In a nutshell: German AI providers promising digital sovereignty frequently rely on US models as the technological foundation of their offerings.
An analysis of 173 German AI providers reveals a discrepancy between marketing claims and technical reality: although they promote digitally sovereign solutions, approximately 40 percent rely on US artificial intelligence models as the basis of their offerings.
In a survey of 173 German providers explicitly advertising digital sovereignty, EU hosting, or German data storage, the results reveal a central contradiction: approximately 40 percent of these providers build their solutions on US-based foundation models – from companies such as OpenAI, Anthropic, or other US-based AI developers.
From the perspective of a Chief Digital Officer, this represents a regulatory and strategic risk factor. The mere geographic relocation of data centers and data processing to German or European territory does not mask the dependence on foreign AI models whose training, governance, and potential security vulnerabilities lie outside one’s own control. A US model can bring data protection, compliance, and geopolitical risks regardless of where the hosting is located.
For CDOs, this means that marketing promises regarding digital sovereignty must be verified against their actual technical implementation. The crucial question is whether a provider develops and trains the model itself, or whether it merely offers foreign models behind a regional wrapper. The former encompasses genuine controllability and IP security; the latter does not.
Source: itwelt.at · Published 22 July 2026
Lumi AI News — AI-assisted curation in accordance with Art. 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.3.