EU AI Policy: A Field Guide

Maps, money, infrastructure, and the open questions Europe still needs to answer.

Europe spent the last decade writing the rulebook for AI while the systems being regulated kept moving. The result is a layered landscape: a strategy here, a funding instrument there, a directive that overlaps with two regulations and an ethics framework. This page is a guided tour through the pieces I’ve written about, organized by question rather than by document name.

Note: this is a working pillar page — the prose below is a scaffold and will be refined over time. Each section links out to a longer essay that goes into detail.

Strategy: what is Europe actually trying to do?

The European AI agenda is easy to caricature (“regulation-first, innovation-last”) and hard to summarize honestly. The Coordinated Plan on AI sets the political ambition; individual member states then translate it into national strategies that don’t always line up.

• Europe’s AI Research Agenda: the Coordinated Plan and Horizon projects, explained — what the EU’s overarching AI strategy actually says, and how it is operationalized through Horizon Europe.
• Lessons from China for the EU on global AI governance — what comparing governance models tells us about Europe’s distinctive bets.

Funding: where the money goes (and where it leaks)

European AI research is well-funded by global standards. The harder question is whether the funding is coordinated enough to produce results greater than the sum of its parts.

• Horizon AI fragmentation: budget overlaps and missed efficiency — analysis of Horizon Europe project data showing where overlapping mandates eat into the effective budget.

Data: the foundation everything else stands on

The EU’s AI ambitions rest on a data ecosystem that is partly built and partly imagined. The Data Governance Act, Data Act, common data spaces, and copyright framework all try to answer the same question from different angles: how does data move in Europe?

• The EU’s Data Strategy: unlocking Europe’s AI future — how the Data Governance Act and Data Act shape the data ecosystem AI models depend on.
• The big open questions on Europe’s AI data policy — what is still ambiguous about training on PII and copyrighted material in the EU.

Infrastructure: AI factories and the compute question

If data is the foundation, compute is the building. The EU’s AI Factories and Gigafactories programs are an attempt to keep frontier-scale infrastructure on European soil — and the site selection problem alone is harder than it looks.

• AI Factories: how Europe plans to centralize AI resources — what AI Factories are, why they exist, and what they are meant to enable.
• AI Gigafactories: Europe’s €20B race and the site selection challenge — the €20B program, the site selection problem, and the open-source tool I built to help reason about it.

Tooling: making the landscape navigable

The policy landscape is hard to read because it lives in PDFs scattered across institutions. Visualization helps.

• PolAI Graph: visualizing the EU AI ecosystem — an interactive graph mapping policies, projects, funds, and organizations to each other.

Where this is going

I am writing this guide because the EU AI conversation tends to ricochet between two unhelpful poles: “Europe is over-regulating itself out of the race” and “Europe is the only adult in the room.” The reality is more interesting. The pieces linked above are an attempt to look at the moving parts honestly — what works, what overlaps, what is still unclear — and to keep updating the picture as the landscape evolves.

If you have a question this page should answer and doesn’t, get in touch.