When wages fall, the state collapses.
The tax system is the invisible architecture of every economic model. In the economy ahead, it becomes the sharpest point of rupture and the greatest point of leverage.
Where do things stand today, and where do they break?
Western welfare states are financed primarily from labour-linked sources. In Switzerland around 70% of AHV revenues come from wage contributions (2025: 71%), in Germany about 40% of government revenue comes from wages and social security contributions; across the OECD, labour-related taxes account for around 50% of the total. At the same time, a growing share of value creation flows to capital income and technology rents, which have historically been taxed at lower rates than labour. The gap between where value is created and where tax is collected has been widening for three decades. AI is accelerating this shift sharply.
If AI automates 20–40% of tasks in the knowledge economy (Goldman Sachs 2023, McKinsey 2024), a third of the tax base disappears before political consensus for alternatives is in place. The result: collapsing welfare systems, a shrinking middle class, social polarisation, all at once. Anyone who does not solve the funding question has no model.
Paths out of the crisis
The stars show how realistic a path is: 1 star = barely feasible, 5 stars = highly feasible.
Robot tax (Pigouvian principle: tax what causes harm)
A tax per robot, AI agent or automated task (Bill Gates 2017, South Korea since 2018).
+ Intuitively fair.
Definition problem; innovation migrates elsewhere.
Trade-off: Attractiveness as a business location declines.
Feasibility: Widely debated politically, but not implemented on any meaningful scale. The definition problem continues to block concrete legislation. South Korea remains the only country to have introduced a scaled-back version of the mechanism.
Higher Capital Gains Taxation
Value creation increasingly comes from capital. Taxation follows it there.
+ Targets the winners.
Capital is more mobile than labour. Global coordination is required.
Trade-off: Investment appetite declines under maximisation.
Feasibility: The OECD minimum tax is already in effect and is being extended in scope. A gradual increase represents the most likely path — technically and politically — over the next ten years.
Land Value Tax (Henry George)
Tax land value rather than development. AI clusters raise land values, and land does not move.
+ The most efficient tax model in theory. Singapore funds ~30% of its government budget through land-related levies.
Politically fraught (property-owner lobby), valuation problems.
Trade-off: Politically costly, economically favourable.
Feasibility: Economically the most robust model, yet politically blocked almost everywhere by property-owner lobbies. Realistic in city-states; in federal democracies, viable only as local pilot schemes.
Data / AI Dividend
Citizens as co-owners of AI infrastructure via a sovereign fund (applying the Alaska/Norway logic to AI).
+ Addresses ownership at its root.
Rarely achievable politically without a boom moment.
Trade-off: Only works if the state enters early.
Feasibility: Methodologically clear, politically sensitive: hyperscalers will fight equity claims. More likely in countries with a sovereign-wealth tradition (Norway, Singapore, Switzerland with an AHV link).
CO₂ as a key pillar
Externalities replace labour taxes. Sweden charges about €120 per tonne, and its economy has grown regardless.
+ Double dividend: climate & distribution.
Insufficient as a standalone solution.
Trade-off: A pillar, not a complete substitute.
Feasibility: EU-ETS (the European carbon emissions trading scheme) and national carbon taxes are already a reality. Scaling to €200+ per tonne sits within the range of what is politically possible. The driver is climate targets, not tax policy as such.
Trigger-based phases
Trigger: First OECD-wide majority in favour of an explicit data levy or strategic-stake mechanism on AI corporations
The political agenda shifts. First pilot legislation in Scandinavia and the EU. Lobbying activity realigns. EU ETS prices begin climbing towards €150/t. Visible in finance-ministry statements, OECD working papers and European Commission proposals.
Trigger: First G7 country establishes a sovereign AI fund with strategic stakes
Concrete implementation of the three pillars. The share of labour taxation declines measurably. Asset managers begin repositioning. Switzerland advances an AHV 2.0 reform; data levies become standard in two to three EU countries. Carbon prices reach the €200/t corridor.
Trigger: The welfare state is funded more than 50% from capital, data and externalities, no longer from labour tax
The new system is the status quo. The old wage-tax architecture is a historical artefact. Tax competition is shifting from location-based taxes to levies on capital and data. Sovereign wealth funds hold substantial stakes in AI champions.
Real-world pilot programmes
Where the ideas behind this pillar are already being tested. "Learn more" opens the explanation without leaving the page.
What we are watching
First genuine political momentum towards a capital-tax architecture. Phase-1 trigger.
Carbon becomes fiscally material. Markets are pricing in its emergence as a primary fiscal pillar.
Taboo broken. Other countries typically follow within 24 months.
Track in the index →SWFs are becoming structural buyers of AI champions. Market moves are likely to follow.
Structural reform of the AHV is visibly under way. Phase-2 trigger for Switzerland.
Our assessment
At Globalance, we regard the tax shift as the most important fiscal policy movement of the next 15 years and as unavoidable. The mathematical logic of a shrinking wage-tax base can be delayed by political means, but it cannot be overridden. Any model that leaves the financing question unanswered is no model at all.
A sustainable architecture rests on three pillars: increased Capital gains taxation with global OECD coordination, substantial CO₂ and externality levies as a dual climate-and-fiscal dividend, and an explicit Data / AI Dividend via a sovereign wealth mechanism. The Land Value Tax is economically the strongest model, but politically it is achievable only in local pilots. We recommend it for city-states and smaller jurisdictions, not as a national solution.
We regard the robot tax as the wrong approach, though the question behind it is the right one. As value creation increasingly comes from automation, public financing must follow. The proposal fails on definition: where automation begins and what counts as a robot is almost impossible to determine in law. Implementation attempts so far show correspondingly little fiscal effect. Levies on capital income and on data achieve the same goal with more reliable tax bases.
From a Swiss perspective direct democracy is both an advantage and a risk. The advantage: a broadly supported reform, such as an "AHV 2.0" with a capital-stock component, can be adopted here more quickly than in EU consensus democracies. The risk: an initiative against higher capital taxes, argued on the grounds of Switzerland's appeal as a business location, can delay the transition by years. That argument is not without foundation. Switzerland lives from its appeal to mobile capital. It stands against a second, equally real consideration: a welfare state without funding is no competitive advantage for a business location either. What matters is whether Switzerland approaches this reform with foresight or only acts under the pressure of crisis.
What this means for investors: First, capital returns and corporate profits are likely to be taxed more heavily than today. Return expectations should reflect that, not ignore it. Second, levies on CO2 and other external costs could rise. Business models that depend on low emission prices carry a valuation risk; efficient providers gain. Third, retirement provision remains the biggest outstanding challenge. A larger capital base in the AHV would increase demand for long-term assets. Fourth, it is worth examining a company's effective tax rate: those paying conspicuously little today have the most to lose when the rules tighten.
Why now: The real problem is not the instrument, but the speed of the transition. If the tax base shrinks faster than the new architecture takes effect, a fiscal vacuum lasting 5–10 years emerges. That is the phase in which democracies tip. History shows this in the Weimar Republic, in post-Soviet transition economies, and in Latin American debt crises. Three years of proactive design are a hundred times more effective than three years of crisis repair.
The strongest objection concerns the time arrow. Capital is mobile; labour is not. Higher capital taxation works only when the major jurisdictions move together, and the OECD minimum tax took around fifteen years to agree on a single rate. A country that begins the transition nationally shifts substance rather than capturing it; one that waits for international alignment may arrive too late. It is a real scenario that both paths prove too slow and the welfare state is instead financed through debt: not as a decision, but as a way of avoiding one.
What we do not yet know
- How can tax competition between jurisdictions be prevented without stifling innovation?
- Which policy mix can be made to stick politically in a direct-democratic system such as Switzerland?
- How does tax law treat synthetic data and foundation models trained on publicly available text?
- Who is best placed to act as custodian of a sovereign AI fund: the state, a cooperative, or a foundation?
- How rapidly can the transition proceed without the fiscal vacuum threatening political stability?
Connected to: 02 · Income, 03 · Value Creation, 07 · Data & AI, 09 · Ecology
Scenarios: Economics & Capital · Geopolitics & Governance · Investment Implications
