🤖 EU AI & Sovereignty Weekly
August 3, 2026
This week’s EU AI & Sovereignty update. The weekend brought no new Commission publications (the latest item remains July 31), so this edition highlights the most important EU AI and digital sovereignty developments from the past weeks, with AI Act enforcement now in force since August 2 as context.
DMA
🔥 HIGH
Commission orders Google AI interoperability on Android — a challenge to Gemini
Under the Digital Markets Act, the Commission issued two binding specification measures forcing Google to give rival AI services equal access to Android features so they can compete with Gemini, and to give third-party search engines access to search data. This matters because it directly targets the competitive moat of a dominant AI ecosystem — a test case for how Europe keeps AI markets open.
DMA
🔥 HIGH
Commission fines Google €890 million for DMA breaches
Google faces two non-compliance decisions: €460 million for self-preferencing in Google Search and €430 million for anti-steering restrictions on Google Play. It is among the largest DMA fines to date, and signals that Europe is serious about enforcing its rules against dominant platforms that favor their own services.
wetgeving
📌 MEDIUM
New guidance for implementing the Cyber Resilience Act — AI and cybersecurity
The Commission published guidance clarifying which products fall under the Cyber Resilience Act and how to meet reporting and risk-assessment duties, with extra attention to micro and small firms. The Commission stresses that frontier AI models with cyber capabilities make swift, correct implementation all the more urgent, ahead of the September 11 reporting deadline.
Commissie
📌 MEDIUM
New GenAI pilots for public administrations launched
Three pilots supporting generative AI in government services — FLOODS & DROUGHTS, EUNOMIA.AI and EuropAI — signed their grant agreements on July 1 under the ‘GenAI for public administrations’ call. They test trustworthy European GenAI solutions and should produce scalable, interoperable results other public services can adopt, feeding directly into the Apply AI Strategy.
sovereignty
🔥 HIGH
EuropAI — reusable European GenAI for public administrations
EuropAI brings together public authorities and technical partners from the Netherlands, Denmark, Belgium and Luxembourg to develop, procure and test sovereign, trustworthy GenAI solutions that are reusable across European government services. By bundling public demand and experience it aims to cut duplication and strengthen European technological sovereignty.
open-source
🔥 HIGH
EUNOMIA.AI — European open-source AI models on sovereign infrastructure
The 36-month EUNOMIA.AI project unites 33 organisations from 13 member states plus Norway to build trustworthy, human-centred GenAI for public services, using European open-source AI models backed by sovereign infrastructure. It integrates legal, ethical, cybersecurity and data-protection requirements throughout and will publish reusable blueprints and governance models.
technology
📌 MEDIUM
MOSAIC paves the way for a European multilingual information space
The EU-funded MOSAIC pilot built and tested technology letting media share content across language borders while keeping full ownership and control. It delivered AI-driven transcription, translation and subtitling in eight languages — including underserved ones like Catalan, Slovene and Estonian — plus multimodal search and metadata enrichment, supporting the European Democracy Shield.
DSA
🔥 HIGH
Commission: TikTok breaches DSA — unsafe accounts for minors
The Commission sent TikTok preliminary findings that minors’ accounts do not meet Digital Services Act safety standards. Minors can set accounts ‘public’, exposing content to anyone, and content of 16–17-year-olds is recommended via the For You feed to random users. The Commission wants default settings changed so minors’ content is only visible to accepted users.
DSA
📌 MEDIUM
Commission fines AliExpress €550 million for DSA breaches
The Commission fined AliExpress €550 million for failing to assess and mitigate risks around the sale of illegal, unsafe or counterfeit products. Executive Vice-President Virkkunen stressed that scale is no excuse and risks must be identified and addressed systematically; periodic penalty payments could follow if the decision is not complied with.
chips
📌 MEDIUM
Commission approves €659 million German state aid for four semiconductor plants
The Commission approved €659 million in German state aid for four ‘first-of-a-kind’ facilities in the semiconductor value chain. The measure strengthens the EU’s position and autonomy in semiconductors, aligns with the Chips Act, and is a concrete step in Europe’s technological sovereignty strategy that underpins its AI ambitions.
📈 Key Trends
- • AI Act enforcement is now in force (since August 2) — expect the first concrete enforcement actions and new complaints and whistleblower tools in the coming weeks.
- • DMA enforcement is escalating — the Google interoperability and €890 million fines show the Commission is actively dismantling dominant AI ecosystems.
- • Sovereign GenAI is moving from policy to practice — the EuropAI and EUNOMIA.AI pilots mark a shift toward concrete, sovereign AI in the public sector.
- • Cybersecurity and AI are converging — the Cyber Resilience Act guidance explicitly responds to frontier AI models with cyber capabilities.
- • Hardware sovereignty is a strategic priority — €659 million in semiconductor state aid reinforces the EU Chips Act ambitions tied to AI.
Wat vond je van dit bericht?
