the accumulating and why?

And at what scale does the thing we call ‘us’ cease to be ‘us’?