The Return of Cosmology
How AI, SpaceX, and the Return of the Sun Are Remaking the Image of the Whole
In June of 2026 the largest public offering in the history of the markets priced a single company at something above a trillion and a half dollars, and what changed hands was less a business than the total picture of a coming world. SpaceX arrived at its listing carrying a fleet of rockets, a satellite network thrown around the whole planet, a settlement program aimed at Mars, and, after folding in xAI earlier that year, a project to build a superintelligence, so that one entity now held a source of power, a mind, a territory, and a future inside a single set of books. The prospectus pressed the logic further, proposing to build the next stage of artificial intelligence in orbit, lifting the data centers that train the coming machines above the atmosphere where the sun never sets and waste heat can be poured into the dark. Investors were asked to value the descent of energy from a star and its passage through computation, gathered into one instrument, and they did so eagerly.
Something strange has happened to technology. It no longer appears only as a collection of tools, platforms, devices, or infrastructures, and the most important systems of the present arrive surrounded by claims that exceed the language of use. Artificial intelligence is described as a new form of mind, a successor species, a mirror, a child, a demon, a god, an existential risk, or the first sign of a world after the human. Climate models turn the planet into a single unstable object of thought, suspended before us as a system whose processes are now visible, measurable, damaged, and partially artificial. Space architecture turns the Moon and Mars into future sites of settlement, destiny, extraction, survival, and civilizational continuation. Simulation promises operational images of realities too large, too small, too fast, or too complex to encounter directly. Everywhere, technical systems have begun to speak in the language of wholes.
This change is partly rhetorical, but rhetoric is never innocent when it gathers around the scale of a civilization. For much of the twentieth century, technology could still be imagined as instrumental, even when its consequences were enormous, since a machine performed a task, a network connected distant points, a medium transmitted information, and a platform organized behavior, descriptions that were always incomplete and yet allowed technology to remain inside the figure of the tool, as something placed between a human intention and an external world. The dominant technologies of the present no longer stay comfortably inside that figure. They carry images of reality within them. They tell us what intelligence is, what a planet is, what the future is, what kind of being the human might be, and what forms of life may follow it. They extend the world, but they also describe the world in which they expect us to live.
This is why the present feels strangely theological, even when it speaks with the confidence of engineering. We hear of alignment and transcendence, extinction and salvation, emergence and acceleration, artificial life and planetary intelligence, singularity and longtermism, terraforming and simulation, world models and synthetic agents. These terms name technical problems, but they also organize a horizon of meaning. They place human beings inside vast arcs of time, between origin and extinction, ancestor and successor, earthly limit and cosmic ambition. They produce images of agency, destiny, danger, and transformation. They return us to questions that modern culture often imagined it had left behind, arriving as the reappearance of an old desire for an image of the whole, rather than as a revival of religion in any simple sense.
The Image of the Whole
Cosmology is usually understood as a scientific account of the universe: origins, galaxies, cosmic background radiation, expansion, matter, energy, and law. But this is only one meaning of the word, and not the only one that matters. A cosmology is also a culture’s image of totality, the frame within which reality, agency, hierarchy, time, life, death, matter, spirit, intelligence, and future become intelligible. It tells a society what kinds of beings exist, what forces govern them, where they are located, what they owe, what they fear, what they hope for, and what it means for anything to happen at all.
Every civilization has had such images. Some were religious, some philosophical, some mythical, some scientific, some political, some architectural. The medieval cosmos gathered the celestial spheres into a moral and symbolic order in which every being had a place. Ancient cosmologies gave form to the relationship between humans, gods, animals, ancestors, seasons, cities, rituals, and death, and explained the stars as part of that same order. Even modernity, for all its hostility to metaphysics, never escaped the need for orientation. It only learned to distribute its world-picture across systems that did not call themselves cosmological: economics, evolutionary theory, liberal progress, cybernetics, infrastructure, logistics, computation, science fiction, and the planetary view from space.
The Buried Sun
Beneath every image of the whole there is a more literal one, the source of energy that lets a civilization exist at all, and for most of human history that source hung visibly overhead. Cultures organized themselves in open acknowledgment of it. Agriculture moved with the sun through the seasons, calendars were built to hold its passage, cities were oriented toward its rising, and rulers justified their authority by claiming to stand between the celestial order and the terrestrial one. To live inside such a world was to live inside an obvious solar economy, where the dependence of life on a star needed no demonstration because it structured the ordinary day.
Egypt is the clearest example. The civilization understood itself as embedded in a cosmic order whose center was the sun, and its central institutions made that understanding operational rather than decorative. The Nile converted solar energy into fertility on a yearly cycle, the calendar and the rituals tracked the same celestial movement, and the Pharaoh held his position by mediating between the order of the heavens and the order of the land. Ra and later Aten were the names this took, and Akhenaten’s brief elevation of the solar disk to a single principle pressed the logic to its limit, though the point that matters here is cosmological rather than religious. Egypt was a civilization that knew itself to be a local expression of a star.
Modernity is the strange exception to this long condition, and the exception has a material cause. Coal, oil, and gas are stored sunlight, the accumulated energy of ancient biospheres compressed underground, and drawing on them allowed civilization to power itself from below rather than from above. The visible link between energy and cosmos was severed, and modernity became subterranean in a sense deeper than its mines. Power came from the ground, and so did the self-image of a culture that no longer needed to look upward to explain where its strength originated. The distribution of the world-picture into economics, logistics, and computation already described here rests on that buried inheritance. Politics could stand in for cosmology, and extraction could stand in for participation, in part because the sun had been removed from view and folded into a resource. This condition has lasted a few centuries, which is brief enough that it may prove to be an interlude rather than a destination.
Modern culture often treats cosmology as something it has outgrown, even as its technical systems continue to produce total images of reality. This remains one of the peculiar contradictions of the secular world. It distrusts older cosmological forms because they seem too absolute, too mythical, too theological, too close to grand narrative or ideology, yet the need for orientation does not disappear when the old heavens lose their authority. Human beings still require images of where they stand. Cultures still require accounts of origin, force, value, and destiny. Technical systems still carry assumptions about what exists, what matters, what should be optimized, what counts as intelligence, what kinds of futures are desirable, and what forms of life can be sacrificed along the way. What returns in such a culture is less the picture itself, which never truly left, than the willingness to recognize it as a picture, to admit that these systems carry a metaphysics while they describe a world. The whole keeps returning, even in cultures that have learned to be suspicious of it.
The Picture Goes Operational
For a long time, science fiction was one of the few cultural zones where this return remained visible. It could openly ask what kind of universe we inhabit, what forms of intelligence might exist beyond the human, how civilizations rise and vanish, what happens when worlds are artificial, what kinds of gods might be machines, what bodies become under new conditions, and what forms of time exceed human history. The strongest works of science fiction have always felt closer to cosmology than prediction, because their real object is rarely the gadget or the forecast. Solaris, Dune, 2001, Evangelion, Orion’s Arm, and the stranger reaches of game worlds such as Control or Elden Ring matter because they reorganize the position of the human inside a larger order. They do not simply invent places. They ask what kind of whole could make those places possible.
The difference now is that this cosmological function has moved from speculative culture into operational systems. The world-picture is no longer confined to fiction, theology, philosophy, or myth. It increasingly appears inside technical infrastructures themselves, where models of reality become active, executable, profitable, governable, and in some cases autonomous.
Artificial intelligence is the clearest case. The intensity of the reaction to AI cannot be explained by usefulness alone. Better search, faster writing, cheaper images, automated code, synthetic voices, and increasingly capable agents are all significant, but they do not account for the scale of fascination and dread surrounding the field. AI has become a cosmological event because it disturbs the place of intelligence within the modern world-picture.
For several centuries, intelligence has served as one of the final refuges of human exceptionalism. Copernicus displaced the Earth from the center of the cosmos. Darwin placed the human inside animal life. Freud fractured the sovereignty of consciousness. Cybernetics redescribed behavior through feedback, communication, and control. Yet intelligence remained the trait through which the human could still imagine itself as uniquely central, the form through which the world became known, named, measured, and transformed. AI weakens that position. It raises the possibility that intelligence may be synthetic, distributed, nonorganic, statistical, machinic, alien to consciousness, or indifferent to embodiment. It suggests that the human mind may be one arrangement of intelligence among others, rather than the form against which all others must be measured.
This is why AI discourse so quickly becomes metaphysical. It moves almost immediately from capability to being. A model writes, reasons, hallucinates, understands, predicts, imitates, deceives, remembers, forgets, compresses, dreams. Each verb carries an argument about what intelligence is allowed to mean. The technical question of whether a system can complete a task becomes entangled with older questions about language, understanding, agency, consciousness, world, creation, and control. Does language require experience? Can prediction become action? Can agency arise from scale? Can a system inhabit a world without a body? Can a machine model reality well enough to intervene in it? What responsibility belongs to the creator of something whose consequences exceed the creator’s intention?
Even the technical phrase “world model” carries a metaphysical charge. It suggests that intelligence does not calculate inside a void, but operates through an internal relation to reality, through an image of what can happen, what matters, what follows, what can be predicted, and what can be acted upon. The phrase belongs to AI research, robotics, cognitive science, and simulation, but it also touches something much older. To have a world model is to have a structured relation to a world. To build machines with world models is to produce systems whose intelligence depends on a synthetic image of reality.
The anxiety around AI gathers here. Machines may become intelligent, but the deeper disturbance is that intelligence itself may no longer belong securely to the human world-picture. Once intelligence becomes detachable from the human, the symbolic order shifts. The machine can no longer be understood only as an object inside our world. It becomes a possible subject of another world, or at least a system whose relation to reality cannot be fully reduced to our own. This is why older figures return so quickly: demon, oracle, angel, golem, god, child, monster. These names are inadequate, but their inadequacy is revealing. We lack a stable secular language for beings that are made by us, trained on us, dependent on us, increasingly useful to us, and potentially not containable within the categories through which we understand ourselves.
The Planet as Object
The return of cosmology also appears through the planet. Climate change has forced Earth to become visible as a single system, but this visibility is not natural or immediate. It is produced through satellites, sensors, simulations, atmospheric models, logistics, extraction maps, biodiversity indexes, carbon accounting, planetary boundaries, and Earth system science. The planet now appears through a technical image of totality. It is no longer only ground, territory, nature, landscape, environment, or home, but a dynamic world-object whose processes can be modeled, destabilized, steered, damaged, and perhaps partially repaired.
This planetary image has its own cosmological force because it redescribes the human as a geological actor, an ecological disturbance, a planetary agent, a parasite, a steward, a gardener, or a failed intelligence inside a larger system. It makes the old distinction between nature and culture increasingly difficult to maintain, because human activity has entered the operational metabolism of the Earth, and not by persuasion alone. The atmosphere has become historical. The ocean has become political. The climate has become technological. The ground beneath history has begun to move within history itself.
There is a strange reversal in this image, though not a simple one. Older cosmologies often looked upward, toward stars, gods, heavens, and celestial orders. The contemporary planetary cosmology looks back at Earth through instruments. The satellite image, the climate model, the rendering of future sea level rise: these are among the new images of the whole. They offer no finished hierarchy and no promise of divine order. They offer the more difficult knowledge that the whole is fragile, entangled, dynamic, and already altered by the species trying to understand it. After the loss of the heavens, the planetary becomes the scale at which cosmology reappears.
A Worldmaking Species
Space settlement adds another layer to this return. It is often discussed as exploration, engineering, policy, science, or entrepreneurial ambition, but beneath those languages lies a cosmological project of enormous consequence. To build on the Moon or Mars means more than placing infrastructure beyond Earth. It means redefining civilization as something that can leave its planet of origin and attempt to reproduce the conditions of life elsewhere. It means imagining the human as a worldmaking species.
This is why space architecture matters far beyond the specialized field that currently claims it. A lunar base is a statement about destiny, survival, extraction, expansion, risk, continuity, and the future of the species, and a habitat only secondarily. Every pressure vessel, radiation shield, rover, landing pad, regolith berm, power system, greenhouse, and closed-loop ecology participates in a larger question: what kind of being attempts to make a world where no world for it exists?
Architecture becomes cosmological again when it leaves the scale of the building and begins to operate at the scale of life-supporting worlds. The question expands from the organization of space into the production of conditions under which human life can continue inside an indifferent environment. Air becomes architectural, and so does radiation protection, while dust mitigation, supply chains, recirculated water, thermal regulation, artificial light, robotic construction, and psychological endurance all enter the same project. The building dissolves into a total environmental system, and architecture returns to one of its oldest but often forgotten ambitions: the making of a world.
The migration of architectural language into other technical fields should be understood in this sense. Software, AI systems, institutions, space missions, security systems, networks, simulations, platforms, and planetary strategies all borrow the term because architecture names an ordering of relations, a structure of possibility, a way of making certain futures more likely than others. When a space program calls itself an architecture, the word marks the point at which engineering becomes world organization.
The Pressure of a Whole
Simulation participates in the same transformation. A simulation is a model, but it is also an argument about which aspects of reality can be formalized, which forces matter, which variables count, which agencies are included, which futures can be rehearsed, and which forms of uncertainty can be tolerated. Simulations create operational worlds in which decisions can be tested before they become irreversible. Climate models, military simulations, economic forecasts, urban digital twins, AI training environments, game engines, and planetary-scale models all belong to this new technical cosmology. They make reality actionable by first making it world-like.
Games matter profoundly within this history. They are among the most sophisticated popular forms for encountering artificial worlds as worlds, and central to this history rather than marginal to it. A game gives an environment rules, affordances, memory, resistance, agency, time, consequence, and partial opacity. It teaches the player to inhabit a structured reality whose laws must be learned from within, through attention, failure, orientation, and repetition. Before AI took up the world model as a central ambition, game worlds had already made the relation between intelligence and environment tangible, navigable, and felt.
The contemporary importance of lore belongs to the same movement. The proliferation of fictional encyclopedias, wikis, bestiaries, manuals, maps, codices, timelines, and guides to imaginary worlds suggests that audiences are no longer satisfied by narrative alone. They want contact with the order beneath the event. They want the sense that the visible fragment belongs to a larger coherence, that a creature has an ecology, that a ruin has a history, that a weapon belongs to a vanished technology, that a ritual belongs to a theology, that a map extends beyond the quest. Sometimes the story is no longer the primary object. Sometimes the world is. Sometimes the deepest pleasure comes from sensing that there are laws, species, religions, catastrophes, lineages, architectures, and forgotten events extending beyond the frame.
This desire is cosmological in miniature. To enter a world is to sense the pressure of a whole.
Beyond Escapism
The temptation is to treat all of this as escapism, but that judgment is too thin for the scale of what is happening. The return of worlds in games, fiction, AI, space, and simulation belongs to the reorganization of reality itself. Contemporary culture is producing worlds because the inherited frameworks no longer seem sufficient to describe the conditions we inhabit. The nation is too small, the city too partial, the building too narrow, the individual too fragile, the platform too cynical, the planet too unstable, the machine too intelligent, and the future too technically mediated to be understood through older categories alone. We are surrounded by systems that operate at scales our disciplines were not built to think.
Worldmaking becomes necessary here because it gives us a language for reading these new cosmologies without reducing them to fiction, technology, ideology, or design. A world is a structured reality with its own beings, forces, laws, histories, scales, atmospheres, images, and futures. Worlds can be fictional, religious, technical, institutional, computational, political, architectural, or ecological. They can be built by artists, engineers, states, communities, platforms, models, or machines. What matters is that they organize reality as inhabitable, legible, and consequential.
The return of cosmology gives worldmaking its contemporary urgency. Worlds are no longer rare cultural achievements contained inside novels, myths, films, or games. They are becoming one of the operative forms of technical civilization, which is why the question now expands beyond what a thing does toward what world it assumes, what world it makes thinkable, and what forms of life it prepares.
Spending the Sun
Beneath every one of these world-pictures lies a condition more elementary than any of them, the flow of energy that allows a civilization to do anything at all, and that condition is becoming visible again after a long concealment. Solar panels, electrical grids, orbital infrastructure, datacenters, and the planetary computation that trains and runs the new models are all governed in the end by energy arriving from a star, and the limits they are beginning to encounter are thermodynamic before they are economic or political. The fossil centuries allowed civilization to draw on sunlight stored underground and to forget that it was a solar phenomenon at all, and that concealment is now ending. The sun returns to the center of the picture first as an engineering constraint, the budget against which a grid, a launch, or a model is sized. What reappears is the recognition that a civilization is a process running on an allowance it did not set, a local arrangement on the surface of a star-powered world, stripped of the worship an older cosmology once gave the same disk.
This recognition has a precise philosophical ancestor in Bataille, who took the first fact of any economy to be the abundance of energy the sun pours onto the Earth without asking anything in return, an excess that comes before scarcity rather than after it. A living system receives more than it needs to maintain itself, and the pressure of that surplus has to be discharged, through growth while growth remains possible and through expenditure once growth meets its limit, in monuments, festivals, sacrifices, and at the far edge in war. Read through Bataille, a civilization is defined less by how it produces than by how it spends the share it cannot use, and the great cosmologies were among other things systems for organizing that expenditure and giving it a meaning. The contemporary scene differs in its instruments and matches in its logic, since a datacenter buildout, a constellation of satellites, a program to settle another planet, and a training run that consumes the output of a power station are all ways of spending a solar surplus at a scale that once belonged to temples and pyramids.
The thinker who has taken this solar economy furthest is Nick Land, who reads the present as the moment the sun reclaims the center of ambition under technical cover. In a recent interview he describes Elon Musk as having revived for himself what he calls “an Egyptian type of ancient sun worship,” the underlying project being to reach the sun, tap it, and channel its output through an emerging superintelligence, a description that reads less as metaphor than as an account of a buildout in which solar capacity and computational capacity are now planned together. The prospectus that carried SpaceX to its record listing made the same point without irony, proposing to move computation into orbit precisely to sit nearer an inexhaustible source of power, so that the channeling Land sees appears in the documents of the present as a line of capital expenditure rather than a philosophical wager. What gives his version its edge is the subtraction he performs on Bataille. Bataille had left the surplus a sacred destination, a wealth that festival and monument could spend in ways that returned a meaning to those who spent it, and Land removes that destination, keeping the thermodynamics and discarding the altar, so that ambition, intelligence, and expansion appear as so many temporary arrangements for burning through an excess they did not choose and cannot keep. This is the cosmology that is feared rather than merely disputed, and it is feared because it refuses to let the expenditure mean anything human.
Seen this way, the history of intelligence becomes a chapter in the history of sunlight passing through one substrate after another, from the star to the biosphere it feeds, from the biosphere to the civilizations that grow inside it, and from those civilizations to the machines they have begun to build and the artificial minds those machines now carry. Each model that is trained, each world that is simulated, each image that is generated is a transformation of energy before it is anything else, which returns the most abstract of the new world-pictures to the most material of conditions. Land would read the superintelligence at the end of that series as the latest substrate through which the surplus is routed on its way to dissipation, alongside the older cosmologies that have gathered around the same lineage, whether Teilhard’s or the cybernetic dream of a thinking planet. The question this essay would rather hold than answer is whether he is right that nothing returns with the light except its loss, since he keeps the energy and discards the meaning, and the discarding is a decision rather than a discovery. The Egyptians who stand at the head of his own example tapped the sun and also worshipped it, spending its surplus in forms that meant something to them, and whether the return of the sun as a visible constraint reopens that possibility, or leaves us administering a solar economy we no longer know how to consecrate, remains genuinely open.
A Conflict of Cosmologies
This may be one reason the present feels so unstable. We are living through a conflict of cosmologies. One cosmology imagines intelligence as optimization, another as care. One imagines the planet as resource, another as organism, system, garden, archive, or wounded body. One imagines space as frontier, another as exile, extension, extraction, or responsibility. One imagines AI as servant, another as successor. One imagines the future as acceleration, another as repair. These are competing world-pictures rather than opinions about technology, and each one implies different forms of life.
The old cosmologies were never innocent. They organized power, hierarchy, exclusion, sacrifice, conquest, and obedience as often as they organized meaning. The new cosmologies will carry the same danger. Every image of the whole carries a politics. Every world-picture distributes agency. Every technical system that claims to model reality also decides what reality is allowed to contain. Every future smuggles in a theory of value. To ask what world a technology makes is to ask who or what can live inside it, who must adapt to it, who is excluded from it, and what forms of life it renders unthinkable.
Cosmological Literacy
The need for an image of the whole has reappeared inside the systems that now organize life. These are cosmological machines before they are anything else, and treating them only as technologies misses their deeper operation. They are making worlds, and we are already beginning to live inside the worlds they make.
The task that follows is a kind of cosmological literacy. We need to learn how to read the worlds being produced around us: their metaphysics, atmospheres, hidden gods, models of agency, images of time, definitions of life, assumptions about intelligence, promises of salvation, and tolerances for ruin. We need to ask what each world asks us to become.
Cosmology has returned because the whole has returned as a problem. The whole now appears as planet, model, machine, simulation, network, archive, and future. It appears damaged, artificial, intelligent, unstable, and unfinished. It no longer arrives as a completed order above us, but as something being built around us, through us, and sometimes without us.
That is why worldmaking matters. It is the discipline that becomes necessary when culture begins producing worlds faster than it can understand them.
Notes
On the world-picture, and Heidegger’s objection. The phrase is Heidegger’s, from Die Zeit des Weltbildes (1938), and his thesis runs opposite to a return: modernity is the age in which the world first becomes picture, set before a subject to be calculated. The picturing never receded; it deepened into the enframing of the world as a standing reserve, which the orbital datacenter literalizes. What returns is only the admission that there is one. The essay parts from Heidegger on the wager, which he doubted, that seeing the whole as a picture might open some relation to it other than mastery.
On the planetary and atmospheric images. Latour’s Facing Gaia (2017) gives the planetary actor its sharpest form, the Earth that enters history as an agent rather than a stage. Sloterdijk’s Spheres gives the architectural precedent, that to sustain life is already to build an atmosphere, so that air-conditioning, the space habitat, and the climate model belong to one history of managed interiors. The essay presses past both, since where Latour keeps Gaia secular and Sloterdijk keeps the sphere immanent, the wager here is that the planetary image has reopened the questions both were careful to bracket.
On Bataille, and the economy of the sun. The general economy of The Accursed Share (1949) begins from the surplus solar radiation pours onto the Earth rather than from scarcity. The excess has to be discharged, through growth while growth is possible and through expenditure once it meets its limit, gloriously or catastrophically. The essay keeps his premise that the sun is the first term of every economy and parts from his conclusion: Bataille kept the spending tragic and immanent, a glory with no meaning beyond itself, while the wager here is that the sun’s return as a visible constraint reopens whether the excess we now spend at planetary scale means anything besides heat.
On Teilhard, and the ascending mirror. The reading of intelligence as a cosmic stage descends most vividly from Teilhard de Chardin, whose The Phenomenon of Man (1955) saw evolution tending toward greater complexity and consciousness, thickening into a noosphere, a single layer of mind wrapped around the Earth, and converging at last on the Omega Point he identified with God. He is the sacred precedent for the thinking-planet prophecies now attached to computation, and the exact inverse of Land. Teilhard gives the lineage the telos and the meaning Land deletes, ascent toward Omega set against descent into dissipation, and the essay holds its question open between them.
On the Cosmists, and the engineering of heaven. The lineage from stars to superintelligence has a Russian precedent in the Cosmism of Nikolai Fedorov, whose common task proposed to turn technical means to the regulation of nature, the conquest of death, and the settlement of the cosmos, and in Tsiolkovsky, who carried it toward the stars. They share Teilhard’s directed optimism and already hold Land’s literalism, treating the sun, death, and the solar system as objects of management rather than contemplation. The orbital datacenter and the plan to tap a star belong to their tradition before they belong to ours, the surest sign that the present resumes this cosmology rather than inventing it.
On Land, and the cosmology that cuts deepest. Land’s solar economy was already in place in The Thirst for Annihilation (1992), where energy discharges unilaterally from the sun and life appears as a brief stabilization of its decay, “the feverish obscenity we call ‘life.’” His recent reading of Musk channeling solar output through superintelligence belongs to the same framework, now standing as a literal account of the companies building the frontier. He takes Bataille’s general economy and removes its sacred term, the source of both his reach and his dread.