When AI Makes Capital More Valuable Than Labour
The most consequential effect of artificial intelligence may not appear in a technology portfolio. It may emerge in the division of economic income between those who work and those who own the systems through which work is performed.
For most of the industrial era, capital and labour have remained mutually dependent. Machinery increased productivity, but factories, logistics networks and offices still required people to operate them. Additional capital eventually produced diminishing returns because human labour remained a limiting factor.
Advanced AI and robotics could weaken that constraint. Software capable of performing cognitive tasks can be replicated at negligible marginal cost, while increasingly autonomous machines extend automation into the physical economy. Should these technologies become genuine substitutes for labour rather than tools used by workers, a larger share of economic output may flow to the owners of models, computing infrastructure, proprietary data and automated companies.
For families managing long-duration capital, the argument reaches beyond deciding how much to allocate to technology. It raises questions about where future returns will accumulate, which forms of ownership provide access to them and how a more capital-intensive economy could alter taxation, regulation and social legitimacy.
The Piketty Scenario Returns In A Different Form
Thomas Piketty’s best-known thesis rests on the idea that wealth tends to concentrate when the return on capital exceeds economic growth. Critics have challenged both its historical consistency and the assumption that such a relationship must persist.
AI introduces another possible route to the same destination.
If machines can replace a substantial part of human labour, capital no longer depends on workers to the same extent. Its marginal value need not decline as quickly when more of it is accumulated. Owners can reinvest a greater proportion of their returns, acquire additional productive assets and transmit those assets to the next generation. Wage income, meanwhile, may lose relative importance.
Such an outcome remains a scenario rather than a forecast. Technological transitions have repeatedly created occupations that were difficult to imagine in advance. AI could increase productivity, reduce costs and stimulate enough new demand to support employment elsewhere.
The distribution of the gains will nevertheless matter as much as their total size. A technology can raise aggregate output while leaving a narrow ownership group with most of the financial benefit.
Switzerland Begins From A Position Of Substantial Capital
Swiss households held CHF 3.28 trillion in financial assets at the end of 2025. Much of this wealth does not sit in directly managed securities portfolios. Occupational pension entitlements, insurance assets and deposits represent a considerable proportion, while listed shares, investment funds and other securities provide more explicit exposure to market returns.
Real estate adds another large store of household wealth. The figures describe a country with significant accumulated capital, but not one in which every household participates in the same way.
Switzerland’s pension system offers broad indirect exposure to financial markets. Employees build occupational pension claims throughout their careers, and pension funds invest across bonds, equities, property and private markets. That structure could distribute part of the productivity gains from AI more widely than in an economy where retirement provision depends almost entirely on public transfers.
The protection is incomplete. Pension beneficiaries do not hold direct control over the underlying assets, and investment limits, conversion rates and demographic pressures affect how market returns reach them. Those who own companies, concentrated equity positions or private-market interests can participate more directly in rising capital values and retain greater discretion over reinvestment.
AI could therefore expand the difference between institutional participation in capital and controlling ownership of it.
Ownership May Matter More Than Sector Classification
A conventional investment response would concentrate on semiconductor companies, cloud providers, data-centre operators and software platforms. These businesses supply the infrastructure behind AI and have already attracted enormous amounts of capital.
A broader reading looks at companies outside the technology sector that can use AI to reduce labour intensity without surrendering the resulting economics to their suppliers.
An insurer that automates claims processing, a pharmaceutical company that shortens parts of drug discovery or an industrial group that combines robotics with proprietary production data may capture more durable value than a business merely subscribing to a widely available AI service.
The distinction lies in ownership of the advantage. Proprietary data, distribution, regulatory approvals, specialised intellectual property and difficult-to-replicate operational systems can allow an established company to retain productivity gains. Businesses without these protections may face the opposite result: lower barriers to entry, increased competition and little power to prevent technology providers from absorbing the margin.
Private-market access becomes relevant here. Some of the most valuable AI infrastructure and application businesses may remain private for longer, supported by venture capital, private equity or strategic investors. Public-market portfolios can still participate through listed suppliers and adopters, but they may not capture the full return generated during the earlier phases of company development.
Families evaluating private investments should be wary of treating an AI label as evidence of scarcity. A model assembled from third-party technology with no exclusive data or distribution may have limited defensibility. The pertinent questions concern who owns the intellectual property, how dependent the company is on external computing providers and whether its economics survive when underlying models become cheaper and more widely available.
The Swiss Corporate Structure Creates A Different Exposure
Switzerland combines global listed groups with a large population of privately controlled companies. Many of these businesses compete through specialised engineering, precision manufacturing, pharmaceuticals, financial services and export-oriented expertise.
This structure gives the country several advantages. High labour costs already encourage investment in productivity. A strong research base, sophisticated industrial clusters and access to international talent support adoption. Labour shortages in healthcare, engineering and skilled technical occupations also mean that automation can relieve genuine capacity constraints rather than merely replace available workers.
The risks are equally specific. A Swiss company may deploy AI developed and hosted abroad, purchase advanced robotics from an international supplier and depend on foreign cloud infrastructure. It can improve its margins while a significant share of the underlying return accrues to technology owners in the United States or elsewhere.
For business-owning families, AI strategy therefore belongs within ownership strategy. Management should be able to explain which productivity gains the company expects to retain, which capabilities it is effectively renting and where dependence on a small number of external platforms is developing.
A lower headcount alone is a poor measure of success. More useful indicators include faster product development, greater capacity without equivalent capital expenditure, improved quality and the creation of proprietary knowledge that remains within the company.
The Social Contract May Become A Portfolio Risk
A sustained shift from labour income towards capital income would eventually produce a political response.
Switzerland finances a substantial part of its social system through employment income, payroll contributions and taxation. A labour market with fewer conventional jobs or weaker wage growth could place pressure on that model even if national output continued to rise.
Governments would then face a choice between allowing inequality to widen, reducing benefits or moving more of the tax burden towards corporate profits, capital gains, wealth, consumption or automated production. The precise policy response would differ by jurisdiction, but owners of capital should not assume that a larger share of national income can move towards assets without changing the rules governing those assets.
Switzerland’s federal structure complicates any simple prediction. Wealth and inheritance taxation remain strongly influenced by cantonal policy, while political change generally proceeds through negotiation and popular votes. This can moderate abrupt intervention, but it also places questions of fairness directly before the electorate.
Private wealth may remain politically defensible when it is associated with investment, employment, innovation and patient ownership. It becomes harder to defend when returns appear detached from broad economic participation.
Stewardship is therefore not separate from investment performance. How companies treat employees during automation, whether productivity gains support training and how owners communicate the purpose of capital may affect the regulatory environment in which future returns are earned.
Succession Could Reinforce The Divide
AI may also change the significance of inheritance.
When a family transfers ownership of a productive company or a diversified portfolio, the next generation receives assets capable of compounding. A household dependent mainly on employment income transfers far less economic leverage, even when its members are highly educated and professionally successful.
Should labour become a weaker source of wealth creation, inherited ownership will carry more weight. The difference between receiving capital and beginning adult life without it could widen across successive generations.
For families already planning succession, this creates two responsibilities. The first is technical: ensure that ownership structures, governance and investment mandates remain suitable for an economy in which intangible assets and technological dependencies become increasingly important.
The second concerns capability. Beneficiaries need to understand how capital produces returns, where those returns originate and which social obligations accompany ownership. A succession plan that transfers assets without preparing future owners may preserve financial wealth while weakening the legitimacy and judgement required to manage it.
Philanthropy may also move closer to the centre of the family’s response. Education, access to technology, workforce transition and entrepreneurial finance offer ways to widen participation in the gains from AI. Such commitments are most credible when integrated with the family’s investment and business conduct rather than treated as compensation for it.
Portfolio Construction Needs More Than An AI Allocation
The possibility of a more capital-intensive economy does not justify indiscriminate exposure to technology. Valuation still matters, technological leadership can change quickly and today’s infrastructure provider may become tomorrow’s commoditised utility.
A resilient approach extends across several layers: the infrastructure required to develop and operate AI; businesses with defensible applications; established companies able to retain productivity gains; energy and physical assets needed to support computing; and investments exposed to the political response that automation may provoke.
Geographical diversification deserves renewed attention. The location of innovation, corporate ownership, taxation and end demand may increasingly diverge. A Swiss investor can benefit from automation at home while depending on returns generated by foreign companies. Currency exposure and political concentration may rise even when the portfolio appears diversified by industry.
Families should also examine the role of human-capital-dependent assets. Some businesses will become more valuable because AI reduces their reliance on scarce expertise. Others may lose their distinction as knowledge becomes easier to reproduce. The analysis belongs at company level rather than within broad assumptions about which professions or sectors will survive.
The Question Is Who Owns The Productive System
AI optimists may prove correct. Productivity could rise sufficiently to create new industries, better-paid work and broader prosperity. The pessimistic scenario, in which ownership concentrates while wages lose economic weight, is neither inevitable nor too remote to ignore.
For private wealth, the appropriate response is not to predict the end of employment. It is to understand the changing relationship between capital, labour and ownership.
A portfolio may perform well because AI raises corporate margins. A family company may become more productive with fewer employees. An inherited stake may compound faster than earned income. Each development can be financially attractive while contributing to a political and social structure that ultimately changes the conditions under which capital operates.
The long-term owner must consider both sides of that equation. AI may make capital more productive. Whether it also makes capital more exposed will depend on how widely its benefits are distributed and how responsibly ownership is exercised.
