Lead Editor · AI Policy
Dr. Elena Vasquez
Former advisor to the EU AI Office (2024–2025). Previously a Senior Policy Analyst at the European Centre for AI Governance in Brussels. Holds a PhD in Technology Law from KU Leuven and an MSc in Computer Science from TU Delft.
Areas of Expertise
Sovereign AI Strategy
National compute infrastructure, data localisation, public sector AI
EU AI Act Compliance
High-risk classification, GPAI obligations, conformity assessment
Geopolitics of AI
US–EU–China technology rivalry, export controls, standards wars
AI Governance
Corporate AI governance, board-level risk, regulatory reporting
Weekly Intelligence Analysis
Published every MondayThe Sovereign Stack Fractures: Three Signals from Brussels, Beijing and Washington
This week's intelligence picture is dominated by a convergent theme: every major AI power is quietly reassembling the supply chain it doesn't yet own. The EU AI Office's draft guidance on GPAI systemic risk thresholds landed Tuesday — and buried in Annex C is a clause that would classify any model with over 10^26 FLOPs training compute as automatically 'systemic'. That catches GPT-5, Gemini Ultra and Llama 4 simultaneously, but not Mistral's latest open-weights release. Intentional asymmetry, or drafting artefact? My read: intentional.
EuroHPC's Quiet Expansion and the Data Centre Sovereignty Paradox
EuroHPC secured €2.1 billion in supplementary commitments this week, but the real story is what those commitments don't cover: inference infrastructure. Europe has bet heavily on training-class supercomputers — LUMI, Leonardo, Jules Verne — while the actual workload shifting to AI in 2026 is overwhelmingly inference. A training cluster that sits idle between training runs is a geopolitical statement, not a business asset.
AI Liability Directive: Why the Commission's Retreat Matters More Than the Advance
The European Commission quietly withdrew the proposed AI Liability Directive from its legislative calendar this week — a decision that received far less coverage than it deserved. The official position is 'consolidation with the Product Liability Directive revision'. The practical effect is that AI companies operating in Europe face a two-year window with no meaningful civil liability framework beyond existing tort law.
Biography
Dr. Elena Vasquez is the Lead Editor and AI Policy Analyst at Global AI Watch. Her work sits at the intersection of technology governance, geopolitics, and legal frameworks — with a focus on how states are building, regulating, and competing over AI capabilities.
Between 2024 and 2025, she advised the EU AI Office on GPAI systemic risk assessment methodology, contributing to the technical annexes of the first GPAI Code of Practice. Before joining the AI Office, she spent four years at the European Centre for AI Governance analysing member state AI strategies and benchmarking European compute infrastructure against US and Chinese investments.
Her PhD research at KU Leuven examined the legal classification of autonomous decision systems under EU administrative law, specifically the tension between algorithmic decisioning and the right to explanation under GDPR Article 22. Her dissertation was cited in the European Parliament's AI Act rapporteur report.
Elena writes the weekly Sovereign Intelligence Digest, published every Monday, which analyses the most strategically significant AI policy developments of the preceding week for executives and policymakers.