Conjecture 1 – Value-Based Erosion (Orthodox Marx Outcome)
In this first scenario, the economy actually gets better at producing things. Output keeps climbing. The problem is that we lose a reliable way to turn all that productive power into broad purchasing power and steady profit growth. You end up with what you might call “productivity without value”: more gets made, but the connections between wages, demand, and social stability start to fray. And it’s not because anyone is trying to break the system. It’s because the incentives keep pushing people and firms toward choices that make sense for them individually, but add up to something destabilising overall.
Through a game theory lens, it looks like a Red Queen race that people talk about as innovation. Every company faces the same basic choice: automate aggressively or fall behind. Slowing down sounds nice in theory, but in practice it’s not really available. If your competitors automate and you don’t, your costs look worse, your prices become less competitive, and you lose share. So firms automate because they have to, not because they’re ideologically committed to it. Over time, that pushes the whole system toward widespread automation, a smaller slice of income going to workers, and tougher price competition. Each company is trying to protect itself, but the combined effect is weaker demand, less pricing power for most businesses, and a more fragile economy.
In an ideal world, you’d manage that transition with coordination: maybe you pace adoption, strengthen income support, invest heavily in retraining, or redesign redistribution in a way that keeps demand healthy. The catch is that individual companies can’t coordinate that on their own. Governments can try, but they run into a structural bind: as labour income shrinks, payroll and income tax receipts soften, while capital and business models move more easily across borders. So the state is being asked to stabilise demand just as its most reliable revenue base narrows. That’s why this scenario tends to swing between policy approaches rather than settling into something durable.
If you zoom out and look at the system dynamics, the engine becomes clearer. There’s a powerful reinforcing loop: invest in AI, productivity rises, competition intensifies, and that pressure forces even more AI investment. It kicks in quickly and speeds up over time. Historically, capitalism also had a stabiliser: jobs and wages supported spending; spending supported demand; demand supported profits; profits funded investment. AI weakens that stabiliser because a smaller share of output flows back into wages. So the economy’s ability to produce can outrun its ability to distribute income. When that gap gets big, you get an “unstable growth machine”: output keeps rising, but value realisation gets harder, volatility increases, and markets become more sensitive to policy signals and liquidity.
Timing is the killer. The automation and competition loop plays out fast. The counter-response is slow. Policy, institutions, and social bargains usually move after the fact, once trust has already been damaged and the set of feasible options has narrowed. So you get a pattern of oscillation: stimulus followed by backlash, deregulation followed by crackdowns, optimism followed by crisis. Not because anyone prefers chaos, but because the feedback structure makes smooth adjustment difficult.
Once you hold those pieces together, the consequences stop looking like a random list and start looking like a connected chain.
On jobs, the first hit is pretty direct: routine office and clerical work erodes quickly, and wages compress across a lot of middle-skilled roles. Over time, the deeper damage shows up: more long-term underemployment, less willingness to retrain because it feels like a risky bet, and a growing sense of “growth without progress.” The labour market doesn’t just shift. It splits, with a small number of highly rewarded roles and a much larger group facing weaker bargaining power, more instability, and fewer paths upward.
For governments, the first effect is fiscal pressure. Wage-linked tax bases weaken, while demand for welfare, unemployment support, and social insurance rises. The second effect is political volatility. Some governments tighten spending to protect credibility. Others spend more to protect legitimacy. Both approaches become unstable when the revenue base is eroding and the underlying coordination problem hasn’t been solved. Over time, debt stress builds—sometimes it breaks into crisis, but more often it shows up as a slow loss of room to manoeuvre. The system gets more brittle because the state is expected to provide stability without the revenue structure it used to rely on.
For investors, the first effect can feel counterintuitive at first: outside of monopoly or bottleneck positions, average profit rates compress. Competition gets harsher, prices come under pressure, and demand is less reliable. Investors respond by crowding into perceived winners and leaning more heavily on financial engineering, speculative assets, and liquidity-driven narratives. The second effect is fragility. When productivity growth no longer translates cleanly into earnings growth, valuation becomes more sensitive to discount rates, policy, and confidence. You get more boom–bust behaviour as capital searches for safe places in an economy that struggles to convert its productive capacity into broad-based demand.
Companies, meanwhile, become more defensive. For most firms, productivity gains don’t turn into durable profit expansion because AI tools spread quickly and competition intensifies. Cost cutting becomes the default strategy, not because it’s inspiring, but because it’s how you survive in an arms race. The second-order result is more consolidation and stagnation. If demand is fragile and differentiation erodes, firms merge to regain pricing power, or they lean on cheap capital just to keep margins intact. Innovation still happens at the frontier, but it doesn’t diffuse as broadly, and the middle of the economy becomes sluggish.
When you map this onto asset classes, the same logic shows up.
Public equities become more split. A small number of dominant firms can sustain high margins because they control distribution, data, platforms, or regulatory moats. Most of the market faces valuation pressure because earnings growth doesn’t keep up with productivity. Indexes might look fine for a while, but the risk shifts toward long periods of weak real returns, because broad earnings power is capped by weak income distribution.
In private equity and growth equity, the usual playbooks get tougher. EBITDA growth becomes less dependable because the same automation forces that lift productivity also intensify competition and weaken demand. Exit multiples compress, leverage gets riskier, and underwriting shifts toward resilience and cash extraction rather than multiple expansion. Roll-ups can still work in fragmented sectors that are operationally messy, locally rooted, or hard to automate, but service models that depend mainly on scaling labour structurally weaken.
Venture capital doesn’t vanish, but it becomes more barbelled. The outliers can be enormous because controlling bottlenecks matters more, while the middle thins out as features commoditise and distribution dominates. The best opportunities cluster around infrastructure and enabling layers—compute, data, security, tooling, integration—rather than startups whose main edge is simply replacing labour for a short window.
Credit becomes more fragile too, because the assumptions lenders usually underwrite—stable pricing, predictable demand, reliable EBITDA—start to wobble. Cycles shorten, default risk rises, and lenders move toward collateral, tighter covenants, and stricter structures. Senior secured and asset-backed lending holds up better than loans that rely on optimistic growth and loose terms.
Real assets also split. Office and urban commercial property tied to dense employment and rising wages face structural repricing as work patterns change and labour income weakens. Regulated infrastructure, by contrast, looks more bond-like: expensive to build, hard to replicate, supported by policy, and less sensitive to demand volatility. And as property and wage-linked revenues soften, city finances and local politics get more strained in lasting ways.
The big takeaway is that none of this depends on technology failing. It just requires the incentives to keep pushing in the same direction while institutions fail to respond fast and strongly enough to preserve a stable conversion from capability into demand and legitimacy. The economy becomes incredibly capable. What lags is the set of mechanisms that make that capability broadly stabilising. You get a world with very high output—and persistently low stability—an economy that can do almost anything, except share the gains in a way that holds over time.
Conjecture 2 – Rentier Capitalism (Scarcity Relocation Equilibrium)
In this second scenario, AI doesn’t really wipe out value. What it does is move where scarcity lives. Instead of value being tied to hours worked or people employed, it shifts toward coordination choke points. Whoever controls distribution, identity, standards, regulation, data, compute, or trust ends up controlling the flow of value. Think of these as toll booths in the economy. The money is still there, often more than before, but it’s earned less by employing people and more by owning the bottlenecks everyone else has to pass through. Profits stop being tightly linked to job creation, and labour quietly stops being the main extraction point.
Once that happens, the broader shape of the economy starts to look familiar. Power and profits concentrate. You get a system where headline profitability stays high even as labour’s share keeps shrinking. Socially, it stops looking like one big middle class and starts looking like two worlds. On one side are the AI owners, platform operators, and the institutions wrapped around them. On the other side is everyone else, still mostly employed, but with weaker bargaining power, more insecurity, and a growing service layer whose job is essentially to support the winner economy. Inequality rises, not through a dramatic jobs apocalypse, but through a slow erosion of the social bargain. Work stops feeling like a reliable source of identity or upward mobility, and class lines start to harden.
If you look at this through a game theory lens, it’s not a clean innovation race at all. It’s more like a contest for choke points, mixed with a regulatory capture game. Firms aren’t just trying to build better products; they’re trying to become unavoidable. The enterprise operating system. The distribution rail. The identity layer. The marketplace everyone has to plug into. The payoff structure is brutal. If you win, you get outsized rents. If you lose, you get leftovers. So the playbook is predictable: scale fast, subsidise aggressively, bundle everything, buy or block rivals, raise switching costs, litigate when needed, and lobby constantly. In that world, the equilibrium tends toward winner-take-most. And for most non-platform companies, the rational move isn’t to fight the giant, it’s to attach yourself to it.
Governments then step in, but not as neutral referees. This becomes a two-level bargaining game. States want legitimacy, tax revenue, and some degree of control. They also want growth, strategic industries, and national champions. Platforms offer exactly that: scale, innovation optics, and geopolitical leverage. What you often get is a negotiated truce. Platforms accept some rules and oversight in exchange for being allowed to dominate. From a systems perspective, the main feedback loop is reinforcing: more scale leads to more rents, which fund more investment, which leads to even more scale. The balancing forces—antitrust, regulation, political backlash—exist, but they tend to arrive late, weak, or already shaped by the firms they’re meant to restrain. The result is a system that looks stable and efficient on the surface, while quietly becoming more brittle underneath.
A helpful analogy here is ecology. This starts to look like an apex predator ecosystem. A few massive players dominate territory and crowd out diversity. Survival depends less on adaptability and more on holding ground and excluding rivals. Innovation slows not because people stop being smart, but because the ecosystem itself gets thinner. Everything works beautifully as long as conditions stay stable. When shocks hit, though, the whole thing can unravel very quickly.
The cleanest way to summarise it is this: value moves, profits don’t disappear. Jobs polarise. Bargaining power weakens. Capital piles into IP, data, compute, and the equities of the dominant firms. You start seeing knock-on effects everywhere. Stock indices become quiet bets on a handful of monopolies. Political risk creeps into asset pricing. Credit splits in two: pristine balance sheets for platform giants, tightening conditions for SMEs. Physical infrastructure like data centres, energy, and networks explodes in importance and starts to look less like private real estate and more like strategic national assets.
What eventually breaks this setup is also fairly clear. Either regulators get serious—real antitrust, real interoperability, real reductions in switching costs—or geopolitics intervenes. Sanctions, security concerns, and national control over data and compute can force fragmentation. If neither happens, the system can run for a long time, compounding rents, until it hits its real failure mode: a legitimacy crisis. Profits keep flowing, but society stops accepting the story that justifies them. When that happens, change tends to be abrupt rather than gradual—breakups, nationalisations, hard regulatory resets, or a fractured techno-world split into blocs. So rentier capitalism isn’t a calm end state. It’s a long, flat plateau that looks stable right up until the ground gives way.
Conjecture 3 - Human-Leverage Economy (Context-Sensitive Augmentation)
In a human-leverage economy, AI doesn’t wipe out work. What it really does is change where the value sits. The payoff shifts away from sheer effort or routine execution and toward good judgment. Yes, a lot of repetitive tasks get automated. That part is real. But the truly scarce input becomes something else entirely: accountable decision-making in environments where mistakes are costly and trust is non-negotiable. Think medicine, finance, critical infrastructure, compliance, security, procurement, safety, complex operations. The constraint isn’t raw intelligence in the abstract. It’s much more concrete than that. It’s who can be trusted to sign their name, explain the reasoning, and carry the consequences when things go wrong. Work doesn’t disappear. It becomes more leveraged. The most valuable output increasingly comes from human judgment amplified by machines, not from humans being replaced by them.
Once you see that, the broader political and economic patterns fall into place. The labor market doesn’t collapse; it re-sorts itself. A distinct tier of high-judgment roles expands: people who design systems, oversee them, integrate them, govern them, audit them, handle edge cases, and take responsibility when reality refuses to behave. At the same time, a lot of routine work shrinks, flattens, or gets compressed into smaller roles. This doesn’t mean wages fall across the board. It means dispersion increases. The skill premium rises sharply. Education and credentialing regain real power, not just as social signals, but as the actual gatekeeping infrastructure that determines who gets access to the leverage layer. In that world, public services like education, training, healthcare, and digital governance stop looking like optional add-ons. They become strategic investments, because they determine how large the judgment class is and whether the system feels fair or legitimate.
If you step back and look at this through a game-theory lens, it’s essentially a coordination problem. The good equilibrium is one where firms invest in augmentation and governance, workers invest in skills, and consumers and regulators reward systems that are transparent, safe, and accountable. The crucial detail is that firms don’t win by eliminating humans. They win by building real trust architecture: human-in-the-loop workflows that aren’t cosmetic, but are explicitly designed around liability, oversight, and legitimacy.
But there’s a catch. There isn’t just one equilibrium. There’s also a worse one, where firms chase cheap automation, regulators lack the teeth to punish corner-cutting, and the system drifts toward instability, erosion, or rent extraction at choke points. Which world you end up in depends heavily on enforcement: liability regimes, standards, audits, and the real ability to penalize shortcutting when it happens.
From a systems perspective, this economy runs on a mix of reinforcing and balancing forces. The reinforcing loop is straightforward. Better AI tools make judgment more productive. That raises the return to skill. People invest more in skill. Augmentation becomes even more valuable. But there’s also a stabilizer: accountability. As risk and liability rise, the demand for human oversight and explainability rises with them, placing a very practical ceiling on full automation. When this balance holds, you get something that looks like slow, steady compounding. Institutions, skills, and tools co-evolve. Growth is slower but sturdier. The labor market is stratified but still functional. And the system is more robust because power and control aren’t concentrated in a single choke point. The core failure mode remains access. If too many people are locked out of leverage roles, society splits into an elite layer and a permanently excluded one.
Evolutionarily, the closest analogy is mutualism. Humans, AI, and institutions co-adapt, and selection favors those combinations that blend machine capability with human judgment in ways that compound learning, reputation, and institutional fit. You don’t get one dominant species that consumes everything overnight. You get many viable niches and a more resilient ecosystem. The biological failure mode is what you might call speciation lock-out: only a narrow elite gains access to the leverage tools and pathways, mobility drops, and stratification hardens. Even so, this remains about as healthy an equilibrium as capitalism can reach without a full redesign. It’s reformist rather than revolutionary.
Once you translate this into practical terms, the implications sharpen quickly. Routine jobs shrink, while judgment-heavy roles expand, followed by second-order effects like education arms races, credential inflation, and tension between leveraged and non-leveraged workers. Governments face mounting pressure to modernize education and training and to treat public services as productive capacity, gradually becoming skills allocators and legitimacy providers, with inequality management moving to the center of policy. Capital flows toward firms that combine talent and AI effectively, especially in judgment-dense sectors, leading to longer investment horizons and lower volatility as value embeds itself in institutions rather than hype cycles. Companies discover that talent systems and AI integration are existential, governance becomes a competitive advantage, and organizations start to resemble hybrids of guilds and platforms, where culture and ethics are strategic assets and innovation is slower but more durable. Markets follow suit: public equities reward trusted, auditable talent-plus-AI businesses; private equity finds operational alpha in governance and talent rather than cost cutting alone; venture capital produces fewer moonshots but more solid outcomes, particularly in vertical AI and decision support; credit expands into services with stronger underwriting; and real assets tied to education, healthcare, logistics, and mixed-use infrastructure regain strategic importance.
And in the end, the whole story really does reduce to a single line: AI makes judgment more valuable, because judgment is where accountability lives. If access to that leverage layer can be widened through training, credible credential pathways, and strong standards, you get a high-trust equilibrium that can compound for a long time. If it can’t, the same mechanics that make the system stable for insiders make it unforgiving for everyone else.
Conjecture 4 - Post Labor Capitalism (Synthetic Labour Regime)
This is basically the endgame of “value without workers.”It’s the point where AI crosses a practical threshold of autonomy and the economy can keep producing high value output with very little human participation in the sectors that matter. And that’s not just another wave of automation. It’s more like a regime change. A speciation event in political economy.
Because for the first time since industrialization, value creation stops depending on the traditional human role of labour. And once that happens, the wage relationship stops being the main way society organizes itself.
In classical capitalism, jobs and wages are the glue that ties value to legitimacy. You can argue about whether it’s fair, but the basic story holds: people work, they get paid, they consume, the system feels broadly legitimate. In a synthetic labour world, that braid snaps.
The arithmetic gets brutally simple: surplus is just output minus cost. And for lots of digital goods and AI mediated services, marginal cost trends toward zero. So the advantage isn’t “who has higher labour productivity” anymore. It’s who has better models, more compute, cleaner data pipelines, stronger distribution, and—this is the big one—stronger permission to operate. Control, access, and legitimacy become the decisive variables. Ownership shifts from “companies that employ people” to “systems that produce.”
That flip also breaks the classic bargaining logic.
The old Marxist frame assumed capital needed labour, and labour had leverage because it could withdraw. In this regime, capital increasingly does not need labour. If workers walk out, production often continues. So labour’s bargaining power collapses, and the core negotiation shifts from capital versus labour to the state versus AI capital—over surplus, control, and legitimacy. The real power sits with whoever owns the models, compute, distribution rails, and the infrastructure stack underneath them.
Socially, humans become economically peripheral.
People remain biologically and culturally central, obviously. But economically, the system needs them less. First order effects are pretty straightforward: structural unemployment at scale, and wage based livelihood stops being the default. Work doesn’t vanish, but it turns into something more optional, niche, artisanal, or symbolic—almost like a leisure economy.
Second order effects are nastier. Identity and routine erode. Status competition replaces labour markets as the main sorting mechanism. And you can get mass alienation even in a world of material abundance, because the social contract breaks. The blunt question becomes: if income isn’t earned through wages, what’s the basis for distribution—and even for citizenship?
Now if you look at it as a system, it’s technically efficient but socially decoupled. Cybernetically incomplete.
The dominant structure is a reinforcing loop: AI autonomy rises → output rises → capital accumulates → AI scale rises → autonomy rises again. Every turn of the loop makes the next turn easier and faster. That’s why this regime can generate explosive output.
But what’s missing is the balancing loop that stabilized capitalism historically: income legitimacy. In a wage society, wages distribute purchasing power and create consent, and that consent stabilizes the system. In a synthetic labour society, output no longer flows back through wages, so that stabilizing negative feedback collapses. Production detaches from social reproduction. Legitimacy detaches from employment. Wealth and capability concentrate fast. Political instability stops being an accident. It becomes structural.
So the strategic interaction turns Hobbesian.
You get a power transition with a commitment problem. The state needs revenue, stability, legitimacy. AI capital controls output and critical infrastructure, and often the distribution rails. Citizens—who no longer have wage leverage—push the state instead.
Two broad equilibria start to look plausible:
Corporate sovereign equilibrium: platforms provide services and stability, and the state tolerates dominance because it’s terrified of disruption. Society runs on corporate provision, and the state becomes thinner.
State capture / nationalisation equilibrium: the state taxes, regulates, or expropriates AI infrastructure and redirects surplus to maintain legitimacy. The state becomes more centralized and coercive because it has to.
Both equilibria are unstable because the commitment problem is knife edge. If firms fear expropriation, they underinvest or move. If the state can’t credibly tax or compel contribution, legitimacy collapses. Citizens excluded from wages force the state to act, which raises expropriation risk. Stability requires either strong legitimacy—credible distribution and trusted institutions—or strong coercion—hard control of capital, information, and unrest. Anything in the middle is fragile.
Underneath all of this is the defining fact: production decouples from population.
Output no longer scales with people, education, or employment. It scales with models, compute, energy, and permission. That single shift fractures the modern world order, because it re sorts which regimes work and which don’t.
Market economies can survive, but only if they mutate away from labour anchored legitimacy.
Their historic strengths—open markets, mobile labour, entrepreneurship—lose force because labour isn’t the scarce input anymore. Markets might still allocate capital efficiently at the production layer, but capital becomes politically dangerous once it stops distributing income through wages.
Early phase looks like a weird paradox: soaring productivity, collapsing employment, asset inflation, rising income insecurity. Consumption decouples from production. Monetary policy loses traction because employment is no longer the main transmission channel. Growth no longer buys legitimacy.
So market systems get forced to re politicize distribution. Income has to flow via non wage mechanisms: universal transfers, social dividends, public ownership of AI infrastructure, compulsory surplus recycling—some version of “the loop must close.” Markets can still run production, but the social contract moves upward into explicit redistribution and entitlement design.
The failure mode is delay. If they cling to wage centric thinking too long, legitimacy collapses: inequality widens, the middle hollows out, electoral volatility spikes. Markets get blamed for outcomes they no longer control, and the response becomes capital controls, punitive regulation, and de facto nationalisation. Markets don’t disappear. They lose autonomy.
Small economies get re sorted the fastest, and the most brutally.
In a labour based world, small countries could compete through education, specialization, supply chain integration, exporting talent. Synthetic labour removes those levers. What matters becomes chokepoints: energy, compute hosting, data jurisdictions, legal shelter, strategic geography.
So small states bifurcate. Some become AI havens—trusted legal regimes, compute hosts, energy exporters, regulatory neutral zones—and punch far above their weight as rent extracting nodes. The rest become managed peripheries: weak currencies, eroding tax bases, shrinking autonomy, living off transfers, remittances, or geopolitical patronage. Post labour capitalism isn’t a gentle adjustment for them. It’s a binary outcome: strategic node or dependency zone.
Authoritarian regimes are, counter intuitively, initially advantaged.
They already operate without labour legitimacy, and they’re comfortable with surveillance, centralized control, non consensual redistribution, and industrial policy. They can substitute AI for human productivity without paying the normal political price of unemployment. Short to medium term, they may look more stable than liberal market systems.
But it’s brittle. Once labour is irrelevant, the population becomes economically redundant. Loyalty must be purchased or enforced entirely through distribution and coercion. If surplus falters—sanctions, energy constraints, inferior models, capital flight—there’s no wage buffer, no civil society shock absorber, no independent economic base. So when authoritarian regimes fail in this world, they tend to fail catastrophically: elite fractures, repression spirals, regime collapse. There’s no gentle path down.
Across all regimes, the game isn’t ideology. It’s feedback design.
The durable systems are the ones that can: control AI production and distribution, recycle surplus credibly into social stability, and maintain legitimacy without labour participation. Everyone else drifts toward one of three outcomes: quiet dispossession under high control, violent expropriation cycles, or loss of sovereignty and relevance.
And that brings you to the deepest implication: post labour capitalism doesn’t abolish markets or states. It breaks the post war settlement that linked work, citizenship, income, and dignity.
Every regime has to answer the same question in a new way: if people are no longer economically necessary, why does the system owe them anything? The systems that can answer that credibly survive. The ones that can’t may still produce enormous output, but they won’t have a society that can meaningfully live inside it.
That’s why the key failure mode isn’t only unrest. It’s mass dispossession without agency. If distribution isn’t redesigned, people lose both income and role, while traditional bargaining channels disappear. Collective action gets harder as surveillance, dependency, and fragmentation rise. You drift into a low legitimacy, high control equilibrium—more like a quiet collapse than a classical revolution.
If you want a better mental model than equilibrium economics, think evolution.
This looks like an invasive species displacement event. Selection pressure rewards replication without dependence on humans: speed, scale, autonomy. The fitness function becomes output per unit of constraint. Humans fall out of the fitness equation. Institutions built for wage society become maladapted. It’s not tidy. It’s a speciation crisis.
Capital markets reshape accordingly, because capital still matters—but the kind of capital that matters changes.
The scarce asset stops being “growth companies with teams.” It becomes control over autonomous productive systems. Venture capital, optimized for a world where humans were the bottleneck, either bifurcates or withers. Team risk collapses as an underwriting variable, time to scale compresses, and capital becomes less patient fuel and more permission to scale infrastructure.
The venture funds that survive become platform adjacent: funding model breakthroughs, securing early compute and data access, locking in distribution and regulatory embedding. Narrative VC—thin apps, founder mythology, labour arbitrage—gets disintermediated by corporate AI labs, state AI stacks, and PE owned AI roll ups. Funds consolidate, returns become even more power law, and venture stops being democratizing. It becomes permissioned. Feudal, basically.
Private equity looks like a winner at first, and a hostage later.
Synthetic labour is the ultimate cost reduction lever, so margins expand and buy and build accelerates. But once portfolios shed workers and become essential infrastructure, they become visible and taxable. Political risk dwarfs operational risk. Returns compress through taxation, price controls, forced nationalisation. End state, PE either becomes regulated infrastructure operation or a quasi state partner. Traditional PE alpha—financial engineering plus labour optimization—dies.
Public markets don’t just re rate in this regime. They re politicize.
A small number of AI sovereign platforms—owning frontier models, compute, and distribution rails—start looking less like tech companies and more like utilities with imperial privileges. They can levy tolls on activity, set rules of access, and extract surplus as a matter of architecture. Cash generation can rival the fiscal capacity of states.
Valuations can go vertical, but what follows isn’t just multiple compression. It’s political exposure. Once you’re systemically necessary, you’re systemically taxable, regulable, and expropriable. Price becomes less about growth and more about permission.
That’s where the Roman and early modern parallels land.
In Rome, the publicani weren’t just contractors—they were private capital performing quasi sovereign functions like tax collection and provisioning, becoming political actors. The East India Company and VOC formalized the next step: chartered corporations with monopoly privileges, administrative capacity, armed force, and the right to make markets.
In a synthetic labour economy, AI platforms drift toward the same structural role. They don’t need to govern in name to govern in fact. Infrastructure control becomes the modern charter. Distribution control becomes the customs house. Compute becomes the fortified port. “Platform economics” becomes toll economics.
Meanwhile, labour heavy incumbents without AI sovereignty—services, retail, logistics, anything with human dense cost structures—get squeezed. Their margins compress, their equity stories unravel, and they get forced into restructuring or acquisition. Competition stops being on comparable terms. It becomes a hierarchy of privileges: chokepoint owners at the top, everyone else negotiating access.
So equity markets stop being mostly about growth and start being about political survivability. Investors price charter risk: expropriation risk, regulatory alignment, national importance, and whether the state treats the firm as partner, adversary, or asset. Capital markets become geopolitical instruments because the underlying assets have become geopolitical.
Private markets fracture the same way.
The classic mid sized private company—fifty to five thousand employees, defensible niche, human execution moat—gets squeezed from both ends. You get micro firms that survive as artisanal, taste and trust businesses: culturally meaningful, economically marginal. And you get AI native micro giants: tiny human cores sitting on massive synthetic labour, scaling without hiring, producing output that rivals mid caps. Entrepreneurship doesn’t disappear. It splits: tiny and human, or vast and inhuman, with little stable middle.
A new capital hierarchy asserts itself, and it looks uncomfortably pre modern.
At the top: owners of frontier models, compute, distribution—new chartered powers. Close behind: states that can grant or revoke permission at scale via coercion or legitimacy. Then: PE owners of essential infrastructure, and a shrinking class of privileged early access VC. Public shareholders without control rights sit lower than they expect, because the real game is no longer shareholder value maximization. It’s regime maintenance.
That’s the meta shift: capital becomes political by default.
Allocation stops being “neutral optimization.” It becomes an implicit vote on sovereignty—what will be tolerated, taxed, regulated like a utility, or treated as an instrument of national power. Finance doesn’t just stop pretending to be neutral. It loses the ability to pretend.
And that brings it back to the core point: Scenario 4 only stabilizes if you rebuild the missing feedback loop.
You need some durable mechanism that reconnects output to social reproduction—credible distribution plus legitimacy without wages. Without that, the system oscillates between coercion, expropriation, and collapse. With it, you can get to something like a post wage social contract—not because it’s morally nicer, but because it becomes cybernetically viable.