The Megawatt Foundation — why the Mittelstand is not AI's casualty but its base

There is a comfortable story about artificial intelligence and the German Mittelstand. It goes like this: the AI economy is the world of large cloud platforms, technology companies, data centres in Nevada or Singapore. The Mittelstand is the old world — precision engineering, special-purpose machinery, family-owned firms with long cycles and even longer decision chains. It will have to adapt, or it will not survive. In this story, the Mittelstand is the one being upended, not the one doing the upending.

I consider this story wrong. Not optimistically wrong. Structurally wrong.

**The question that comes before the AI pitch**

Anyone who looks at the current development with a sober eye asks a question that most AI strategy papers omit: what does this economy actually run on, physically? Large language models and AI systems at industrial scale consume substantial amounts of energy. Data centres must stand somewhere, must be cooled somewhere, must be connected to the grid somewhere. The ideas economy is not immaterial. It depends on a physical base that it does not itself produce.

That base — energy, land, cooling capacity, grid connection, industrial infrastructure — is not in the hands of the technology companies in Germany. It sits where it has sat for generations: in the industrial Mittelstand and in the regions it has shaped.

This is not consolation for an industry under pressure. It is a strategic inventory.

**Sovereign infrastructure requires industrial ground**

The debate about AI sovereignty in Europe has so far been conducted mainly in terms of software and models. Whose models? Whose data? Whose legal framework? These are the right questions. But underneath them lies a more fundamental one: whose infrastructure?

A European AI system running on American or Asian cloud platforms is sovereign until the moment the export-control authority or the operator's compliance department decides who still has access. That this is not a theoretical risk has been demonstrated in recent experience — providers notifying their European customers with a few hours' notice that certain models were no longer accessible.

Sovereign compute capacity on European soil requires sites with stable energy supply, industrial reliability, and cooling capacity. It requires organisations with experience in load management, in combined heat-and-power systems, in energy-intensive processes. It requires what the Mittelstand, in many cases, already has.

I am not asserting that the Mittelstand will operate data centres tomorrow. I am asserting that the Mittelstand already carries the competence and infrastructure on which a sovereign AI economy must build — if this is recognised as a strategic question before it is decided by default.

**Competitive advantage as a dormant position**

What is genuinely remarkable about this situation is not the risk facing the Mittelstand. It is the position it already holds.

Industrial sites with direct grid connections, waste-heat recovery, stable energy contracts, and operational excellence in energy-intensive processes are not easily replicated. They have been built over decades. Those who have them, have them. Those who do not, will not build them in three years.

This infrastructure does not carry the label "AI strategy." It appears in annual reports as a production site, an operating cost block, a legacy asset of the industrial past. That the same infrastructure might be the physical precondition of the digital future — this is the position I would describe as dormant. Not lost. Dormant.

A dormant position is an advantage only when it is recognised and claimed before others value it differently.

**The question management should be asking now**

When a chief executive or board member of a mid-sized industrial firm surveys the AI debate, they typically ask the wrong question. The wrong question is: how do we integrate AI into our existing processes? That is a legitimate operational question. It is not the strategic one.

The strategic question is: which of the resources we control — infrastructure, energy, land, process expertise, grid connection — are structurally relevant to the AI economy? And do we know what they are worth?

Those who do not ask this question answer it anyway, by omission. They administer a competitive advantage without knowing they hold it. And they risk selling it to someone who recognised it first.

The thesis is simple: megawatts powered our past. Ideas power our future — on the foundation of the megawatts the Mittelstand already provides. The question is not whether this foundation exists. It does. The question is whether management recognises it for what it is: a first-mover position in an economy that is searching for its physical base.

Those who wait until others have recognised it will negotiate from the second chair.