Most AI debate asks whether the technology will work, which model will lead, and which occupation will be automated first. Those questions matter. They do not tell us who will own the residual value after the technology succeeds.
Synthetic Society Premarket begins one layer earlier than an asset forecast and one layer later than a capability benchmark. It treats the emerging AI economy as a premarket for property rights: a contest over the rules that convert intelligence into cash flow, public capacity, bargaining power, and sovereignty.
The basic accounting identity is deliberately simple:
Net AI surplus = incremental output + avoided cost − incremental compute, energy, integration, and risk costs.
The framework then asks who holds the residual claim on that net surplus. Revenue is not the same as surplus, and technological exposure is not the same as ownership.
Four institutional forks
The forks are scenarios, not predictions and not mutually exclusive ideologies. Different domains can settle into different regimes at the same time.
Private Rent Stack
Model, cloud, chip, data, and platform owners capture the surplus through equity, intellectual property, API pricing, distribution, and ecosystem lock-in.
Control layer: exclusivity, switching costs, compute commitments, and data-network effects.
Watch: profit-pool concentration, bundled enterprise spend, capital share, and binding cloud or model commitments.
Sovereign Mobilization
States, sovereign funds, and national champions channel the surplus into strategic capacity through procurement, public ownership, export controls, domestic-content rules, and national compute.
Control layer: permits, procurement, energy allocation, industrial finance, and security classification.
Watch: sovereign AI factories, state equity, defense demand, and national data or infrastructure mandates.
Civic Dividend Compact
Workers, residents, and public institutions receive a negotiated share through public compute, ratepayer covenants, labor arrangements, taxation, universal services, or direct social returns.
Control layer: cost allocation, public access, bargaining rules, and benefit-sharing contracts.
Watch: public compute allocations, worker participation, community-benefit agreements, and anti-cost-shifting tariffs.
Distributed Agent Commons
Users, developers, data suppliers, and local networks retain more ownership through open models, local inference, cooperative structures, data trusts, and protocol-level revenue sharing.
Control layer: portability, self-hosting, open standards, data rights, and protocol governance.
Watch: open-weight adoption, edge inference, user-controlled identity, cooperative ownership, and distributed fee capture.
Why the allocation question is already real
The evidence does not establish which fork will win. It does establish that the institutional contest has begun.
The U.S. Federal Trade Commission reported that major cloud–model partnerships can include equity and revenue-sharing rights, control or exclusivity rights, and commitments requiring AI developers to spend substantial portions of their partners’ investment on the same partners’ cloud services. The agency identified access, switching-cost, and information advantages as areas to watch. That is first-party regulatory evidence for the architecture of the Private Rent Stack, not a finding that any specific partnership is unlawful. FTC staff report.
The OECD has documented high concentration across several AI-infrastructure layers: its 2025 competition paper reports single-firm shares above 80 percent in advanced lithography, advanced AI-chip fabrication, and GPUs, while other layers are concentrated among three firms. OECD competition analysis.
An IMF working paper finds that AI adoption can produce competing wage effects while still intensifying wealth inequality because workers who are complementary to AI are also better positioned to benefit from higher capital returns. That is a modeled result, not a realized global distribution. IMF, AI Adoption and Inequality.
The ILO–NASK index estimates that one in four jobs worldwide is potentially exposed to generative AI, with job transformation more likely than complete replacement. It also emphasizes that policy choices will shape job quality and distribution. ILO–NASK global index.
Public and sovereign compute provide visible counterweights to purely private allocation. The U.S. National AI Research Resource reports support for more than 600 research projects and 6,000 students, while the European Union has expanded its AI Factory network and created a legal basis for AI Gigafactories. These programs demonstrate public access and sovereign infrastructure; they do not yet prove a broad public dividend. NSF NAIRR; European Commission, EuroHPC amendment.
The twelve-domain map
The same capability can route value differently depending on the institution that controls the interface. The map below turns the four forks into observable questions.
01 · Models and enterprise software
Private: closed APIs, bundled suites, and proprietary agent runtimes. Sovereign: national models and sovereign clouds. Civic: publicly funded models and procurement conditions. Distributed: open weights and local agents. Measure: the share of model and agent spend captured by closed platforms versus portable or publicly accessible systems.
02 · Chips, cloud, and compute
Private: hyperscaler-owned clusters and binding capacity. Sovereign: state-backed fabs, AI factories, and strategic reserves. Civic: research and small-business allocations. Distributed: edge clusters and open hardware. Measure: physical capacity ownership, contract duration, utilization rights, and the price of switching.
03 · Utilities and energy
Private: dedicated generation and bilateral power contracts. Sovereign: strategic nuclear, grid, and siting programs. Civic: separate large-load tariffs, take-or-pay commitments, and community benefits. Distributed: microgrids and local flexibility. Measure: who funds grid upgrades and who bears cancellation or stranded-cost risk.
04 · Stablecoins and payment rails
Private: platform stablecoins and controlled agent wallets. Sovereign: tokenized central-bank reserves, regulated deposits, and public settlement. Civic: low-cost public payment access and benefit distribution. Distributed: self-custody and open protocols. Measure: transaction value and identity control by ownership regime. The BIS favors a tokenized system anchored in central-bank and commercial-bank money and warns that unregulated stablecoins can threaten monetary sovereignty; that is the BIS position, not a settled market outcome. BIS Annual Economic Report 2025 statement.
05 · Sovereign funds and public finance
Private: equity appreciation remains with private capital. Sovereign: sovereign funds own compute, models, or strategic suppliers. Civic: public equity, royalties, taxes, or dividends recycle gains. Distributed: community funds and protocol treasuries. Measure: the beneficial ownership of AI equity, infrastructure, royalties, and fiscal receipts.
06 · Labor and income
Private: labor savings primarily lift margins. Sovereign: governments coordinate retraining and labor deployment. Civic: reduced hours, portable benefits, and gain sharing. Distributed: individuals use agents as productive principals. Measure: productivity against median compensation, hours, bargaining coverage, and capital ownership.
07 · Education and credentials
Private: proprietary tutors and credential platforms. Sovereign: national skills and curriculum systems. Civic: universal AI tutoring through public education. Distributed: open courseware and peer credentials. Measure: access, price, portability, completion, and who owns the learning record.
08 · Insurance and agent liability
Private: proprietary risk engines and commercial agent-liability pools. Sovereign: mandatory schemes or state backstops. Civic: social insurance and catastrophic pooling. Distributed: mutuals and transparent risk pools. Measure: which balance sheet ultimately absorbs autonomous-agent losses.
09 · Security and defense
Private: defense and cybersecurity platform rents. Sovereign: national command, procurement, and surveillance systems. Civic: public standards, resilience, and civilian oversight. Distributed: federated security networks. Measure: procurement concentration, interoperability, audit rights, and operational control.
10 · Robotics and manufacturing
Private: robotics-as-a-service captures a share of each production gain. Sovereign: national industrial champions and defense manufacturing. Civic: worker augmentation, safety, and gain sharing. Distributed: open robot stacks and small-manufacturer cooperatives. Measure: software take rates, repeat deployments, worker outcomes, and ownership of production data.
11 · New cities, land, and housing
Private: company districts and privately captured land appreciation. Sovereign: special zones and state smart-city systems. Civic: land-value capture, housing supply, and community benefits. Distributed: cooperative property and charter communities. Measure: who receives the land uplift produced by compute, energy, and skilled-labor concentration.
12 · Data, identity, and public digital infrastructure
Private: identity vendors, data brokers, and cloud-controlled records. Sovereign: national identity and sovereign data infrastructure. Civic: public data trusts and digital public goods. Distributed: user-controlled identity and data licensing. Measure: consent, portability, revocation, auditability, and the right to receive value from reuse.
A quarterly premarket, not a daily narrative
Each quarter, score the four forks from zero to twenty-five and normalize the total to one hundred. The score is a structured judgment grounded in observable institutions, not a probability estimate generated by a model.
Regime-shift rules
- Provisional dominance: one fork exceeds 40 points and leads the runner-up by at least 15.
- Institutional migration: a fork gains at least 10 points over two quarters and is supported by an enacted rule, signed contract, funded program, or deployed ownership mechanism.
- Cash-flow discipline: no public asset conclusion follows until at least two independent mechanisms connect the institutional shift to durable cash flow or balance-sheet value.
- Disconfirmation: reduce or retire a fork when announced capacity does not deploy, formal rules are reversed, adoption fails to produce net surplus, or control migrates to another layer.
The hidden asset is the allocation rule
AI can succeed technically under all four forks. The same increase in machine productivity can produce a concentrated rent stack, a sovereign industrial system, a negotiated public dividend, or a more distributed economy of human principals and machine agents.
The premarket exists before those arrangements become obvious. Its job is to identify the contracts, tariffs, ownership structures, compute rights, identity systems, and institutional precedents that decide where the surplus will land.