
Here is this week’s EU AI & Sovereignty news roundup, covering the landmark Tech Sovereignty Package, AI Act amendments, compliance realities, and more.
-
EU Launches Tech Sovereignty Package — Chips Act 2.0 and Cloud & AI Development Act — On June 3, 2026, the European Commission unveiled its most ambitious tech policy package since the AI Act. It includes the Chips Act 2.0 (semiconductors), the Cloud and AI Development Act (tripling datacenter capacity in 5-7 years), an EU Open Source Strategy, and a Strategic Roadmap for digitalization and AI in energy. This represents the single largest shift in EU tech sovereignty policy to date — moving from regulation-only to active industrial investment and infrastructure building.
Source: European Commission — Digital Strategy -
Digital Omnibus on AI — Political Agreement Reached (May 7, 2026) — The Council and European Parliament reached a provisional political agreement on the Digital Omnibus for AI — the first amendments to the AI Act since its adoption in 2024. The package includes targeted simplification (reducing regulatory overlap), deadline extensions for high-risk AI, and new prohibitions. This is part of Omnibus VII, the broader EU simplification package. The agreement signals that even as the EU doubles down on AI regulation, it recognizes the need to adapt rules to industry realities.
Source: Council of the EU -
High-Risk AI Deadline Pushed: August 2, 2026 → December 2, 2027 (Provisional) — The Digital Omnibus provisionally moves the high-risk AI compliance deadline (Annex III) to December 2, 2027. The original August 2, 2026 date remains legally in effect until the Omnibus is published in the Official Journal. National supervisory authorities across all 27 member states will then gain enforcement powers over high-risk AI systems covering recruitment, credit scoring, biometrics, critical infrastructure, law enforcement, and education. Companies should treat the original date as their target — the extension is not yet law.
Source: Aurora Trust AI -
European Parliament Votes in Favor of AI Act Amendments (June 16, 2026) — The European Parliament voted 423-57 to formally amend the AI Act. The changes — part of the broader Digital Omnibus — include deadline extensions, a new prohibition on certain AI practices (effective December 2, 2026), and clarified transparency obligations for AI-generated content. The overwhelming majority (423-57) shows broad cross-party consensus that the AI Act needs pragmatic adjustments without abandoning its core principles.
Source: National Law Review -
78% of Enterprises Unprepared for AI Act Compliance — Vision Compliance Report — A large-scale study by Vision Compliance across financial services, healthcare, tech, manufacturing, energy, retail, telecom, and transport found that 78% of organizations have taken no meaningful steps toward AI Act compliance. With enforcement looming (August 2026 / December 2027), this is a wake-up call for the entire EU market. The gap between regulatory ambition and organizational readiness has never been wider — expect a surge in compliance consulting and AI governance tools in the coming months.
Source: Global Fintech Series -
EU Cybersecurity Package (January 20, 2026) — NIS2 Alignment — The Commission published a new cybersecurity package aligning the Cybersecurity Act with NIS2. Goals: simplifying jurisdiction rules, streamlining ransomware attack data collection, and strengthening cross-border supervision coordination via ENISA. Part of the broader digital sovereignty policy framework — cybersecurity and AI sovereignty are increasingly intertwined as AI systems become critical infrastructure.
Source: European Commission — Cybersecurity Q&A -
Netherlands: AI Act Implementation Law in Consultation; Supervision Structure — The Dutch Cabinet took a step forward with the Implementation Act for the AI Regulation, which was in public consultation until June 1, 2026. This law governs national supervision of the AI Act, to be carried out by multiple authorities (AP, RDI, etc.). The ‘Implementation Framework’ for responsible use of algorithms supplements the AI Act. The Netherlands is positioning itself as a frontrunner in AI governance implementation among EU member states.
Source: Rijksoverheid -
Dutch Proposal: ‘Sovereign Tech Stack’ as Guiding Concept — The Netherlands proposed organizing the EU Tech Sovereignty Package around one clear concept: a ‘Sovereign Tech Stack’. This aims to counter fragmentation of European tech initiatives and realize a coherent, open-source-based digital infrastructure for the EU. It is a pragmatic framing that could give the sprawling Tech Sovereignty Package a unifying narrative — making it easier for citizens and businesses to understand what the EU is building.
Source: iBestuur -
AI Literacy (Article 4 AI Act) — Challenge for European Organizations — The Digital Omnibus proposal amends Article 4 (AI literacy) of the AI Act: the binding obligation for providers is being revised. Academic research (Hanze University) shows AI literacy adoption remains a major challenge for organizations. The revised provision aims to make it more workable for SMEs. AI literacy is the foundation of the entire AI Act — if organizations do not understand AI, they cannot comply with the rest of the regulation.
Source: Hanze University Research -
EU Open Source Strategy as Sovereignty Instrument — The third pillar of the Tech Sovereignty Package is the EU Open Source Strategy. It aims to counter vendor lock-in, strengthen European digital autonomy, and promote interoperable digital infrastructure in both public and private sectors. This builds on earlier initiatives such as Next Generation Internet (NGI) and ISA² programs. For AI sovereignty specifically, open-source AI models and frameworks are a strategic priority — reducing dependence on Big Tech proprietary ecosystems.
Source: Global Policy Watch
Key Trends This Month
- From regulation to investment: The Tech Sovereignty Package marks a fundamental shift — the EU is no longer just regulating AI but actively building infrastructure (datacenters, semiconductors, open-source) to compete globally.
- AI Act in flux: The Digital Omnibus shows the AI Act is not static — first amendments are already being adopted before the original deadlines have passed, reflecting the speed of AI development.
- Compliance gap widening: With 78% of enterprises unprepared and enforcement approaching, expect a compliance services boom across the EU.
- Member states stepping up: The Netherlands’ Sovereign Tech Stack proposal and implementation law show national governments are taking initiative beyond Brussels.
- Open source as strategy: Open-source AI infrastructure is now explicitly framed as a sovereignty instrument — a major policy shift reshaping Europe’s AI ecosystem.
Wat vond je van dit bericht?