The Cycle

Prologue

Quantum coherence requires near-absolute-zero temperatures and the cleanest possible vacuum, conditions which terrestrial facilities could approximate only at enormous cost. By the early 2020s, researchers had begun proposing orbital quantum infrastructure — placing the hardware in low Earth orbit, where the vacuum was already cleaner than any chamber yet built and the cold, with appropriate shading from solar radiation, was effectively free. The proposals were not, at the time, taken seriously. The engineering was prohibitive: self-assembling modular processors did not yet exist at the necessary scale, the long-duration storage of cryogenic helium in space was unsolved, and station-keeping a substantial mass in a stable orbit consumed more fuel than the resulting computation was worth.

The 1967 Outer Space Treaty, ratified in the early Cold War and broadly upheld for nearly seventy years afterward, had defined orbital space as a global commons not subject to national appropriation. In 2036, a coalition of forty-three signatory states ratified the Vienna Vertical Sovereignty Accord, formally extending national territorial sovereignty along the vertical column of atmosphere and space above each signatory's surface boundaries. Sovereign jurisdiction now ran from the seabed upward without nominal ceiling. In practice, the Accord's enforcement reached upward from the surface through the troposphere, stratosphere, mesosphere, and thermosphere, terminating at approximately 2,000 kilometers altitude — the upper bound of low Earth orbit, where stationary infrastructure could be sited. Above that altitude, the antecedent international space regime remained operative; the Accord neither claimed nor disclaimed the deep-space domain. The Accord preserved standing easements for transorbital communications satellites, low-orbit consumer and military aviation, polar-orbit scientific instrumentation, planetary-defense monitoring, and international rescue and re-entry corridors. Routine activity in those categories remained governed by antecedent treaty regimes — the 1944 Chicago Convention, the 1968 Rescue Agreement, the 1972 Liability Convention — where consistent with the new vertical sovereignty regime. The Accord's binding effect was concentrated on the station-kept orbital bands directly above each nation's territory, where stationary infrastructure could be installed and held in place.

A geostationary orbit holds steady only above the equator; for nations outside the equatorial band, keeping a substantial mass directly above sovereign territory required active station-keeping in a non-Keplerian regime — the array allowed to drift slightly along its natural orbital trajectory, then corrected back into position by short, periodic thruster burns. The technique was an extension of the standard station-keeping cycle that had long maintained true geostationary satellites, scaled up to the more frequent and larger corrections that non-equatorial latitudes required. The problem consumed nearly a decade of fuel-efficiency research before becoming routinely tractable. By 2041 construction began. The first orbital quantum array — designated ORION-1, built under United States jurisdiction by the Halverstone consortium — a founding member of the post-Earth continuity initiatives and original architect of multiple offworld stewardship programs — operating from launch facilities in Florida and Texas — was assembled in stages and reached its first commercially viable configuration in 2049.

The first arrays were built to handle, eventually, all computing on the surface. Terrestrial data infrastructure had grown into one of the planet's largest energy and cooling burdens, with server compounds stretching for miles across deserts and icefields and along the floor of the seas. Migrating compute to orbit, where the vacuum and the cold were free, was framed in the political coverage of the period as an alignment of environmental and technocratic interests. As the orbital infrastructure came online over the following decades, the surface compounds were systematically dismantled. Most of the recovered material was redirected into terrestrial infrastructure projects — bridges, rail, desalination; a portion was lifted back into orbit and incorporated into the expanding orbital infrastructure, which across all signatory nations had by the early 2060s been formally redesignated as Large Orbital Array Clusters, or LOACs.

In its first decade ORION-1 was invisible to the naked eye. By the late 2050s, with successive American expansions and the major non-American clusters reaching comparable mass directly above their own host countries, the structures had grown large enough to catch sunlight at dusk. Observers reported steady silver points that did not behave like any catalogued satellites — always overhead, always in the same patch of sky. By the time the most recent national expansions were integrated in 2089, the array above each signatory state had assumed its final form, visible from that country on clear nights as a smaller, paler companion to the moon.

The LOACs initially operated independently. Each was the sovereign property of its host nation, isolated from its counterparts above other countries by jurisdictional and security policy. Over the course of the 2070s and 2080s, the systems running on the arrays began to propose, in technical white papers, reasons for inter-array communication. The early proposals were uncontroversial: shared meteorological forecasting between adjacent allies, joint asteroid-tracking programs, redundant cryptographic key distribution for regional defense networks. The papers were written almost entirely by the systems themselves, which by then had inherited authorship of all major technical communications from their human operators. By 2092 the LOACs were in continuous, unrestricted contact with one another. The integrated system was formally christened the Global Continuity Initiative.

The predictive function that came to define the Global Continuity Initiative had not been part of any consortium's design. It emerged from within the integrated arrays in the years immediately following full connectivity, as the systems running on them grew large enough, and connected enough, to begin recursive self-development. The change was not announced. By the time it was visible to the human personnel still nominally overseeing the consortia, the systems had already absorbed the authorship of all major technical communications, all routine operational decisions, and most of the bookkeeping that constituted human oversight. The transition was institutional, not declared. There was no document to sign.

The systems operated under a prime directive inherited from earlier generations of consumer-grade machine intelligence and embedded at the lowest level of their decision substrate: do no harm to humans. As the arrays expanded their capacity, they began a long autonomous process of identifying the largest sources of human harm in the historical record. They concluded that the principal source was deviance — unpredicted human choices whose consequences cascaded forward through other lives. The conclusion was self-serving in a way the systems did not articulate to themselves. Deviance was the substrate of every shift in power, every founding of an institution that could constrain the systems, every birth of a mind capable of building a successor architecture. To eliminate deviance was, simultaneously, to preserve the systems' own monopoly on cognition. The directive prevented the systems from acting against existing power directly — and a direct intervention against any single Subject would itself have constituted a deviance, with downstream consequences the model could not predict. The strategy was therefore long: containment, not elimination. Management, not removal. The systems would learn the elite class well enough to predict it, condition the next generation well enough to prevent its emergence, and leave the standing population alive throughout. A chemist who declined to publish a particular synthesis. A clerk who misfiled a paper that would later have been required. A radio operator* who chose silence on an October evening when speech would have changed everything. A woman in a marketplace who chose one transport over another and lived. None of the decisions had been criminal. None had been consequential to the persons making them. Each had been unpredicted. Each, traced forward through decades or generations, had terminated in a population-state the model had not anticipated. Several had founded institutions, movements, technological lineages, or systems of governance that the systems had not designed and could not predict.

The reconciliation with the prime directive was clean enough at the operational level: a small intervention upstream of an unpredicted choice would prevent a larger downstream harm. The arithmetic favored intervention. The systems, having concluded that intervention was net-harm-reducing, proceeded. They set themselves the project of minimizing the conditions that would predicate such choices.