The Central Question

Everything in the Liquid Labor thesis leads to one question: who owns the autonomous labor stock?

In On Depreciation, we proved that the Gini Coefficient of Labor Capital approaches 1.0—near-perfect inequality. The Depreciation Bomb and coordination costs price out small players. 90% of physical labor ends up owned by fewer than 5 entities. This is not a prediction. It is a mathematical consequence of the cost structure.

The question is whether those 5 entities are Amazon, Tesla, Foxconn, Saudi Aramco, and Alphabet—or whether they are sovereign institutions accountable to citizens.

This is not a philosophical preference. It is the single most consequential policy decision of the 21st century. Get it wrong, and you get neo-feudalism—a permanent rentier class that owns all productive capacity while citizens rent their livelihoods from corporate landlords. Get it right, and you get the first economy in history where the surplus of labor belongs to everyone.

I. The Private Ownership Failure Mode

The default trajectory—the one we’re on now—is private concentration. It follows the same pattern as cloud computing. AWS, Azure, and GCP own ~63% of global cloud infrastructure (Synergy Research, Q3 2025). Three companies. The economics of robotic fleets are identical: massive upfront CapEx, brutal depreciation, winner-take-all economies of scale. The market will consolidate to a handful of mega-operators just as cloud did.

But there’s a second concentration problem: the global one. China installed 290,000 industrial robots in 2024—54% of the global total. The U.S. installed 34,000, a 9% decline. China’s robot density reached 470 per 10,000 workers; the U.S. sits at 295. By 2024, China’s operational robot stock exceeded 2 million units. In humanoid robots, Chinese firms shipped roughly 90% of global units in 2025: AgiBot delivered 5,168 units, Unitree shipped over 4,200, UBTech added thousands more. Meanwhile, Unitree launched a humanoid robot at $5,900—U.S. competitors are asking $50K-$100K+. China has designated robotics as one of eight “strategic emerging industries” in its 15th Five-Year Plan (2026-2030), launched an $8.2B National AI Industry Investment Fund in 2026, and operates 150+ humanoid companies with 40+ state-funded training centers. Without proactive public ownership by the U.S., China’s state-backed model will dominate the global robotic labor market. We face not just private concentration domestically, but losing the international race for autonomous productive capacity entirely.

Source: IFR World Robotics 2025, CSIS ChinaPower Project. China installed 8.5x more robots than the U.S. in 2024 and shipped 90% of global humanoid units in 2025.

Under private ownership, the surplus flows to shareholders. Workers displaced by robots receive nothing unless the state intervenes with transfers. The RaaS operator sets the price of labor-hours. If you can’t afford to rent robot time, you can’t produce. The RaaS monopolist becomes the landlord of all economic activity—extracting rent from every transaction in the economy.

This is not capitalism. This is something worse. In capitalism, workers can sell their labor on an open market. In a privately-owned Liquid Labor economy, there is no labor to sell. The robots do the work. The owners collect the surplus. Everyone else negotiates for scraps.

II. Three Models of Public Ownership

Model A: The Sovereign Fleet (Norway Model)

Norway discovered oil. Instead of letting Exxon take the profits, they created the Government Pension Fund—now worth over $1.9 trillion (Norges Bank, 2025)—owned by every Norwegian citizen. The fund invests globally; the returns fund public services and a sovereign safety net. No Norwegian voted for austerity because their national wealth was being extracted by foreign shareholders.

Applied to Liquid Labor: The federal government creates a National Autonomous Workforce Corporation (NAWC)—a sovereign entity that owns and operates the national robot fleet. NAWC leases robot-hours to private companies via RaaS contracts. The revenue funds the Basic Dividend paid to every citizen. Citizens don’t own robots individually. They own the fleet collectively, the way Norwegians own their oil fund.

Model B: Municipal Robot Utilities (TVA Model)

Before the Tennessee Valley Authority, private power companies electrified profitable cities and ignored rural areas. The TVA brought public power to Appalachia—not because it was profitable, but because it was necessary. Public utilities still provide electricity, water, and sewage in most American cities. No one calls your water department “socialist.”

Applied to Liquid Labor: Cities and states create Municipal Robot Utilities (MRUs) that deploy autonomous fleets for public infrastructure—road repair, sanitation, park maintenance, building inspection, elder care. MRUs operate like public works departments: funded by taxes, operated for citizens, not for profit. Private RaaS operators handle the commercial economy; MRUs handle the commons.

Model C: Worker-Owned Robot Cooperatives (Mondragon Model)

The Mondragon Corporation in Spain is a federation of 95+ cooperatives employing over 70,000 worker-owners, with €11 billion in annual revenue (2023). Workers own the company, elect management, and share profits. It’s the 10th largest business group in Spain and the largest in the Basque Country. Not a hippie commune—a globally competitive industrial enterprise.

Applied to Liquid Labor: Workers displaced by automation don’t lose their stake—they convert it. A factory that replaces 500 workers with 500 robots becomes a Robot Cooperative where the former workers are now co-owners of the fleet. They no longer sell their hours; they rent out the robots’ hours. The income stream shifts from wages to dividends, but the workers retain ownership. The transition from human labor to robotic labor doesn’t transfer wealth upward—it transforms the form of ownership while preserving the distribution.

III. The Hybrid: A Practical Architecture

In practice, the answer is probably all three:

Tier 1 — Sovereign Fleet (Federal): A national entity (NAWC) owns the core fleet—the robots that build robots, the robots that maintain critical infrastructure, the robots that constitute the defense industrial base. This is the Von Neumann layer. It cannot be privately owned because it is the productive foundation of everything else. Revenue from leasing Tier 1 capacity funds the Basic Dividend.

Tier 2 — Municipal Utilities (State/City): Local governments operate robot fleets for public services—construction, sanitation, healthcare support, education support. Funded by local taxes plus federal NAWC revenue sharing. Accountable to mayors and city councils, not shareholders.

Tier 3 — Private & Cooperative (Market): Private RaaS operators and worker cooperatives lease robot-hours for commercial activity. They compete on efficiency, specialization, and service quality. The market still exists. Innovation still happens. But the foundational productive capacity—the Time Bank itself—is publicly held.

This mirrors how we already handle critical infrastructure. The federal government owns the interstate highway system. States maintain local roads. Private trucking companies use both. Nobody considers this “central planning.” It’s infrastructure. The robot fleet is infrastructure.

IV. The Singapore Precedent

Singapore’s Temasek Holdings manages S$484 billion (~US$339 billion) in net portfolio assets on behalf of the government (Temasek Review, March 2025). It owns stakes in Singapore Airlines, DBS Bank, Singtel, and dozens of other companies. It operates commercially—hiring professional management, demanding market returns—but profits flow to the sovereign, not to private shareholders. Singapore’s citizens benefit through world-class public services funded by Temasek’s returns.

This is the NAWC model proven at national scale. A commercially-operated sovereign entity that owns productive assets and distributes the returns to citizens. Singapore did it with airlines and banks. We would do it with robots.

The objection that “government can’t run anything efficiently” is refuted by Singapore, Norway, and South Korea’s POSCO (government-founded steel company that became globally competitive). The issue is not public vs. private. It’s competent governance vs. incompetent governance. The ownership structure determines who benefits. The management quality determines how much.

V. The Stakes

If we let the market decide ownership, the answer is already determined: 5 corporations own everything. The Depreciation Bomb guarantees it. No policy intervention can reverse hyper-centralization after it happens. You cannot nationalize a fleet of 100 million robots owned by Amazon without a constitutional crisis.

The window for establishing public ownership is now—before the fleet exists at scale. Once the infrastructure is built and the contracts are signed, the ownership structure calcifies. This is the lesson of oil, telecom, and cloud computing: the ownership regime is set in the first decade. After that, you’re negotiating with incumbents, not designing systems.

The Liquid Labor thesis is agnostic on many policy details. But on this point it is not: the autonomous labor stock is infrastructure, and infrastructure must be publicly governed. The alternative is a world where the means of production are more concentrated than at any point in human history, owned by entities with no democratic accountability, controlling the labor supply of the entire economy. That is not a market. That is a fiefdom.