Value Creation & Ownership
Widely distributed value creation and ownership across the Mittelstand, yet succession planning and digitalisation are stalling.
Where does Germany stand, and where does it break?
The Mittelstand accounts for around 99 per cent of German companies and more than half of all employees subject to social insurance contributions; roughly 1,500 hidden champions lead their global markets. Ownership is frequently family-held and long-term in orientation, complemented by co-determination on supervisory boards and at establishment level. Germany's value-creation model distributes ownership and decision-making power comparatively broadly — through family entrepreneurs, workforces and regional rootedness rather than through anonymous capital markets. At the same time, equity culture remains underdeveloped: only a minority of households hold shares or fund units, and private wealth is highly unequally distributed and tied predominantly to real estate rather than productive capital.
AI shifts value creation from labour to capital — specifically to those who own the models, data centres and data. These assets are currently held predominantly by a small number of mostly foreign platform corporations, whilst Germany's Mittelstand procures AI rather than owning it. A new concentration of ownership looms: the most productive capital stock of the future may belong neither to the broader population nor to the Mittelstand, with the result that automation returns flow abroad and to a narrow group of owners. Without broad participation, the economy ahead will deepen an already high degree of wealth inequality and hollow out the foundations of Germany's co-determined economic model.
Paths for Germany
Employees become co-owners of their companies' capital through tax-advantaged participation schemes, allowing them to share in the returns from automation. For Germany, this builds on the tradition of co-determination and fits naturally with its Mittelstand structure. The improved tax-exempt allowances for employee capital participation introduced in 2024 represent early steps, but Germany remains behind international peers. The impact is limited as long as only profitable firms willing to offer participation schemes take part and the lion's share of AI returns accrues to external platforms.
Trade-off: Broader participation versus limited reach
Feasibility: Culturally well-suited, but needs to be made considerably bolder
A state-managed, broadly diversified fund accumulates capital assets on behalf of all citizens. It distributes returns or feeds into retirement provision. This addresses the fault line directly, as it gives the population a stake in precisely the capital stock that benefits from AI — irrespective of personal savings. For Germany, this could build on the existing Generationenkapital, which could be developed into a genuine citizens' fund. Challenges include financing, the political independence of management and market risk.
Trade-off: Broad participation versus state control of capital
Feasibility: The Generationenkapital is the natural starting point
Germany and Europe are building their own data centres, models and data infrastructure. This means the AI capital stock is held at least partly domestically and value creation remains within the country. For the Mittelstand, the decisive question is whether it owns AI or merely leases it. Only domestic infrastructure keeps returns on capital within the country. Feasibility is constrained by the technological lead and capital strength of US and Asian providers. Without massive, coordinated investment, procuring foreign systems will remain the norm.
Trade-off: Technological sovereignty versus enormous investment requirements
Feasibility: Realistic only at a European level and with a long-term commitment
Subsidies, cooperative models, and shared computing infrastructure enable family-owned businesses and hidden champions to develop and own AI applications themselves rather than simply procuring them. This is central to Germany because the Mittelstand forms the backbone of employment and regional value creation. Sector-specific expertise and proprietary data represent a genuine competitive advantage — provided they are translated into owned AI. Without pooling, individual firms lack the scale to build their own models.
Trade-off: Decentralised strength versus scale disadvantages
Feasibility: Cooperative pooling could close the scale gap
Trigger-based phases
Trigger: The Mittelstand and large corporations integrate AI primarily as purchased software from external providers.
Productivity rises. Yet the value created at the AI layer is extracted as licence and cloud fees. The ownership question is not yet perceived as a structural problem. Early debates on data sovereignty and dependency are taking shape.
Trigger: A growing share of profits visibly shifts away from labour and manufacturing towards the owners of AI platforms.
Wealth inequality and dependence on a small number of providers are becoming political issues. Calls for participation models, sovereign infrastructure, and a citizens' fund are gaining traction.
Trigger: Policymakers and the business community determine through which channels the broader population gains a stake in the AI capital stock.
Germany faces a choice between broad participation — via ownership stakes, a citizens' fund, and sovereign infrastructure — and growing concentration. This decision will determine the distribution of wealth and the future of the co-determination model through to 2040.
Real-world pilot programmes
Status: In force. Higher tax allowances for employee share ownership have applied since 2024
The raised allowance and deferred taxation ease participation models, particularly for start-ups and the Mittelstand. It is a genuine German lever for giving employees a stake in productive capital, though its reach has so far been modest.
Status: Not enacted; debate open (state-owned funded investment vehicle)
The fund gives the state, and indirectly citizens, a stake in global productive capital. It is the natural starting point for organising broad participation in the AI capital stock — provided it is conceived as something more than a tool for containing pension costs.
Status: Under construction (shared industrial data spaces)
Initiatives such as Catena-X and Manufacturing-X pool industrial data so that Mittelstand companies and large corporates can build their own sovereign AI applications. They represent the most concrete approach to translating Germany's competitive advantage in expertise and data into owned AI assets.
Status: Established in Norway; to be assessed as a model for Germany
The Norwegian Government Pension Fund demonstrates how a country can accumulate and manage capital returns on behalf of all its citizens. For Germany, it serves as a reference model for how a citizens' fund could organise broad participation in global productive capital.
What we are watching
The labour share measures what proportion of value creation accrues as labour income and what proportion as return on capital. This measure matters because a declining share is the core symptom of the shift from labour to capital. A reversal would mean that, without broad capital participation, inequality grows structurally. For policymakers and investors alike, it is the benchmark for whether the returns from automation are being shared.
This measures how broadly the population participates in productive capital at all. The threshold matters because AI returns flow through capital, and those who hold no capital are excluded. If the ratio remains low, the gains from automation concentrate among a few. An upward shift would signal broader participation.
This indicator shows whether the AI capital stock resides domestically or whether returns flow abroad. The threshold matters because only domestic infrastructure keeps value creation and ownership within Germany. If the share remains low, the Mittelstand remains a tenant rather than an owner of key technology. An upward shift would strengthen technological and economic sovereignty.
The coefficient measures how unevenly private wealth is distributed. The threshold matters because the economy ahead rewards capital ownership and may amplify existing inequality. An upward shift towards even greater concentration signals that participation models are failing. For political stability, this is one of the most sensitive early indicators.