How wide is the gap, really?

The headline number is encouraging. In May 2026, 54.5 percent of German companies were using AI, up from 40.9 percent a year earlier, with a further 16 percent planning to start. Adoption is no longer a question of whether. It is happening across the economy.

The detail is less comfortable. Broken down by size, 47.2 percent of Mittelstand firms use AI against 67.2 percent of large companies. That is a twenty-point gap, and it is not closing on its own. Large firms have the teams, the budgets, and the dedicated functions to adopt quickly and govern as they go. Many Mittelstand companies are adopting with neither.

Why is governance the harder half of the gap?

Adoption can be bought. Governance has to be built. German companies are far more aware of their data-protection duties than their AI obligations: surveys put GDPR awareness around 82 out of 100, but AI Act readiness at roughly 56. That gap matters, because the EU AI Act does not wait for readiness.

From August 2026 the transparency obligations bind. Companies that use AI to make or shape decisions about people have to be open about it, and AI-generated content has to be marked. For a large firm with a compliance function, that is a project. For a Mittelstand company that adopted AI without one, it is a scramble.

What does closing the gap actually take?

The firms that close it treat governance as part of adoption, not a later add-on. Before a tool goes live, they know what it touches, what decisions it influences, and who is accountable for it. This does not require a large team. It requires the discipline to ask the questions early, while the answers are still cheap to act on.

The alternative is the expensive path: adopt fast, govern never, and then meet the August deadline with unmapped systems and no documentation. That is how a competitive lag becomes a compliance liability. The Mittelstand does not have to be behind. It has to decide that governance is part of the work, starting now.

Read next: Should You Still Bet Your Company on a Single AI Model?  ·  Is AI Still a Bet, or Is the Revenue Already Here?