Germany’s public AI landscape—state by state.
Place three public evidence lanes side by side: published AI system profiles, regionally assigned procurement and physical infrastructure. You get inspectable evidence—not an invented ranking.
- all 16 German states
- exact state identifiers
- sources for every lane
- no generative AI
Side by side ≠ connected.
A state with more published profiles, notices or route kilometres is not automatically “more AI-ready”. The three sources measure different things, periods and administrative events. That is why there is no composite score.
Procurement reflects published regional assignments—not necessarily the eventual deployment location.One map, three separate readings.
Map colour represents published AI system profiles with an exact state identifier only. Select a state to inspect all three lanes separately.
Public AI systems
82system profilesPublished profiles from the AI System Passport, counted only with an explicit controlled state identifier.
Public procurement
1,358noticesAI-related notices with a published state assignment. This does not prove operation or impact.
Physical infrastructure
16infrastructure measuresAbsolute source values for digital infrastructure. They measure neither quality nor available AI compute.
Permanent state evidence instead of a fleeting map.
Every state has a linkable dossier. You can use this short reference for the complete landscape.
i6eal (2026): German state AI evidence monitor—three separate public evidence lanes for 16 German states, projection generated 20 July 2026. https://i6eal.de/en/tools/ki-laender-lagebild/
Clearly explained
Does a high procurement count mean more AI is in use?
Why might a state have zero system profiles?
Can infrastructure values be read as a quality ranking?
Are the three lanes combined into a score?
Does the collector use generative AI?
Need a defensible regional evidence product?
We build inspectable data products with stable identifiers, visible boundaries and sources that remain verifiable.