🤖 EU AI & Sovereignty
Wednesday, July 29, 2026
The EU Digital Omnibus enters into force, Microsoft and Mistral expand their partnership with billions in sovereign AI investment, Portugal launches its first open-source AI model, and the 2 August transparency deadline is now just days away. Here is today’s EU AI & Sovereignty update.
wetgeving
🔥 HOOG
📜 EU Digital Omnibus on AI Enters Into Force — Major AI Act Overhaul
The EU Digital Omnibus — the largest revision of the AI Act since its original adoption — entered into force on 27 July 2026. Key changes include extended deadlines for high-risk AI systems (now mid-2027), eased compliance for SMEs, and a ban on AI “nudifier” apps by December 2026. The transparency obligations for AI-generated content, however, go into effect on 2 August 2026 — just 4 days from now. This is the most significant regulatory shift in EU AI governance this year, balancing stricter rules with much-needed flexibility for European businesses.
wetgeving
🔥 HOOG
⏰ AI Act Transparency Deadline: 4 Days to 2 August 2026
In just four days, the AI Act’s transparency rules take effect. All AI systems that generate or manipulate content must comply with labelling and transparency obligations. The European Commission has finally published the definitive Code of Practice, and Google became the first major tech company to sign it. Companies that fail to comply risk fines of up to 3% of global annual turnover. This is a critical deadline that will affect every company deploying generative AI in the EU.
investeringen
🔥 HOOG
🤝 Microsoft & Mistral Expand Strategic Partnership — Billions for Sovereign AI
Microsoft and Mistral have significantly expanded their partnership with a multi-billion euro investment. Microsoft will integrate Mistral’s models into Azure AI and Copilot Studio, specifically targeting European regulated industries. The partnership is explicitly positioned as “sovereign AI” — European companies retain control over their data while accessing frontier models. Mistral is now valued at approximately €20 billion, cementing its position as Europe’s primary AI champion. This deal signals that sovereign AI is no longer just a political ambition but a commercially viable strategy.
investeringen
📌 MEDIUM
🏭 Samsung Weighs €1 Billion Mistral Stake in Dual Chip-and-AI Sovereignty Play
Samsung is negotiating an equity stake of up to €1 billion in Mistral AI, which is targeting a €20 billion valuation. The investment is seen as a “dual chip-and-AI sovereignty play” — Samsung wants to link its AI chip operations (Exynos, foundry) with Mistral’s models to create a European-Asian AI alternative independent of US and Chinese dominance. If completed, this would further strengthen Mistral’s position and create a unique transcontinental AI alliance.
lidstaten
📌 MEDIUM
🇵🇹 Portugal Launches Amália — First Open-Source AI Model as a Public Good
Portugal has launched “Amália”, the country’s first open-source AI model, named after the legendary fado singer Amália Rodrigues. The model is positioned as a “public good” and was developed with government funding. Portugal joins a growing list of European nations — France (Mistral), Germany (Aleph Alpha), Spain — developing their own models to reduce strategic dependence on the US and China. This grassroots approach to sovereign AI, driven by individual member states rather than Brussels, is an interesting counterpoint to top-down EU regulation.
Commissie
🔥 HOOG
🛡️ CEPA: “Avoiding a European AI Dark Age” — Warning of Falling Behind
The Center for European Policy Analysis (CEPA) warns that without drastic measures, Europe faces an “AI Dark Age”. The think tank argues that the combination of over-regulation, lack of venture capital, and an AI talent shortage is structurally leaving Europe behind the US and China. The report calls for a European “AI Manhattan Project” — a coordinated investment of at least €50 billion in AI infrastructure, compute capacity, and talent development. This is a sobering counterpoint to the regulatory optimism around the Digital Omnibus.
technology
📌 MEDIUM
🔵 Google Signs EU AI Act Code of Practice on AI Transparency
Google has become the first major US tech company to sign the EU AI Act Code of Practice for transparency of AI-generated content. The company commits to labelling AI content in search results, YouTube, and other Google services. This is a significant step — it shows that big tech is taking EU regulation seriously, even as some parties continue to lobby against stricter rules. The timing, just before the 2 August deadline, is strategic and sends a strong signal to the rest of the industry.
Commissie
📌 MEDIUM
🌍 EU AI Act Drives Global AI Governance — The Brussels Effect in Action
A new study by the Thomson Reuters Foundation demonstrates the “Brussels Effect” in action: countries outside Europe — including Brazil, Japan, South Korea, and Canada — are increasingly adopting elements of the EU AI Act in their own legislation. The study concludes that the EU, despite its limited AI compute capacity, is becoming the de facto global standard-setter for AI regulation. This strengthens the EU’s position in the sovereign AI debate but also raises questions about whether regulation alone is sufficient without corresponding investment in AI capacity.
wetgeving
📌 MEDIUM
🏛️ Europe’s AI Crossroads: Regulation vs. the Race to Keep Up
EU Reporter analyses the fundamental tension Europe faces: the world’s strictest AI regulation (AI Act + Digital Omnibus) on one hand, and a growing gap with the US and China on AI capacity on the other. The article argues that the Digital Omnibus is a necessary relaxation, but Europe still struggles with fundamental problems: lack of compute, brain drain to the US, and fragmented investment. The key question is whether relaxed rules alone are enough to let European AI champions like Mistral grow to compete globally.
Commissie
📌 MEDIUM
🧠 CEPS Task Force: EU Needs an “Apply AI Strategy”
The influential Brussels think tank CEPS has launched a Task Force for an “Apply AI Strategy” — a strategy focused not on building frontier models, but on AI adoption and application in European industries. The task force emphasises that Europe’s real opportunity lies not in winning the frontier model race (dominated by the US and China), but in smartly applying AI in Europe’s strong sectors: manufacturing, healthcare, energy, agriculture, and financial services. This pragmatic approach may be Europe’s most realistic path to AI leadership.
📈 Key Trends
- • Digital Omnibus in effect — The largest AI Act revision is now law, extending deadlines for high-risk systems while tightening transparency rules. The 2 August deadline is just 4 days away.
- • Sovereign AI goes commercial — Microsoft-Mistral’s multi-billion partnership and Samsung’s potential €1B stake show sovereign AI is no longer just political rhetoric but a real market strategy.
- • Member state-led AI development — Portugal’s Amália joins a growing list of national AI models, suggesting a bottom-up approach to European AI sovereignty alongside top-down EU regulation.
- • Regulation vs. capacity gap widens — CEPA’s “AI Dark Age” warning and CEPS’s “Apply AI” strategy highlight the growing tension between Europe’s regulatory leadership and its compute/talent deficit.
- • Brussels Effect confirmed — The EU AI Act is becoming a global standard, with Brazil, Japan, Korea, and Canada adopting its elements. Europe sets the rules, but can it build the technology?
Wat vond je van dit bericht?
