Cost Anchors and Shared Stakes: Securing Mass Economic Agency

Doctrine in one sentence

Mass economic agency is preserved by anchoring the unit costs of housing and energy through supply expansion and competition, capturing observable public returns from AI-related infrastructure investment for broad distribution or service provision, and maintaining tax, competition and welfare rules on auditable bases that function regardless of the precise rate of AI substitution in any workflow.

Executive summary

The Unit Cost Dominance thesis identifies a conditional risk: if AI plus verification delivers cognitive output at materially lower unit cost across wide task ranges, the postwar wage-demand circuit weakens without any requirement for total job disappearance or coordinated global restraint. The policy package responds with measures that remain effective under continued or accelerated adoption while delivering ordinary public benefits if displacement proves limited.

Seven interlocking planks are proposed. They rest on observable administrative bases: energy consumption or revenue thresholds for taxation, planning permissions and public land for housing, existing investment vehicles and procurement contracts for public stakes, and competition law for interoperability. No plank requires governments to measure the AI share of individual workflows or to preserve commercially unnecessary employment.

In the United Kingdom the first 24 months focus on repurposing the National Wealth Fund, accelerating planning reform already in train, prioritising Competition and Markets Authority cases in AI markets, and adjusting Universal Credit capital rules plus modest reallocation within existing AI compute budgets. The European Union and member states use the InvestAI facility and AI Factories to embed return conditions, enforce high-risk provisions of the AI Act in employment contexts, and coordinate competition and state-aid-compliant investment. Years 3 to 5 activate dormant fiscal or distributional triggers only if independent statistical indicators cross pre-specified thresholds.

Fiscal arithmetic is bounded. Existing capacity in the National Wealth Fund (£27.8 billion) and EU mobilisation targets (€200 billion for AI including €20 billion for gigafactories) can be redirected toward equity or revenue-share structures whose returns support either small citizen-level payments or expanded public services. Modest broadening of the UK Digital Services Tax base or a 1 percentage point corporation tax adjustment offers further headroom, but precise yields require Office for Budget Responsibility scoring. The package cannot by itself fund large-scale unconditional income replacement; it augments real purchasing power through lower essential costs and limited upside sharing.

Distribution is addressed on the cost side (housing and energy absorb a large share of lower- and middle-income budgets) and through competition that limits platform rents. Durability features include statutory review cycles, independent statistical triggers, anti-avoidance rules attached to observable bases, and dispersal of administrative power across UK Parliament, EU institutions and member states. Political acceptability rests on a growth-plus-distribution coalition that can point to immediate benefits for households and businesses alongside insurance against contingent risk.

The design passes the ten tests: it is adoption-compatible, uses auditable bases, respects jurisdictional competences, identifies payers and modelling gaps, explains agency channels, incorporates anti-capture provisions, identifies plausible sequencing and compensation, distinguishes time horizons, supplies falsifiable indicators, and introduces mechanisms not reducible to simple transfers or job subsidies.

The policy package

The package comprises seven planks. Each attaches to an observable administrative object and carries positive value under baseline economic conditions.

Public AI and digital infrastructure vehicles (National Wealth Fund in the UK; InvestAI and AI Factories in the EU) are re-mandated to take equity stakes, revenue shares or convertible instruments whose financial returns are earmarked for either small annual or periodic citizen-level distributions or equivalent service expansions (broadband, training credits, energy rebates). Returns are not conditioned on measuring AI use in private workflows.
Statutory housing supply acceleration through further planning reform, public land release, and infrastructure coordination targets a sustained increase in completions. Lower housing costs directly increase real disposable income for the median household irrespective of employment status.
Energy and grid investment, including accelerated renewables, storage and interconnection, caps and reduces unit energy costs for households and for compute-intensive activity. Because AI training and inference are energy-intensive, public supply expansion limits the extent to which energy rents capture productivity gains.
Competition and interoperability mandates, using existing CMA and EU Digital Markets Act/competition powers, require data portability, model interface access and non-discriminatory access terms for foundation models and key platforms above defined turnover or user thresholds. This reduces lock-in rents and supports new entrants without requiring any AI-specific levy.
Public procurement conditions attach transparency, auditability, human oversight where safety or rights are engaged, skills transition clauses and public data contribution requirements to contracts above defined values. These conditions are contract-specific, verifiable and non-discriminatory.
Welfare adjustments reform capital limits in means-tested benefits and create portable transition accounts or credits funded from general revenue or reallocated AI budgets. Accounts support retraining, relocation or self-employment without requiring proof of AI displacement.
Tax neutrality and base-broadening reviews examine interactions between full expensing for plant and machinery and large-scale AI capital expenditure, alongside modest parameter adjustments to the existing Digital Services Tax (currently 2 per cent on qualifying UK revenues of large search, social and marketplace services above thresholds). Any changes use revenue or energy-consumption thresholds that are already reported or auditable.

These planks are mutually reinforcing: lower housing and energy costs increase the real value of any wage or transfer income; competition and public stakes limit rent extraction; procurement and welfare adjustments support adjustment without freezing existing occupational structures.

United Kingdom: first 24 months

Amend the National Wealth Fund framework document and strategic priorities to require that a defined share of new digital and technology investments take forms that generate attributable public returns (equity, royalties or profit shares). Design a distribution mechanism (small annual credit or service top-up) for consultation and pilot in year 2. Use part of the existing £27.8 billion capacity and any new capitalisation.
Deliver the housing targets embedded in the Planning and Infrastructure Act 2025 and updated standard method (approximately 370,000 net additional homes per year nationally). Release further public land, streamline approvals for high-supply areas, and link infrastructure funding explicitly to affordability outcomes.
Instruct the Competition and Markets Authority to prioritise AI platform and foundation-model cases with a focus on interoperability and data access remedies under the existing digital competition regime.
Issue guidance and consult on Universal Credit capital rules to reduce the sharpness of the £16,000 limit and taper for transitional cohorts, funded within existing departmental expenditure limits or by modest reallocation from AI compute budgets.
Implement the Compute Roadmap with explicit public-stake conditions on any new sovereign or co-funded facilities. Begin collection of energy and capacity data from large data-centre operators to support future observable bases.
Commission joint Treasury-OBR modelling of revenue effects from a 1 percentage point corporation tax adjustment and from parameter changes to the Digital Services Tax, with publication before the end of the period.
Establish an independent statistical monitoring panel (building on ONS and DSIT series) to track AI adoption, headcount change in exposed occupations, real household disposable income, housing affordability and energy prices.

European Union and member states: first 24 months

Adapt InvestAI governance and the AI Factories programme to require that a share of public or joint funding takes equity or revenue-share forms whose returns flow to member states for use in citizen-level measures or services. Nineteen AI Factories and associated antennas are already operational or selected; new gigafactory facilities provide a natural insertion point.
Accelerate enforcement of the AI Act high-risk obligations in employment and worker-management contexts (recruitment, performance evaluation, task allocation and termination decisions) from the applicable dates, with guidance emphasising human oversight and documentation rather than prohibition.
Issue Commission guidance and support member-state action under the Digital Markets Act and competition rules to require interoperability and data access for AI systems meeting turnover or user thresholds.
Coordinate use of existing European Social Fund, Just Transition and recovery funds for portable skills and transition support, without creating new permanent EU-level income entitlements. Member states retain primary responsibility for recurrent transfers.
Require state-aid notifications for national AI compute or gigafactory support to include public-return conditions consistent with single-market rules.
Begin technical work on observable bases (energy consumption reporting, large-model registration thresholds) that could support future fiscal measures if triggers are met, while respecting Article 113 and unanimity requirements for new own resources.
Establish EU-level monitoring indicators drawing on Eurostat AI use data and labour-force statistics, with annual public reporting.

Years 3 to 5 and dormant triggers

Dormant provisions activate only on evidence of sustained pressure on the wage-demand circuit. Indicative triggers, to be refined by independent statistical bodies, could include:

ONS or Eurostat data showing AI adoption above 40 per cent of businesses combined with headcount reduction in AI-exposed occupations exceeding the economy-wide average by a defined margin for two consecutive years, or
real median household disposable income failing to rise in line with measured productivity growth in cognitive-intensive sectors for a sustained period, or
housing or energy costs as a share of median disposable income rising despite productivity gains.

Activation could scale the size of public-return distributions, authorise further tax-parameter changes, or expand procurement conditions. Conversely, if employment in exposed occupations and real wage growth track productivity, the fiscal and distributional elements can be held at baseline or scaled back through scheduled review.

Statutory reviews occur at 36 and 60 months with published OBR-style scoring. Provisions that prove unnecessary or counterproductive can be repealed by primary legislation (UK) or ordinary legislative procedure where competence exists (EU).

Funding and fiscal arithmetic

The United Kingdom possesses immediate levers: reallocation within the National Wealth Fund and existing AI compute allocations (£500 million Sovereign AI Unit envelope and supercomputer commitments), modest parameter adjustment to the Digital Services Tax (which raised approximately £944 million in 2025-26), and corporation tax at 25 per cent main rate. A 1 percentage point corporation tax increase would require OBR modelling; historical yields suggest several billion pounds annually, but precise interaction with full expensing and behavioural response must be scored before enactment. Public debt near 95 per cent of GDP constrains large new borrowing.

The European Union cannot run a budget deficit and new own resources normally require unanimity. InvestAI and EuroHPC resources (overall supercomputing and AI Factory investment reaching €10 billion in the 2021-2027 period plus the €200 billion mobilisation target) provide the primary vehicle. Member states can adjust national tax and transfer parameters more rapidly.

The package does not claim to fund full income replacement for a large displaced workforce. Its arithmetic rests on cost reduction (housing and energy can represent 30-40 per cent of lower-income budgets) plus limited upside capture. Detailed dynamic scoring of interactions with existing welfare, housing benefit and energy support is required before legislation. Where arithmetic is missing, the proposal states the gap rather than assuming closure.

Political coalition and public case

A viable coalition spans growth-oriented Conservatives and Labour in the UK, and Christian Democrat, liberal and social-democratic groupings in the EU and member states. Business organisations can support infrastructure investment and competition that prevents dominant-platform lock-in. Households gain from lower housing and energy costs. Trade unions and centre-left parties gain portable support and public stakes without defence of every existing job.

The ordinary-language public case is straightforward: AI that lowers production costs should also lower the cost of living and give the public a direct stake in the infrastructure it helps finance. The alternative is that concentrated owners of platforms, land and energy capture the gains while essential costs remain high.

Losers are owners of scarce land and concentrated digital platforms, plus any firms whose tax treatment is adjusted. Compensation occurs through grandfathering of existing investments, transition periods, and the fact that lower input costs benefit most businesses. Sequencing begins with low-cost, high-visibility actions (planning delivery, competition cases, procurement guidance) before fiscal elements.

Durability and anti-capture design

Capture is addressed by dispersing power: UK primary legislation can be altered by future parliaments; EU measures rely on internal-market and competition competences that are harder to unwind unilaterally. Independent statistical triggers reduce ministerial discretion. Procurement conditions are published and challengeable. Tax bases use thresholds already administered by HMRC or equivalent.

Avoidance is limited by focusing on turnover, energy consumption or planning permissions rather than self-reported AI intensity. Offshoring is countered by consumption-side measures (housing, energy) and by competition rules that apply to services used in the jurisdiction. Hostile future governments face statutory review clauses and the political cost of repealing visible cost-reduction measures. Concentration of administrative power is avoided by using existing regulators (CMA, national competition authorities, procurement bodies) rather than creating a single AI income administrator.

Legal and institutional obstacles

United Kingdom: primary legislation for National Wealth Fund mandate changes, planning adjustments and tax parameters is straightforward. Universal Credit changes are within secondary or primary powers. No major Treaty constraint.

European Union: direct-tax and new own-resource decisions require unanimity. Competition and internal-market measures, state-aid frameworks and use of existing funds are more accessible. Procurement conditions must relate to the subject matter of the contract and be non-discriminatory. AI Act high-risk obligations are already legislated; implementation dates are known.

Member states retain primary tax and welfare competence and can move faster on recurrent support. State-aid rules constrain selective national subsidies but permit general measures and public-investment vehicles when properly structured.

Failure modes, review and exit rules

Principal risks: slower-than-expected adoption renders public-return mechanisms small while cost-reduction measures remain valuable; faster adoption overwhelms the scale of returns that can be captured from £27.8 billion or €20 billion vehicles; political reversal before triggers mature; successful lobbying that weakens competition or procurement conditions; measurement disputes over trigger definitions.

Review points are statutory at 36 and 60 months with published data and OBR-style assessment. Exit or scaling-down rules: if independent monitors report that real median disposable income and employment in exposed occupations are rising at or above productivity growth for two review periods, fiscal planks may be held constant or reduced by order subject to affirmative resolution. If housing completions meet targets and affordability indices improve, supply-side effort can be rebalanced toward maintenance rather than acceleration. Triggers are published in advance and revisable only with independent statistical sign-off.

Feasibility table

Public AI infrastructure returns (NWF/InvestAI equity and revenue shares) | High – existing vehicle and remit already include digital and technology; framework amendment is administrative and legislative | Medium – requires governance adaptation of InvestAI and EuroHPC structures and member-state agreement on return earmarking | High – parallel national co-financing or complementary vehicles straightforward | 12-24 months for mandate and pilot design | Agreement on distribution mechanism (dividend versus services) and additionality rules | Strengthens public balance sheets and catalyses growth investment regardless of displacement scale

Housing supply acceleration | High – Planning and Infrastructure Act 2025 and updated standard method already provide vehicles; delivery is the issue | Low-direct – limited EU competence; support via state aid and funds only | High – national planning and land policy dominant | Immediate continuation of existing reforms | Local political resistance and construction-sector capacity | Directly reduces largest non-discretionary cost for most households; benefits accrue under any productivity path

Energy and grid investment | High – existing policy and regulatory frameworks; National Wealth Fund remit covers clean energy | Medium-High – state-aid compliant investment and TEN-E frameworks available | High – national energy policy and permitting | 12-36 months for major projects | Grid connection queues and planning for transmission | Caps energy costs for households and AI compute; improves security and decarbonisation irrespective of AI outcomes

Competition and interoperability mandates | High – CMA digital regime and DMCC Act already operational | High – DMA and competition rules in force; guidance can be issued rapidly | Medium – national authorities implement but single-market dimension constrains unilateral action | 6-18 months for prioritised cases and guidance | Evidence thresholds and appeals in complex markets | Reduces lock-in and rents; promotes entry and lower prices under baseline competition policy

Public procurement conditions | High – central and devolved procurement already used for policy goals; transparency conditions are standard | Medium-High – must be contract-related and non-discriminatory under EU rules | High – national procurement rules can go further within EU constraints | 6-12 months for revised guidance and model clauses | Risk of legal challenge if conditions stray beyond subject matter | Builds transparency and skills spillovers; improves public purchasing efficiency regardless

Welfare capital and transition adjustments | High – Universal Credit parameters changeable by Parliament or regulation | Low for EU-level entitlement; Medium for coordination of existing funds | High – member states control core benefits | 12-24 months for consultation and phased change | Fiscal cost and interaction with work incentives | Improves transition support and reduces arbitrary cliffs; valuable even if displacement is modest

Tax neutrality and base broadening (DST parameters, capital-allowance review) | High – Parliament controls rates and bases; DST already administered on observable revenue | Low – unanimity for new own resources or major harmonisation | High – national tax policy | 18-36 months for review and any legislation | Behavioural response and international tax coordination | Broadens revenue base and corrects distortions; any adjustment has conventional fiscal value

What is genuinely new here

The package does not restate the diagnosis or default to job subsidies, reduced working time or unconditional transfers as automatic remedies. It prioritises the cost side of household budgets (housing and energy) because these are large, observable and amenable to supply policy. It uses public infrastructure investment to capture upside through equity and revenue shares rather than attempting to tax an unmeasurable AI intensity. It treats competition and data access as distribution mechanisms that lower prices and entry barriers without requiring fiscal transfers. Triggers and reviews are attached to published statistical series rather than discretionary ministerial judgement. Every element is designed to be useful if the thesis proves overstated and to be scalable, within fiscal and legal bounds, if pressure on the wage circuit materialises.

Bottom line

The proposal offers a bounded-regret route to preserving mass economic agency. It works with adoption rather than against it, uses only observable administrative objects, respects the differing legal capacities of the United Kingdom, EU institutions and member states, and identifies payers and modelling requirements explicitly. It cannot by itself replace the scale of the wage-demand circuit if unit-cost dominance proves very broad and rapid. It can, however, lower the cost of living, share some of the returns from public-backed infrastructure, and maintain competitive pressure that passes productivity gains to households and new entrants. These steps remain valuable under a wide range of AI outcomes and can be adjusted or exited on the basis of evidence rather than ideology. Detailed fiscal and distributional modelling by the Office for Budget Responsibility, equivalent EU bodies and national finance ministries is the necessary next step before legislation.