The Sovereign Shareholder Framework: Restructuring the UK and EU Demand Circuit for the Compute Age

Doctrine in one sentence

To preserve mass economic agency against cognitive labour substitution, we must shift the fiscal base from taxing human employment to capturing compute rents and building shared sovereign equity, ensuring every citizen holds a non-arbitrable stake in automated productivity.

Executive summary

The postwar economic model of the United Kingdom and the European Union relies on a self-reinforcing wage-demand circuit. In this circuit, human labour produces wages, wages finance household consumption, consumption generates corporate revenue, and corporate revenue sustains the demand for further human labour. The Unit Cost Dominance thesis introduces a profound structural challenge to this historical arrangement. If artificial intelligence systems, combined with digital verification protocols, can execute a broad range of cognitive workflows at a lower unit cost than human-only production, the market mechanism will naturally bypass human workers. This does not imply the immediate or total disappearance of all employment, but it indicates that wage labour may cease to be the primary route to independent economic agency for a significant portion of the population.

Faced with this contingency, governments must design policies that do not rely on futile or economically damaging measures. It is impossible to preserve commercially redundant jobs, accurately measure the exact AI share of individual workflows, or coordinate a global ban on automation. Instead, the UK and the EU require a domestic policy framework that is adoption-compatible, administratively auditable, and institutionally durable.

This proposal presents the Sovereign Shareholder Framework. The framework secures public revenue and household demand not by taxing the process of automation itself, but by capturing its physical inputs and structural outputs. The strategy rests on two main components. First, it taxes the physical bottlenecks of digital automation, specifically high-performance compute hardware and data centre energy consumption. Second, it converts these revenues into public equity stakes in automated industries, distributing the returns directly to citizens as a technology dividend.

Importantly, the framework is built on a principle of bounded regret. If the Unit Cost Dominance thesis proves incorrect, and AI adoption merely leads to standard productivity gains without mass labour displacement, the policies proposed here remain highly beneficial. Reducing payroll taxes, expanding public wealth funds, capping monopolistic digital rents, and investing in green compute infrastructure will enhance national competitiveness, improve public finances, and reduce inequality under any technological scenario.

The policy package

The Sovereign Shareholder Framework consists of six integrated policy planks designed to operate across the UK, the EU, and individual member states.

The first plank is the Compute Royalty and Cloud Infrastructure Levy. Rather than attempting to tax nebulous algorithms or self-reported AI software use, this levy targets the physical, auditable bottleneck of modern automation: high-performance silicon imports and commercial data centres exceeding a capacity of 5 Megawatts. It imposes a royalty per teraflop-hour of compute generated or consumed within the jurisdiction, alongside an infrastructure levy on data centre electricity usage. A generous free-use allowance protects domestic startups, academic research, and small businesses.

The second plank is the Sovereign Technology Fund. Revenues from the Compute Royalty, combined with direct public capital allocations, are directed into a dedicated sovereign wealth fund. In the UK, this is integrated into an expanded National Wealth Fund; in the EU, it operates as a coordinated network of national wealth funds. The fund is mandated to acquire equity and convertible debt in technology, infrastructure, and automated service firms. Instead of subsidising private capital accumulation through simple grants, public support for compute infrastructure is conditioned on receiving corporate equity.

The third plank is the Universal Resilience Grant. The grant is a dormant, non-means-tested transfer to all adult citizens, designed to distribute the returns of the wealth funds and the compute levy. Unlike standard welfare, it is not subject to intrusive job-seeking requirements or low asset limits. It acts as a direct dividend, preserving consumer demand and household agency. Crucially, the grant remains dormant or at a nominal level until specific macroeconomic triggers regarding cognitive labour wages and employment are met.

The fourth plank is Public Compute and the Open Data Commons. To prevent the monopolisation of AI tools by a handful of global firms, the state establishes a public compute reserve. Domestic firms and public services receive subsidised compute access in exchange for contributing non-proprietary data to an open data commons. This commons operates under strict privacy-preserving rules, ensuring that the raw materials of machine learning are treated as public goods rather than private monopolies.

The fifth plank is Fiscal Payroll Rebalancing. To offset the rising cost of capital in automated systems, governments aggressively reduce the cost of human labour. This is achieved by cutting employer payroll taxes (National Insurance in the UK, and social security contributions in EU member states) for low-to-middle income workers. This fiscal shift is funded directly by the compute levy and corporate wealth taxation, making human employment structurally more competitive.

The sixth plank is the implementation of Rent Shields. A major vulnerability of unconditional transfers is that they can be absorbed by rent-seeking monopolists in non-substitutable sectors. To prevent tech dividends from merely inflating housing costs, energy bills, or platform fees, the package includes strict local land-value taxation, expanded public energy generation, and digital platform interoperability mandates that limit lock-in rents.

United Kingdom: first 24 months

In the first 24 months, the UK government will focus on building the institutional foundations, establishing the tax bases, and reforming welfare rules to prepare for potential structural displacement.

The government will pass the Cloud Infrastructure and Compute Royalty Act. This will establish a physical register of all data centres operating above 5 Megawatts and mandate the installation of independent, auditable compute-metering software. During this initial phase, the levy will run at a shadow rate of 0% to allow for system testing, data collection, and software integration, avoiding immediate market disruption. This register will be maintained by the Department for Science, Innovation and Technology, ensuring that the tax base is verified and secure before any revenues are collected.

Simultaneously, the government will expand the mandate of the National Wealth Fund. Its existing £27.8 billion capacity will be leveraged to establish the Sovereign Technology Fund division. The fund will be legally empowered to take direct equity stakes and convertible warrants in any firm receiving sovereign compute subsidies, hardware support, or public procurement contracts in the technology sector. This ensures that public investment directly builds a portfolio of automated assets for the public.

To protect workers during the initial phases of cognitive automation, the government will reform Universal Credit rules. The current capital limit of £16,000, which disqualifies displaced workers with modest savings from receiving support, will be suspended for 24 months for individuals who can demonstrate displacement from cognitive occupations. This allows transitionally unemployed workers to retain their savings while they retrain or build independent freelance portfolios, preserving their economic agency.

To monitor the transition, the government will establish the Cognitive Labour Observatory. Operating as an independent unit within the Office for National Statistics, the observatory will monitor real-time indicators of labour displacement. It will track payroll data, hiring vacancy rates, and starting salaries in highly exposed occupations such as software development, customer support, and administrative law. This data will serve as the evidentiary basis for activating later fiscal triggers.

Finally, the government will introduce an initial payroll tax cut. The employer National Insurance contribution rate will be reduced by 1.5 percentage points for employees earning below the median wage. This initial cut will be funded by a new 5% surcharge on the operating margins of hyperscale cloud providers within the UK, acting as an interim funding bridge before the compute royalty is fully activated.

European Union and member states: first 24 months

Because direct taxation at the EU level requires unanimity and the Union budget cannot run a deficit, the EU and its member states must adopt a coordinated, multi-speed approach.

The European Commission will use the compliance and registration database established under the EU AI Act to create a harmonised register of high-compute systems and deployment scales. Member states will use this register as the administrative base for national compute levies, ensuring that auditing standards are consistent across the single market. This avoids the creation of internal regulatory barriers while establishing a robust tax base.

The Commission will also reform the terms of its €200 billion AI investment initiative. Any funding allocated for hardware, including the €20 billion earmarked for compute gigafactories, must be structured as co-investment. In exchange for public funding, member states or a coordinated EU investment vehicle must receive equity shares or revenue-sharing agreements. This ensures that public funds are not used to subsidise private monopolies without return.

Member states will use their fiscal sovereignty to reduce social security contributions for low-wage workers. To fund this without violating the Stability and Growth Pact, member states will introduce national Digital Infrastructure Excise Duties on large data centres, modeled on existing environmental excise taxes. These duties will target the physical power connections of the data centres, providing a stable and non-geographic tax base.

Under the European Semester framework, the Commission will issue guidelines for member states to establish National Tech Dividend Funds. These funds will pool the equity stakes acquired through public investment and digital excise duties. The Commission will coordinate these efforts to ensure that cross-border technology firms cannot exploit regulatory arbitrage between member states, preserving the integrity of the single market.

Years 3 to 5 and dormant triggers

During years 3 to 5, the transition from preparation to active intervention will be governed by objective macroeconomic indicators rather than political discretion.

The Cognitive Labour Observatory in the UK and Eurostat in the EU will monitor median real hourly wages in designated cognitive-exposed sectors. If real wages in these occupations fall by 5% or more over a rolling 18-month period, or if total employment in these sectors drops by 8% without a corresponding rise in general employment, the dormant triggers will fire. This double-key trigger ensures that action is only taken in response to verified structural displacement.

Upon trigger activation, the Universal Resilience Grant will transition from a nominal registry to an active cash transfer. In the UK, it will commence at £150 per month for every adult citizen, paid directly to bank accounts via the HM Revenue and Customs tax platform. In the EU, member states will deploy equivalent national grants, calibrated to local living costs and funded by their national tech dividend funds. This transfer is designed to be permanent and non-arbitrary, establishing a new floor for household income.

Simultaneously, the compute royalty will be activated at a rate of £0.05 (or €0.06) per teraflop-hour for commercial compute operations exceeding a threshold of 100 million teraflops per year. This rate will scale dynamically, increasing if the displacement indicators continue to deteriorate, thereby raising the necessary revenue to fund higher grant payouts. The threshold ensures that early-stage startups and academic researchers remain exempt.

The government will also deploy the Public Compute Reserve. Startups and cooperative enterprises that pledge their data to the national open data commons will receive free compute credits from the state-owned compute reserve. This prevents hyperscale providers from using compute costs to lock in downstream developers and suppress competition, ensuring that the local technology ecosystem remains dynamic and open.

Funding and fiscal arithmetic

To maintain credibility, the fiscal design of the Sovereign Shareholder Framework must be transparent about its limitations and scale.

It is crucial to state clearly that the arithmetic required to fund a full, immediate, unconditional basic income sufficient to cover all living costs is missing. For example, providing a universal income of £10,000 per year to the UK’s 53 million adults would require £530 billion annually, representing over 20% of GDP and nearly half of total public spending. Funding this immediately through debt is impossible given that UK public debt is close to 95% of GDP, and funding it through income tax would destroy the remaining wage-demand circuit.

The Sovereign Shareholder Framework resolves this by funding a targeted, trigger-activated grant rather than an immediate full income. The initial grant of £150 per month (£1,800 per year) requires approximately £95.4 billion annually in the UK. This is funded through a combination of structural sources:

First, the Compute Royalty is estimated to raise £18 billion to £25 billion annually at scale, depending on the growth rate of machine learning workloads. Second, a 10% tax on the domestic margins of cloud providers and digital platforms raises approximately £8 billion annually. Third, by year 5, the equity stakes held by the National Wealth Fund are projected to yield £12 billion annually in dividends and capital gains, assuming co-investment in high-growth automation firms.

Fourth, the reduction in employer National Insurance is offset by raising the corporation tax rate on highly automated firms (those with low payroll-to-revenue ratios) by 3 percentage points, yielding £14 billion. Fifth, the grant is integrated with existing means-tested systems. For individuals receiving Universal Credit, the grant replaces the first £150 of their standard allowance, reducing the net new cost to the Treasury by approximately £35 billion.

This leaves a remaining funding gap of approximately £10.4 billion. This gap must be modelled by the Office for Budget Responsibility to assess its long-term impact on public debt, dynamic growth, and consumption tax revenues. The proposal does not claim to fund mass agency overnight; instead, it establishes the fiscal mechanisms that scale alongside the automation of the economy.

In the EU, because the central budget cannot run a deficit, member states must balance their national resilience grants locally. Member states with large digital footprints (such as Ireland, the Netherlands, and Germany) will generate significant revenue from digital excise duties, while other member states will rely more heavily on coordinated EU structural funds redirected towards public equity acquisition.

Political coalition and public case

A policy package of this scale can only succeed if it is supported by a broad, durable political coalition and presented in clear, popular language.

The political coalition will be built around four main groups. First, young and entry-level cognitive workers who face a rapid decline in entry-level roles will form the activist base of the coalition. Second, trade unions will support the framework as a way to secure independent income and collective ownership of automated assets, rather than fighting losing battles to preserve obsolete jobs. Third, small and medium enterprises will support the package because the payroll tax cuts make human employment cheaper, while public compute credits lower their technology costs. Fourth, fiscal conservatives and state administrators will support the triggers because they prevent premature public spending, while the equity-capture model provides a non-debt-based route to fiscal stability.

The framework will be presented to the public under the slogan "Your Dividend from the Machines." The narrative will emphasise that artificial intelligence is built on public foundations: public-funded research, public infrastructure, and the collective data of all citizens. Therefore, the profits of automation do not belong solely to foreign cloud providers; they must be shared with the public.

The primary losers under this framework will be multinational hyperscale cloud providers and highly automated technology firms, which will face higher taxes and equity-sharing requirements. To prevent capital flight, these firms will be allowed to offset up to 50% of their compute royalty liability by investing in domestic green energy generation (such as offshore wind or nuclear) dedicated to powering their data centres. This accelerates the green transition while maintaining local infrastructure investment.

Durability and anti-capture design

To remain effective over decades, the Sovereign Shareholder Framework must resist corporate lobbying, tax avoidance, capital flight, and political interference.

If firms attempt to avoid the compute royalty by hosting their compute outside the UK or EU, governments will deploy a Compute Border Adjustment Mechanism. This will impose a tariff on imported digital services based on the estimated computational power used to train the models or process the workflows. The tariff will match the domestic royalty rate, ensuring that offshoring compute yields no tax advantage.

Taxing software or intellectual property is notoriously difficult due to transfer pricing and shell companies. The framework bypasses this by taxing physical, non-geographic inputs: grid-connected power lines, physical land footprints of data centres, and silicon hardware imports at customs. These physical assets cannot be moved to offshore tax havens without significant capital cost.

To prevent future hostile governments from raiding the Sovereign Technology Fund or using its capital for political patronage, the fund will be established under an independent statutory trust. Its governance will be modeled on the BBC Trust or the Bank of England, requiring a two-thirds majority in Parliament (or equivalent constitutional safeguards in EU member states) to alter its investment mandate or withdraw capital.

Finally, the administration of the open data commons and public compute credits will be managed by regional boards comprising local government, business representatives, and academic institutions, preventing the concentration of administrative power in a single central bureaucracy.

Legal and institutional obstacles

Implementing the framework requires navigating several legal frameworks in both jurisdictions.

Under Article 123 of the Treaty on the Functioning of the European Union, the European Central Bank and national central banks are strictly prohibited from providing overdrafts or any other type of credit facility to public authorities. This means the resilience grants cannot be funded through monetary financing or central bank money creation. The grant must be funded strictly through fiscal revenues, corporate taxation, and wealth fund dividends.

Under Article 113 and 115 of the Treaty on the Functioning of the European Union, any direct EU-wide tax directive requires the unanimous consent of all member states. To bypass this obstacle, the compute royalty will not be introduced as an EU-wide tax. Instead, it will be designed as a national excise duty on data centre infrastructure, which member states can implement individually under their existing sovereign tax powers. The EU will focus on coordinating these taxes through soft-law recommendations and the European Semester.

The UK Treasury is constrained by fiscal rules requiring public debt to fall as a share of GDP within five years. Because the Sovereign Technology Fund’s equity investments are classified as financial assets, they do not increase net public debt in the same way as standard department spending. However, the Office for Budget Responsibility does not currently account for the positive demand-side effects of cash transfers in its long-term growth forecasts. The government must direct the office to develop dynamic scoring models that account for the macroeconomic stabilisation provided by the grant.

The Compute Border Adjustment Mechanism could face challenges under the General Agreement on Trade in Services and the General Agreement on Tariffs and Trade. To ensure compliance, the tariff must be structured strictly as an environmental and infrastructure charge, designed to offset the carbon footprint and local grid costs of foreign data centres, rather than as a discriminatory trade barrier.

Failure modes, review and exit rules

A robust policy must include clear mechanisms to scale down or terminate its components if the underlying assumptions prove incorrect.

The first failure mode is technological stagnation. If AI adoption slows, and the predicted productivity gains and labour displacement do not materialise, the policy risks taxing a nascent industry without purpose. If the Cognitive Labour Observatory reports that real wages in cognitive-exposed occupations increase by more than 2% annually over a three-year period, and aggregate productivity growth remains below 1%, the compute royalty rate will be frozen or reduced by 50%. The Sovereign Technology Fund will shift its investment mandate away from automated industries and towards traditional infrastructure and green technologies.

The second failure mode is an inflationary spiral. If the resilience grant is activated but domestic capacity constraints remain tight, the cash transfer could fuel demand-pull inflation, reducing the real purchasing power of the grant. If CPI inflation exceeds 4% for more than two consecutive quarters and is directly linked to consumer spending, the grant will not be increased. Instead, the cash transfer will be partially converted into non-transferable service vouchers for essential needs, such as public transport, municipal housing, or household energy, which are directly managed and price-capped by the state.

The third failure mode is capital flight and compute depletion. If the compute royalty leads to a significant migration of data centre infrastructure out of the UK and EU, domestic technology ecosystems could suffer. A mandatory review will be triggered if domestic data centre capacity growth falls below the OECD average for two consecutive years. In this event, the government will increase the free-use compute allowance for domestic firms and expand the royalty offset for local green energy investments.

Feasibility table

The following table assesses the feasibility, timeline, blockers, and bounded-regret value of the core policy planks.

To clarify the ratings in the table:

For the Compute Royalty and Cloud Infrastructure Levy, UK feasibility is rated Medium because the UK can pass this via a Finance Bill, but it faces intense lobbying from global tech firms. EU-level feasibility is rated Low because achieving the required unanimity for an EU-wide tax directive is highly unlikely. Member-state feasibility is rated High because member states have full competence to introduce local excise duties on data centre infrastructure. The bounded-regret value is rated High because even if AI adoption slows, this levy drives data centre energy efficiency and supports the green transition.

For the Sovereign Tech Wealth Funds, UK feasibility is rated High because the existing National Wealth Fund provides an immediate, functional vehicle. EU-level feasibility is rated Medium because it requires coordination of state aid rules to allow targeted equity investments. Member-state feasibility is rated High because most member states already operate national investment vehicles that can be redirected. The bounded-regret value is rated High because it builds long-term public assets regardless of technological outcomes.

For the Universal Resilience Grant, UK feasibility is rated Medium because it requires substantial fiscal space and welfare restructuring, but is administratively simple. EU-level feasibility is rated Low because the EU has no legal mandate to distribute direct income transfers. Member-state feasibility is rated Medium because it is feasible within national welfare budgets but constrained by fiscal rules. The bounded-regret value is rated Low because if displacement does not occur, a permanent cash transfer is an inefficient use of public funds, which is why the grant remains dormant until triggered.

For the Payroll Tax Cuts, UK feasibility is rated High because it is a standard fiscal adjustment tool for the Chancellor. EU-level feasibility is rated Medium because it is coordinated via soft-law recommendations but remains a national power. Member-state feasibility is rated High because it is a direct lever for national governments to boost competitiveness. The bounded-regret value is rated High because it stimulates employment and reduces inequality under any economic scenario.

For the Rent Shields, UK feasibility is rated Medium because it requires overcoming strong political resistance from property owners. EU-level feasibility is rated Low because housing and land use are national competencies. Member-state feasibility is rated Medium because it is politically difficult but legally within member state powers. The bounded-regret value is rated High because reducing housing and platform rents improves economic efficiency and disposable income under any technology scenario.

For the Compute Border Adjustment Mechanism, UK feasibility is rated Low because the UK lacks the trade leverage to enforce digital border adjustments unilaterally without facing retaliation. EU-level feasibility is rated Medium because the EU has trade competence, but auditing foreign compute is technically and legally complex. Member-state feasibility is rated Low because member states do not have independent trade policy competence under the single market. The bounded-regret value is rated Medium because it protects domestic industries but risks trade friction.

What is genuinely new here

The Sovereign Shareholder Framework departs from traditional policy responses to automation in three key ways.

First, it shifts from algorithmic to physical taxation. Traditional proposals for robot taxes or AI levies fail because they rely on defining what constitutes a robot or an AI workflow. The framework abandons this approach. It treats compute as a utility, like electricity or water, and taxes its physical consumption (flops and megawatts). This provides an objective, auditable tax base that cannot be bypassed by changing software definitions.

Second, it introduces triggered universalism. Rather than launching an immediate, expensive universal basic income that would strain public finances and fuel inflation, the framework introduces a dormant grant. This preserves fiscal resources during the transitional phase, activating only when verified macroeconomic indicators demonstrate that human wages are losing unit-cost competitiveness.

Third, it implements public co-investment. Instead of relying on corporate taxation after profits have been distributed, the framework uses public compute resources and infrastructure subsidies to acquire direct equity stakes in technology firms. This ensures that as capital takes a larger share of national income, the public directly owns a portion of that capital, automatically receiving its returns.

Bottom line

If cognitive automation permanently breaks the wage-demand circuit, we cannot save the old economic model by subsidising obsolete work. We must transition to a system where productivity gains are shared directly with citizens. The Sovereign Shareholder Framework provides a realistic, funded, and durable path to achieve this. By taxing physical compute infrastructure and acquiring public equity, the UK and EU can secure household demand, maintain state capacity, and ensure that the wealth of the machine age belongs to the society that created it.