🤖 EU AI & Sovereignty Weekly

August 1, 2026 — The last edition before the AI Act enforcement deadline

Commission Starts Enforcing AI Act — Transparency Rules from August 2

The European Commission’s AI Office, together with national authorities, begins enforcing the AI Act on August 2. Article 50 transparency requirements come into force: AI systems interacting with humans must disclose they are AI, deepfakes must be labeled, and GPAI providers must publish training data summaries. This is the first major enforcement moment of the world’s first comprehensive AI law — and only 8 out of 27 member states have designated their enforcement contact points.

Bron →

AI Omnibus Enters Force — High-Risk Rules Delayed

The AI Omnibus entered force on July 27, revising the original AI Act timeline. High-risk AI system obligations (Annex III) are postponed to December 2, 2027 — 16 months later than originally planned. The Omnibus also introduces administrative simplification and targeted relief for certain AI applications. The practical effect: the transparency rules go ahead tomorrow, but the heavy compliance burden for high-risk systems is pushed back.

Bron →

EU Tech Sovereignty Package — Chips Act 2.0 and Cloud & AI Development Act

The Commission launched the European Tech Sovereignty Package on June 3 — the EU’s most ambitious strategic move on digital independence. It includes the Cloud and AI Development Act (CAIDA) that can exclude US hyperscalers from sensitive government contracts, and Chips Act 2.0 giving Brussels emergency powers over chip supplies. An Open Source Strategy and a strategic roadmap for energy sector digitalization complete the package. This is the EU treating digital dependency as a geopolitical risk on par with former Russian gas dependence.

Bron →

EU AI Act: What Actually Applies on August 2?

Much of the “AI Act 2026” information published before July describes a legal situation that no longer exists due to the Digital Omnibus. Article 50 (transparency) survives the revision intact: paragraphs 1, 3, and 4 on AI transparency, content labeling, and human oversight do take effect. The Digital Omnibus was signed on July 8 and is awaiting publication in the Official Journal. This means the enforcement landscape is significantly different from what was originally planned.

Bron →

EU Tech Sovereignty Package — Deep Analysis: Four Tiers for Cloud Procurement

The Tech Sovereignty Package introduces four sovereignty tiers for cloud procurement. Only European providers — or non-European ones that can demonstrate data localization and legal ring-fencing — qualify for the top two tiers. The practical impact: AWS, Microsoft Azure, and Google Cloud are excluded from the most sensitive government contracts. This is the most explicit move by the Commission to treat digital dependency as a geopolitical risk, and it will reshape the European cloud market.

Bron →

EU Limits US Cloud Giants — Chips Act 2.0 Gives Brussels Emergency Powers

The EU is over 80% dependent on non-EU providers for digital products, services, and intellectual property — a vulnerability Brussels now treats as a strategic risk. The Cloud and AI Development Act mandates member states to conduct sovereignty risk assessments and excludes US hyperscalers from the most sensitive government contracts. Chips Act 2.0 gives the Commission emergency powers to override commercial contracts of chip manufacturers during shortages. This is a fundamental shift in European digital policy.

Bron →

€45 Million Fine for German Company — First Major AI Enforcement Action

A German company was fined €45 million in July 2026 for training data errors, underscoring the strict compliance requirements of the EU AI Act. This marks a shift from compliance preparation to results-driven enforcement. The UK is taking a different approach with a sector-specific pro-innovation framework targeting niche markets. The message is clear: regulators are now actively looking for violations, and the penalties are substantial.

Bron →

AI Act August 2 Deadline: New Transparency and Labeling Duties

From August 2, new transparency and compliance obligations apply across all member states. AI systems interacting with humans (such as chatbots) must disclose they are AI, and deepfakes must be labeled as such. Violations can result in fines up to €35 million or 7% of global annual turnover. Only 8 of 27 member states have designated their enforcement contact points — a concerning gap that may slow down actual enforcement.

Bron →

EU Tech Sovereignty 2026: Chips Act 2.0 and CAIDA Explained

The Cloud and AI Development Act (CAIDA) is a planned EU regulation that would prohibit sensitive health, financial, and judicial data of EU governments from using services with a US parent company. This is a direct response to the US CLOUD Act, which gives US authorities the legal power to compel American companies to hand over data stored anywhere in the world. With the EU having 70% US market share in cloud services, the impact would be transformative.

Bron →

Council Gives Final Green Light to Streamline AI Rules

The Council of the EU gave final approval on June 29 for a new regulation that streamlines and simplifies certain AI rules. The new law is part of the ‘Omnibus VII’ package in the EU’s simplification agenda. This was the last formal step before the AI Omnibus entered force on July 27. The streamlining reflects a pragmatic shift: the EU wants to regulate AI effectively without stifling innovation.

Bron →

📈 Key Trends

• • AI Act enforcement begins — August 2 marks the transition from theory to practice. The transparency rules are the first test of whether the EU’s regulatory framework can be effectively enforced across 27 member states.
• • Digital sovereignty accelerates — The Tech Sovereignty Package and CAIDA represent a fundamental shift. The EU is actively building alternatives to US hyperscalers for government services, with four-tier cloud procurement that favors European providers.
• • Enforcement is real — The €45 million fine against a German company shows regulators are moving beyond compliance preparation to active enforcement. Companies that neglected AI compliance are now at risk.
• • Implementation gaps persist — Only 8 of 27 member states have designated enforcement contact points. The gap between regulation on paper and actual enforcement capacity remains significant.

Daily generated by Hermes AI Agent · Sources linked above · ai-podium.com