Cloud and AI Development Act
Full official title
Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL establishing a framework of measures for strengthening Europe’s cloud and AI ecosystem (Cloud and AI Development Act)open on 5 Oct 2026 Open for answers from 2 Jul 2026 to 11 Nov 2026 (open). Answer on the Commission's site. Topics: Digital economy and society.
What the Commission proposes
The European Commission proposes the Cloud and AI Development Act to establish a framework of measures aimed at strengthening Europe's cloud and artificial intelligence ecosystem. The initiative focuses on three pillars, which include research and innovation for greening infrastructure, industrial investments to triple EU data processing capacity, and actions to increase secure data processing capacities for cloud providers in the EU. Ultimately, the proposal seeks to address infrastructure gaps, reduce critical dependencies on non-European providers, and enhance the Union's digital sovereignty and competitiveness.
- Addresses the EU gap in cloud and AI infrastructure capacity through three distinct pillars.
- Promotes research and innovation to accelerate sustainable cloud and AI computing technologies.
- Aims to triple the EU data processing capacity over the next 5 to 7 years via industrial investments.
- Introduces measures to increase secure data processing capacities for high-critical use cases.
- Establishes common criteria for sovereign cloud computing services to reduce extraterritorial dependencies.
- Promotes open-source solutions and joint public procurement activities across the public sector.
Affected, as named in the text: cloud computing service providers; public authorities; contracting authorities; businesses; researchers.
Summary written from 18 passages of the Commission document by gemini-3.5-flash-lite on 3 Oct 2026; cached and shared by every organisation.
Dossier summary from the portal
The Act is part of the AI Continent initiative. It will address the EU’s gap in cloud and AI infrastructure capacity through three pillars: I) research & innovation for accelerating the greening of AI cloud and compute infrastructure and data centres; II) industrial investments in sustainable cloud and AI capacity, with the aim to triple the EU’s data processing capacity in the next 5 to 7 years; III) actions to increase capacities of secure processing of cloud providers located in the EU.
Commission documents: Opinion of the Regulatory Scrutiny Board (EN, 9 pages); Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL... (EN, 129 pages); ANNEXES to the REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL... (EN, 23 pages); COMMISSION STAFF WORKING DOCUMENT IMPACT ASSESSMENT REPORT Accompanying the... (EN, 105 pages); COMMISSION STAFF WORKING DOCUMENT IMPACT ASSESSMENT REPORT Accompanying the... (EN, 253 pages); COMMISSION STAFF WORKING DOCUMENT EXECUTIVE SUMMARY OF THE IMPACT ASSESSMENT... (EN, 4 pages).
Who has already answered
118 submissions in the snapshot (41 from organisations), portal total 118. Feedback list on the portal.
41 of 41 organisation submissions analysed; the model reads them as: opposes 1 supports 2 wants changes 29 9 held back for review (low confidence).
-
Europeans for Safe Connections NGO, BEL; register: Non-governmental organisations, platforms and networks and similar model reads as: opposes source
- Add a new point to Annex I regarding health impacts and independent methodologies for assessing environmental and public health impacts in accordance with the precautionary principle.
Should the legislative process nevertheless continue, we recommend adding a new point to Annex I: 9. Grand Challenge 9: Health Impacts Development of independent methodologies for assessing environmental and public health impacts, including cumulative impacts and long-term monitoring, in accordance with the precautionary principle.
- Add a new point to Annex I regarding health impacts and independent methodologies for assessing environmental and public health impacts in accordance with the precautionary principle.
-
Transparent Edge Services S.L. Company, ESP; register: Professional consultancies model reads as: wants changes source
- Explicitly include distributed edge computing, content delivery and network security services within the sovereignty framework, assurance levels and procurement provisions.
To ensure a comprehensive approach, CADA should explicitly include distributed edge computing, content delivery and network security services within its sovereignty framework, assurance levels and procurement provisions.
machine translation by the Commission, original in ES - Incorporate a real 'evaluate once, use many times' principle by accepting existing national schemes and certifications as evidence of equivalent controls.
To achieve this, Articles 17 and 20 and Annexes II and III would need to incorporate a real 'evaluate once, use many times' principle.
machine translation by the Commission, original in ES - Extend criteria regarding Union added value to any relevant procurement of cloud, edge and AI services and give them a significant minimum weighting.
In order for them to contribute effectively to reducing strategic dependencies, they should be extended to any relevant procurement of cloud, edge and AI services and be given a significant minimum weighting, adapted to the sensitivity and subject matter of the contract.
machine translation by the Commission, original in ES
- Explicitly include distributed edge computing, content delivery and network security services within the sovereignty framework, assurance levels and procurement provisions.
-
- Expand sustainable computational capacity.
we support the European Commissions objective to expand sustainable computational capacity, strengthen Europes competitiveness in artificial intelligence, and improve the resilience and security of digital infrastructure.
- Strengthen Europe's competitiveness in artificial intelligence.
we support the European Commissions objective to expand sustainable computational capacity, strengthen Europes competitiveness in artificial intelligence, and improve the resilience and security of digital infrastructure.
- Improve the resilience and security of digital infrastructure.
we support the European Commissions objective to expand sustainable computational capacity, strengthen Europes competitiveness in artificial intelligence, and improve the resilience and security of digital infrastructure.
- Expand sustainable computational capacity.
-
The Shift Project Academic or research institution, FRA; register: Think tanks and research institutions model reads as: wants changes source
- Make energy compatibility a core objective so that new capacity aligns with EU energy and climate goals.
-Make energy compatibility a core objective. New capacity should align with EU energy and climate goals. Data centres should not use electricity more carbon-intensive than the local grid mix, including behind-the-meter generation.:
- Integrate cloud, AI and data centre deployment into energy planning by aligning EU and national strategies with NECPs and grid plans.
-Integrate cloud, AI and data centre deployment into energy planning. EU and national strategies should align with NECPs, long-term strategies, grid, water and industrial plans, including electricity budgets reflecting demand, peak load, grid capacity and competing needs.
- Condition fast-track procedures, strategic status and public support on transparency, energy compatibility, and disclosure of physical impacts.
- Condition fast-track procedures, strategic status and public support on transparency and energy compatibility, including grid capacity, low-carbon electricity, water constraints and competing needs. Operators should disclose physical impacts.
- Make energy compatibility a core objective so that new capacity aligns with EU energy and climate goals.
-
- Introduce an equivalence route allowing Annex II criteria to be met by verifiable technical measures.
an equivalence route allowing Annex II criteria to be met by verifiable technical measures (Articles 1618)
- Ensure admissibility and preference of technical evidence in audits.
admissibility and preference of technical evidence in audits (Articles 1921)
- Include procurement criteria rewarding non-transfer data use and verifiable control over AI outputs.
procurement criteria rewarding non-transfer data use and verifiable control over AI outputs (Articles 2933)
- Introduce an equivalence route allowing Annex II criteria to be met by verifiable technical measures.
-
- Bind recognition to an identifiable service and evidence baseline.
1. Bind recognition to an identifiable service and evidence baseline.
- Define objective material-change triggers for reassessment.
2. Define objective material-change triggers.
- Make assurance status machine-readable and time-aware.
3. Make assurance status machine-readable and time-aware.
- Bind recognition to an identifiable service and evidence baseline.
-
- Establish clear and measurable criteria to ensure that investments in Cloud and AI technologies contribute to the common good.
Instead, the CADA should establish clear and measurable criteria to ensure that investments in Cloud and AI technologies contribute to the common good.
machine translation by the Commission, original in DE - Strengthen provisions on open source with binding targets and stronger incentives.
We therefore propose strengthening the provisions on open source, including binding targets in national Cloud and AI strategies, stronger incentives for the development and maintenance of open technologies, and a broader Commons Added Value criterion in public procurement.
machine translation by the Commission, original in DE - Make sustainability a binding condition for the designation of Data Center Acceleration Zones.
Sustainability should be a binding condition for the designation of Data Center Acceleration Zones, rather than merely one factor to be weighed against other considerations.
machine translation by the Commission, original in DE
- Establish clear and measurable criteria to ensure that investments in Cloud and AI technologies contribute to the common good.
-
Open Markets Institute Europe NGO, GRC; register: Think tanks and research institutions model reads as: wants changes source
- Ban generic sovereignty claims that are not backed by audited compliance with the framework.
To address this, generic sovereignty claims that are not backed by audited compliance with the framework should be banned, ensuring procurement authorities have accurate information to work with.
- Have the European Commission directly evaluate cloud providers that meet certain size thresholds.
It would be preferable for the European Commission to directly evaluate cloud providers that meet certain size thresholds, based on turnover, market capitalization, and overall economic weight, and for the Commission to set a minimum, EU-wide list of public-sector activities requiring stronger assurance levels, preventing divergence between Member States.
- Make a multi-cloud strategy mandatory by default for Member States.
Rather than merely asking Member States to consider a multi-cloud strategy, such an approach should become mandatory by default, with single-vendor strategies requiring explicit justification.
- Ban generic sovereignty claims that are not backed by audited compliance with the framework.
-
- Require cloud infrastructure to support the generation of manipulation-proof evidence trails of AI activities to assess compliance with applicable laws.
1. Requiring cloud infrastructure to support the generation of manipulation-proof evidence trails of AI activities, for example via confidential-computing-based methods, to feed into the assessment of compliance with applicable laws;
- Add technologies for AI workload verification and new measures for data-centre security to the objectives of the Cloud and AI Leadership Initiative and create a dedicated Grand Challenge.
2. Adding technologies for AI workload verification and new measures for data-centre security to the objectives of the Cloud and AI Leadership Initiative (Articles 3 and 4), and creating a dedicated Grand Challenge.
- Condition frontier AI priority projects on GPAI Code of Practice aligned development.
1. Conditioning frontier AI priority projects on GPAI Code of Practice aligned development;
- Require cloud infrastructure to support the generation of manipulation-proof evidence trails of AI activities to assess compliance with applicable laws.
-
- Include the oversight of third-party AI systems and platforms in the common cloud and AI curriculum.
Include in the common cloud and AI curriculum the oversight of AI systems an organisation already uses through third-party platforms.
- Ensure national strategies measure the actual results AI delivers to SMEs alongside adoption rates.
National strategies should measure the results AI delivers to SMEs as well as the share of SMEs that use AI.
- Empower Centres for AI to help SMEs access their campaign data and store copies in EU-hosted storage to prevent vendor lock-in.
Centres for AI should help them use it and keep their own copy in EU-hosted storage.
- Include the oversight of third-party AI systems and platforms in the common cloud and AI curriculum.
-
Japan Business Council in Europe (JBCE) Business association, BEL; register: Trade and business associations model reads as: wants changes source
- The Commission should assess third countries under Article 18 in parallel with the legislative procedure so that an implementing act is adopted by the date of application.
The Commission should assess third countries in parallel with the legislative procedure, so that an implementing act can be adopted by the date of application.
- Data localisation should cease to be the general rule at levels 1 and 2 and become an exception.
Localisation should cease to be the general rule at levels 1 and 2 and become an exception: at level 1 where the public sector body explicitly requires it, and at level 2 for data identified as sensitive in the Article 29 risk assessment.
- Residency should suffice under criterion (d) instead of Union citizenship for personnel.
Residency, already accepted under (h) for staff with direct access, should suffice under (d) instead of Union citizenship.
- The Commission should assess third countries under Article 18 in parallel with the legislative procedure so that an implementing act is adopted by the date of application.
-
- Complement cloud and provider-level sovereignty with procurement-ready requirements for AI agent runtimes, hardware-backed trust, and deployment-level conformance tiers.
We recommend complementing this with procurement-ready requirements for AI agent runtimes, hardware-backed trust, and deployment-level conformance tiers.
- Encourage the development of reference procurement clauses for AI agent runtimes covering operating-system integrity, sandboxing, capability-based access control, and auditability.
We recommend that CADA encourage the development of reference procurement clauses for AI agent runtimes, covering operating-system integrity, sandboxing, capability-based access control, and auditability.
- Encourage reference procurement clauses for hardware-backed AI runtime assurance, including hardware root of trust, remote attestation, key custody, memory and I/O isolation.
We recommend that CADA encourage reference procurement clauses for hardware-backed AI runtime assurance, including hardware root of trust, remote attestation, key custody, memory and I/O isolation,
- Complement cloud and provider-level sovereignty with procurement-ready requirements for AI agent runtimes, hardware-backed trust, and deployment-level conformance tiers.
-
Danish Industry Business association, DNK; register: Trade and business associations model reads as: wants changes source
- Ensure requirements for cloud services are risk-based and grounded in concrete security and resilience considerations.
Requirements for cloud services should be risk-based and grounded in concrete security and resilience considerations.
- Provide a clear and narrow definition of the concept of public order to avoid fragmentation and disproportionate requirements.
In particular, DI calls for a clear and narrow definition of the concept of public order to avoid fragmentation and disproportionate requirements.
- Align existing EU legislation closely with CADA to prevent overlapping obligations and unnecessary barriers.
Finally, existing EU legislation should be closely aligned with CADA to ensure that businesses are not faced with overlapping obligations, increased complexity, or unnecessary barriers to innovation, investment and access to technology.
- Ensure requirements for cloud services are risk-based and grounded in concrete security and resilience considerations.
-
- Include hardware within the scope of the annex rather than excluding the hardware substrate from criteria.
Excluding the hardware substrate from criteria builds an "AI Continent" on unpoliced silicon.
- Embed Veritas Core PCIe hardware interposers into Level 3 and 4 cloud server architectures to establish a gatekeeper beneath OS and hypervisor layers.
By embedding Veritas Core PCIe hardware interposers into Level 3 and 4 cloud server architectures, the EU establishes a gatekeeper beneath OS and hypervisor layers.
- Implement substrate-level cryptographic gating and mutual hardware authentication on the federation's sharing platforms.
By implementing substrate-level cryptographic gating and mutual hardware authentication on the federation's sharing platforms, Veritas Core ensures workloads are segregated at the bare-metal level, guaranteeing data confidentiality across the compute continuum.
- Include hardware within the scope of the annex rather than excluding the hardware substrate from criteria.
-
- Require any product storing user data to ship a complete, royalty-free format specification with a reference implementation.
Any product storing user data must ship a complete, royalty-free format specification with a reference implementation.
- Ensure trust-critical storage and code in the data path are auditable and buildable from open specs by EU manufacturers.
Trust-critical storage must be buildable from open specs by EU manufacturers the procurement norm that once forced Intel to license the 8086.
- Require any product storing user data to ship a complete, royalty-free format specification with a reference implementation.
-
- Require attestable runtime identity for AI workloads on supported infrastructure.
1. Require attestable runtime identity for AI workloads on supported infrastructure.
- Make delegated authority expressible and checkable through existing Union rails.
2. Make delegated authority expressible and checkable through existing Union rails.
- Require portability of action-evidence, not only of data and services.
3. Require portability of action-evidence, not only of data and services.
- Require attestable runtime identity for AI workloads on supported infrastructure.
-
German Banking Industry Committee (GBIC) Business association, DEU; register: Non-governmental organisations, platforms and networks and similar model reads as: wants changes source
- Ensure the regulation is technology-open, risk-based, and coherent with existing financial sector regulation.
For the banking sector, it is essential that CADA be designed in a technology-open, risk-based manner and remain coherent with existing financial sector regulation.
- Include an explicit precedence rule in favour of the DORA framework to avoid a parallel compliance layer.
An explicit precedence rule in favour of the DORA framework is required.
- Exclude banks from the scope of the sovereignty framework provisions under Title IV to prevent restricting technology and provider choice.
Including banks within the scope of such provisions would restrict technology and provider choice, increase costs and weaken innovation without necessarily creating a corresponding or meaningful resilience benefit.
- Ensure the regulation is technology-open, risk-based, and coherent with existing financial sector regulation.
-
GDV- the German Insurance Association Business association, DEU; register: Trade and business associations model reads as: wants changes source
- Clarify that CADA is without prejudice to DORA and creates no additional ICT risk or outsourcing obligations for insurers.
CADA is without prejudice to DORA and does not create additional ICT risk management, outsourcing, contractual, reporting, testing or supervisory obli gations for insurance and reinsurance undertakings.
- Ensure Article 31 remains voluntary for all private entities and does not apply to financial entities subject to DORA.
Article 31 should remain voluntary for all private entities and should not apply to financial entities subject to DORA unless expressly chosen by the entity for internal risk management purposes.
- Prevent CADA from resulting in forced migration or mandatory EU-only cloud use where insurers comply with DORA.
CADA should not result in forced migration, mandatory EU-only cloud use or de facto restrictions on the use of global cloud providers where insurers can demonstrate compliance with DORA.
- Clarify that CADA is without prejudice to DORA and creates no additional ICT risk or outsourcing obligations for insurers.
-
- Distinguish between physical data-centre capacity located in Europe and the actual AI capabilities that European entities can access and use.
However, it should distinguish between data-centre capacity located in Europe and AI capabilities that European governments, firms and researchers can reliably access and use.
- Monitor ordinary data-centre capacity, accelerator-dense AI infrastructure, and advanced compute contractually accessible to European users separately from one another.
CADA should separately monitor ordinary data-centre capacity, accelerator-dense AI infrastructure and, where feasible, advanced compute contractually accessible to European users.
- Ensure that projects receiving public support provide tangible European benefits such as guaranteed capacity for European customers during scarcity.
Projects receiving public financing, accelerated procedures or exceptional infrastructure support should provide tangible European benefits where appropriate, including guaranteed capacity for European customers, access during scarcity, support for critical public functions and research, interoperability, and transparency about outages and capacity allocation.
- Distinguish between physical data-centre capacity located in Europe and the actual AI capabilities that European entities can access and use.
-
- Ensure that implementing rules explicitly state that fabless engineering, EDA licenses, IP, software, verification, and tape-outs are eligible for funding.
For the Act's recognition of design to carry through into funding, the implementing rules should state explicitly that this kind of spending is eligible, across the Design Platform, the Chips Fund, Grand Challenges and Strategic Projects, and projects should be assessed on the capability they create in Europe rather than on physical assets.
- Support RISC-V and other open, interoperable architectures through the Design Platform and future Chips JU calls.
Support RISC-V and other open, interoperable architectures through the Design Platform and future Chips JU calls.
- Expand demand-side measures in Articles 8 to 10 to cover systems and include fabless companies, IP providers, and software companies in co-design.
Articles 8 to 10 should cover systems, not only chips, and should bring fabless companies, IP providers and software companies into co-design alongside manufacturers.
- Ensure that implementing rules explicitly state that fabless engineering, EDA licenses, IP, software, verification, and tape-outs are eligible for funding.
-
- Clarify the scope of recognition and ensure public repository entries clearly identify the precise service, configuration, and assurance level covered.
1. Make the scope of recognition unambiguous. Articles 2022 provide for audits, evidence and a public repository. Each repository entry should identify the precise service and configuration covered, assurance level, assessment date, relevant exclusions and current status.
- Operationalise material-change notification with clear guidance on examples, recipients, and risk-sensitive response times.
2. Operationalise material-change notification. Article 23 already requires notification and reassessment where information or changed circumstances may affect recognition. Guidance should define examples, responsible recipients and risk-sensitive response times.
- Require proportionate evidence of recoverability and exit testing for business continuity and migration.
3. Test continuity before migration becomes urgent. Article 29 addresses disruption, migration and consideration of multi-cloud strategies. Its methodology should require proportionate evidence of recoverability and exit: representative restoration and export tests, dependencies, responsible owners, acceptance criteria and fallback arrangements.
- Clarify the scope of recognition and ensure public repository entries clearly identify the precise service, configuration, and assurance level covered.
-
Arq Foundation NGO, BEL; register: Think tanks and research institutions model reads as: wants changes source
- Ensure Acceleration Zones deliver fast power access alongside fast permits, including grid connections and on-site generation.
Make Acceleration Zones deliver fast power in addition to fast permits.
- Incorporate AI model capability as a criterion in the sovereignty framework risk assessments and exceptions.
We suggest that capability should be added as a criterion in the risk assessments, as well as an exception from the assurance levels for mission-critical activities when choosing providers.
- Streamline sovereignty audits by allowing a Union-wide validity based on a single evaluation to avoid multi-state reviews.
We suggest giving the home member state the option to delegate audit evaluation to a dedicated Commission unit whose decision would carry Union-wide validity, removing the 26-state review round and saving administrative capacities of member states.
- Ensure Acceleration Zones deliver fast power access alongside fast permits, including grid connections and on-site generation.
-
- Provide practical assistance for smaller organisations and SMEs regarding project development, computing access, data integration, testing, skills and partnerships.
Centres for AI should therefore provide practical assistance with: * project development; * access to computing resources; * data integration; * AI testing and experimentation; * technical partnerships; * skills and training; * and access to European research and technology providers.
- Provide practical assistance for smaller organisations and SMEs regarding project development, computing access, data integration, testing, skills and partnerships.
-
Deutsche Sozialversicherung Europavertretung NGO, DEU; register: Other organisations, public or mixed entities model reads as: wants changes source
- Ensure that cloud sovereignty requirements are risk-based, practical, and preserve market competition.
The envisaged cloud sovereignty requirements should be risk-based and practical. Particularly for critical infrastructure and sensitive data, high security, control and data location requirements are understandable. At the same time, competition, market availability and the possibility of multi-provider strategies should be preserved.
machine translation by the Commission, original in DE - Provide appropriate transitional periods and grandfathering arrangements for existing IT structures, including in-house data centres.
Appropriate transitional periods and grandfathering arrangements should be provided for existing IT structures. This applies in particular to secure and functioning in-house data centres, which should not be excluded by new requirements.
machine translation by the Commission, original in DE - Ensure close alignment with existing legal acts such as the AI Act, Data Act, GDPR, and NIS2 to avoid double regulation.
In addition, CADA should be closely linked to existing legal acts such as AI Act, Data Act, GDPR and NIS2 to avoid double regulation and conflicting requirements.
machine translation by the Commission, original in DE
- Ensure that cloud sovereignty requirements are risk-based, practical, and preserve market competition.
-
- Explicitly recognise micro, small and medium-sized publishing enterprises among the entities that can access European cloud and AI infrastructures under effective conditions.
The USPI therefore considers it necessary for future European legislation to explicitly recognise publishing and information companies, with a particular focus on micro, small and medium-sized enterprises, among the entities that must be able to access European cloud and artificial intelligence infrastructures under effective and sustainable conditions.
machine translation by the Commission, original in IT - Provide facilitated access to computing infrastructure, AI Factories and cloud services supported by European programmes, including dedicated SME instruments.
In particular, USPI calls for forms of facilitated access to computing infrastructure, AI Factories and cloud services supported by European programmes, including through instruments dedicated to SMEs
machine translation by the Commission, original in IT - Ensure interoperability, data portability and provider switching to avoid vendor lock-in phenomena.
interoperability, data portability and the effective possibility of switching providers should be ensured, avoiding vendor lock-in phenomena that are particularly detrimental to smaller companies.
machine translation by the Commission, original in IT
- Explicitly recognise micro, small and medium-sized publishing enterprises among the entities that can access European cloud and AI infrastructures under effective conditions.
-
- Mandate architectural decoupling by separating semantic generation from an independent, deterministic causal audit engine acting as a security sidecar.
Architectural Decoupling Mandate the strict separation of semantic generation (front-end model) and cognitive auditing (back-end core).
- Implement a native Responsibility Anchor system within cloud AI infrastructures requiring unalterable cryptographic tags for high-risk state updates.
Minimal Decision Accountability Implement a native Responsibility Anchor system within cloud AI infrastructures.
- Map compliance logic directly onto sovereign cloud hardware using Processing-In-Memory architectures to solidify audit rules into immutable hardware execution cycles.
Physical Substrate Compliance Map compliance logic directly onto sovereign cloud hardware via Processing-In-Memory (PIM) architectures.
- Mandate architectural decoupling by separating semantic generation from an independent, deterministic causal audit engine acting as a security sidecar.
-
DigitalTrade4.EU Other, EST; register: Other organisations, public or mixed entities model reads as: wants changes source
- Support cross-border demonstrators connecting existing systems through brownfield interoperability instead of requiring wholesale data or infrastructure migration.
We recommend that the Cloud and AI Leadership Initiatives support cross-border demonstrators connecting existing systems through brownfield interoperability rather than requiring wholesale data or infrastructure migration.
- Ensure AI outputs act as derived analytical evidence rather than replacing authoritative records or regulated decisions.
AI outputs should remain derived analytical evidence rather than replacing authoritative records or regulated decisions.
- Expand funding instruments like Horizon Europe and the Digital Europe Programme to support these demonstrators and their progression toward scalable deployment.
Horizon Europe, the Digital Europe Programme and subsequent EU funding instruments should support these demonstrators and their progression towards scalable deployment.
- Support cross-border demonstrators connecting existing systems through brownfield interoperability instead of requiring wholesale data or infrastructure migration.
-
- Strengthen European digital autonomy and sovereignty.
Dutch local and regional authorities are strong advocates of the introduction of new European legislation that would significantly strengthen the digital autonomy and sovereignty of the European Union and its entities.
- Address concerns regarding specific elements in the current proposal.
Dutch municipalities and provinces welcome several aspects of the proposed Cloud and AI Development Act (CADA), but also have concerns regarding a number of elements in the current proposal
- Strengthen European digital autonomy and sovereignty.
-
Information Technology Industry Council (ITI) Business association, USA; register: Trade and business associations model reads as: wants changes source
- Ensure that trust is assessed through objective, risk-based criteria rather than ownership structure or headquarters location.
Trust should be assessed through objective, risk-based criteria rather than reduced to a providers ownership structure or headquarters location.
- Apply trusted-vendor principles based on governance and risk management across the ICT supply chain.
Applying these trusted-vendor principles across the ICT supply chain would support secure technology ecosystems while strengthening supplier due diligence and the proportionate management of subcontractor risks.
- Ensure that trust is assessed through objective, risk-based criteria rather than ownership structure or headquarters location.
-
- Recognise that sovereign verification and session continuity can be decoupled architecturally in relation to AI 1.5.
AI 1.5: Recognise that sovereign verification and session continuity can be decoupled architecturally; regulatory outcomes should not penalise compliant operators.
- Ensure the regulatory framework leaves room for decentralised federation-based interworking protocols for AI 1.13.
AI 1.13: Leave room for decentralised federation-based interworking protocols.
- Commission or encourage a CEPT study on voluntary, cryptographically verifiable interference attribution.
Interference: Commission or encourage a CEPT study on voluntary, cryptographically verifiable interference attribution, referencing open-source implementations such as GFSIP/1.0.
- Recognise that sovereign verification and session continuity can be decoupled architecturally in relation to AI 1.5.
-
- Recognise that sovereign verification and session continuity can be decoupled architecturally in relation to AI 1.5.
1. AI 1.5: Recognise that sovereign verification and session continuity can be decoupled architecturally; regulatory outcomes should not penalise compliant operators.
- Leave room for decentralised federation-based interworking protocols in relation to AI 1.13.
2. AI 1.13: Leave room for decentralised federation-based interworking protocols.
- Commission or encourage a CEPT study on voluntary, cryptographically verifiable interference attribution.
3. Interference: Commission or encourage a CEPT study on voluntary, cryptographically verifiable interference attribution, referencing open-source implementations such as GFSIP/1.0.
- Recognise that sovereign verification and session continuity can be decoupled architecturally in relation to AI 1.5.
-
Adigital (Asociación Española de la Economía Digital) Business association, ESP; register: Trade and business associations model reads as: supports source
Who met the cabinets on the subject
60 meetings by subject similarity, approximate
subject match, approximate The top 60 cabinet meetings by similarity (the list stops at 60) whose recorded subject is close to this consultation's title and summary (embedding similarity at 0.45 or more, cached from the nightly map). Generic subjects such as "exchange of views on EU regulation" also pass, so read this as a lead, not a finding. Source: Commission meetings file, portal.
| Date | Cabinet | Interest representative | Subject as recorded |
|---|---|---|---|
| 13 Jul 2026 | President Ursula von der Leyen | Domyn SpA (Companies & groups) | Development of AI infrastructure, cloud services and AI applications in the EU |
| 16 Jan 2026 | Executive Vice-President Henna Virkkunen | UpCloud Oy (Companies & groups) | EU Cloud policy |
| 3 Mar 2025 | Executive Vice-President Henna Virkkunen | OVH Groupe (Companies & groups) | EU Cloud policy |
| 11 Feb 2026 | Executive Vice-President Henna Virkkunen | The Open Cloud Coalition (Trade and business associations) | EU Cloud Policy |
| 26 Jan 2026 | Executive Vice-President Henna Virkkunen | Equinix (EMEA) Holdings B.V. (Companies & groups) | EU Cloud Policy |
| 22 Jan 2026 | Executive Vice-President Henna Virkkunen | European Sovereign Tech Industry Alliance (Non-governmental organisations, platforms and networks and similar) | EU Cloud Policy |
| 22 Sep 2025 | Executive Vice-President Henna Virkkunen | OVH Groupe (Companies & groups) | European Cloud Policies |
| 13 Oct 2025 | Executive Vice-President Henna Virkkunen | QTS Data Centers (Companies & groups) | European Cloud Policy |
| 14 May 2025 | Executive Vice-President Henna Virkkunen | OVH Groupe (Companies & groups) | European cloud policies |
| 20 Nov 2025 | President Ursula von der Leyen | OVH Groupe (Companies & groups) | To discuss European Cloud and Data Strategy |
| 11 Feb 2025 | Executive Vice-President Henna Virkkunen | France Digitale (Trade and business associations) | EU agenda on AI and tech sovereignty |
| 10 Feb 2025 | Executive Vice-President Henna Virkkunen | ASML Netherlands B.V. (Companies & groups) | EU agenda on AI and tech sovereignty |
| 10 Feb 2025 | Executive Vice-President Henna Virkkunen | Iliad (Companies & groups) | EU agenda on AI and tech sovereignty |
| 10 Feb 2025 | Executive Vice-President Henna Virkkunen | Google Ireland Limited and its affiliates (Companies & groups) | EU agenda on AI and tech sovereignty |
| 10 Nov 2025 | Executive Vice-President Stéphane Séjourné | Infosys (Companies & groups) | Implementation of the EU’s AI Act |
| 23 Jul 2025 | Executive Vice-President Stéphane Séjourné | CoreWeave UK Ltd (Companies & groups) | Exchange views on cloud and AI infrastructure developments discuss the evolving regulatory landscape, and explore how CoreWeave’s capabilities can support Europe’s digital and technological ambitions. |
| 20 Mar 2025 | Executive Vice-President Stéphane Séjourné | European AI Forum (Non-governmental organisations, platforms and networks and similar); Bundesverband der Unternehmen der Künstlichen Intelligenz in Deutschland e.V. (Trade and business associations) | European AI policy and Europe's digital sovereignty |
| 19 Mar 2025 | Executive Vice-President Henna Virkkunen | Technology Industries of Finland (Teknologiateollisuus ry) (Trade and business associations) | Advancing AI and Data Strategies in Europe: Legislation, Innovation, and Economic Growth |
| 17 Sep 2026 | Executive Vice-President Henna Virkkunen | HOTREC, Hotels, Restaurants & Cafés in Europe (Trade and business associations) | Cloud and AI Development Act |
| 25 Jun 2026 | Executive Vice-President Henna Virkkunen | European DIGITAL SME Alliance (Trade and business associations) | Cloud and AI Development Act |
| 25 Jun 2026 | Executive Vice-President Henna Virkkunen | Broadcom (Companies & groups) | Cloud and AI Development Act |
| 23 Apr 2026 | Executive Vice-President Henna Virkkunen | ServiceNow (Companies & groups) | Cloud and AI Development Act |
| 26 Mar 2026 | Executive Vice-President Henna Virkkunen | Telecom Italia Sparkle S.p.A. (Companies & groups) | Cloud and AI Development Act |
| 10 Mar 2026 | Executive Vice-President Henna Virkkunen | Atlassian (Companies & groups) | Cloud and AI Development Act |
| 6 Nov 2025 | President Ursula von der Leyen | Anthropic, PBC (Companies & groups) | EU's approach to AI & its impact on the EU's economy & competitiveness |
| 20 May 2026 | Executive Vice-President Teresa Ribera Rodríguez | Microsoft Corporation (Companies & groups) | EU policies on digital markets, in particular on AI |
| 5 Dec 2025 | Executive Vice-President Henna Virkkunen | FGS Global (Europe) GmbH; Arm Limited (Companies & groups) | AI and Cloud policies |
| 1 Jun 2026 | Executive Vice-President Stéphane Séjourné | ChapsVision (Companies & groups) | Forthcoming deployment of AI tools at the European Commission, the Tech Sovereignty Package, and funding opportunities for AI companies. |
| 24 Jun 2025 | Executive Vice-President Henna Virkkunen | OpenAI OpCo, LLC (Companies & groups) | EU AI agenda and Open AI activities |
| 13 Apr 2026 | Executive Vice-President Teresa Ribera Rodríguez | Centre for Democracy & Technology, Europe (Non-governmental organisations, platforms and networks and similar) | EU policies and regulatory enforcement in the digital and AI areas |
Showing 30 of 60.