Summary Under the proposed Cloud and AI Development Act (CADA), Grand Challenge 2 is the strategic pillar dedicated to "Building end-to-end hardware and software cloud stacks, including AI tools, infrastructure, services and management layers to bridge the Union's critical capacity gaps" (Annex I(2)). As proposed, this initiative specifically targets the development of "AI servers powered by semiconductors and quantum technologies designed and manufactured in the Union for distributed and decentralised cloud and edge computing for AI." This challenge directly operationalises Article 4(2) of the proposal, which mandates the Cloud and AI Leadership Initiatives to develop secure, resilient open cloud computing stacks and AI-optimised servers based on EU-designed processors, accelerators, and quantum technologies.

Detail

Grand Challenge 2 represents a fundamental shift in the EU's approach to digital infrastructure, moving beyond software interoperability to secure the entire hardware-software vertical. It is one of eight "grand challenges" identified in Annex I of the CADA proposal, designed to address major technological and industrial challenges of strategic relevance for the Union (Article 6(2)).

The Core Definition: End-to-End Sovereignty

The proposal defines the scope of Grand Challenge 2 explicitly in Annex I(2):

"Building end-to-end hardware and software cloud stacks, including AI tools, infrastructure, services and management layers to bridge the Union's critical capacity gaps."

This definition underscores that the challenge is not limited to the application layer or the management plane. It encompasses the full stack, from the physical hardware (servers, storage, networking) to the software layers (operating systems, middleware, AI tools). The primary objective is to reduce the Union's dependence on third-country providers by creating a self-sufficient ecosystem capable of supporting the massive computational demands of frontier AI.

Hardware Focus: AI Servers and EU Semiconductors

A distinctive feature of Grand Challenge 2 is its explicit focus on hardware sovereignty. Annex I(2) states that the initiative involves:

"building AI servers powered by semiconductors and quantum technologies designed and manufactured in the Union for distributed and decentralised cloud and edge computing for AI."

This requirement links CADA directly to the EU's semiconductor strategy. The proposal envisions AI servers that are not merely assembled in Europe but are powered by processors, accelerators, and quantum technologies that are designed and manufactured within the Union. This is intended to ensure that the critical compute resources underpinning the EU's AI ambitions are resilient to supply chain disruptions and external geopolitical pressures. The challenge also highlights the need for "next-generation ultra-high density and long-term data storage," addressing the physical constraints of data growth.

Operational Objective 2: The Legal Mechanism

Grand Challenge 2 is the strategic umbrella for operational objective 2 set out in Article 4(2) of the CADA proposal. While Annex I defines the what, Article 4(2) defines the how through specific operational actions the Cloud and AI Leadership Initiatives must undertake.

Article 4(2) mandates the following specific measures to achieve the cloud stack objective:

  1. Open Cloud Stacks: The Initiatives must "develop and pilot secure, resilient and performant open cloud computing stacks covering on-device edge, connectivity, data and AI tools, backend and service layers for strategic sectors" (Article 4(2)(a)). This ensures that the stacks are not proprietary black boxes but are built on open standards to prevent vendor lock-in.
  2. EU-Designed Hardware Integration: The proposal requires the development of "AI-optimised servers and baseline software based on processors, accelerators and quantum accelerators designed and manufactured in the Union, alongside next-generation ultra-high density and long-term data storage" (Article 4(2)(b)). This directly translates the hardware ambition of Annex I(2) into a legal obligation for the Initiatives.
  3. Data Availability via Middleware: To support AI development, the Initiatives must "boost data availability for AI via open-source middleware platforms underpinning common European data spaces" (Article 4(2)(c)). This connects the cloud stack to the broader data strategy, ensuring that the infrastructure can effectively serve the data spaces required for AI training and inference.
  4. Open-Source Foundations: The proposal mandates fostering "the creation of open-source software foundations supporting open-source components" (Article 4(2)(d)). This is critical for ensuring that the software layers of the stack remain transparent, auditable, and adaptable.
  5. Cataloguing Solutions: Finally, the Initiatives must "establish a catalogue of European open cloud computing solutions developed under points (a) to (d) of this paragraph" (Article 4(2)(e)). This catalogue will serve as a reference for public and private sector bodies seeking sovereign alternatives.

Synergies and Co-Design

The proposal emphasizes that these stacks must be developed through "co-design and cross-optimisation of hardware and software development" (Recital 14). This approach is intended to ensure that the AI-optimised servers are not just generic hardware running generic software, but are specifically tuned for AI workloads. The proposal also notes that the Cloud and AI Leadership Initiatives should work in synergy with the "Chips for Europe Initiative 2.0" to foster the development of these technologies (Recital 14).

The emphasis on "open" in Article 4(2) is not merely rhetorical. Recital 15 clarifies that the Initiatives should "promote the development of technologies relying on open standards, open specifications and open source" to foster innovation and resilience. This aligns with the broader CADA objective of using open source as a lever for technological sovereignty (Recital 15).

What this means for you

For technology leaders, infrastructure providers, and investors, Grand Challenge 2 signals a massive, state-backed push to re-engineer the EU's cloud infrastructure from the silicon up.

For CTOs and Cloud Architects

If you are responsible for infrastructure strategy, Grand Challenge 2 suggests that the future EU cloud landscape will be defined by interoperability and sovereignty.

  • Pilot Opportunities: The proposal explicitly mentions "pilot programmes" to demonstrate the capabilities of European open cloud stacks in strategically important sectors (Annex I(2)). Organizations involved in strategic sectors (e.g., healthcare, energy, defence) should monitor these pilots for early access to sovereign infrastructure.
  • Stack Integration: The focus on "on-device edge, connectivity, data and AI tools" (Article 4(2)(a)) implies that future EU-backed solutions will be modular. You may be able to integrate EU-based edge layers with backend services more seamlessly, reducing the friction of multi-cloud strategies.
  • Hardware Constraints: Be prepared for a shift towards AI servers powered by EU-designed semiconductors. If your current infrastructure relies heavily on non-EU hardware, you may face a transition period where new, sovereign alternatives become available and potentially preferred in public procurement.

For SMEs and Startups

Grand Challenge 2 is designed to lower barriers to entry for smaller players.

  • Modular Participation: You do not need to build a full data centre to participate. The proposal's focus on "open-source software foundations" and "open cloud computing stacks" (Article 4(2)) allows SMEs to contribute specific layers, such as edge AI tools, data middleware, or management layers.
  • Funding and Catalogue: By developing components that align with Article 4(2)(b) (e.g., software optimised for EU processors) or Article 4(2)(c) (middleware for data spaces), SMEs may gain access to funding via the Cloud and AI Leadership Initiatives. Furthermore, inclusion in the "catalogue of European open cloud computing solutions" (Article 4(2)(e)) could provide significant market visibility and credibility.

For Hardware and Semiconductor Providers

There is a clear, legally backed demand signal for EU-designed chips.

  • AI-Optimised Hardware: The mandate to develop "AI-optimised servers... based on processors, accelerators and quantum accelerators designed and manufactured in the Union" (Article 4(2)(b)) creates a specific market for hardware that is not just assembled in Europe, but designed and manufactured there.
  • Quantum Integration: The explicit mention of "quantum technologies" (Annex I(2) and Article 4(2)(b)) indicates that quantum computing is a priority area for integration into the cloud stack, offering opportunities for quantum hardware and software providers.

Common misconceptions

Misconception 1: Grand Challenge 2 is only about software. It is a common error to assume "cloud stacks" refers solely to software layers like Kubernetes or OpenStack. Annex I(2) explicitly includes "hardware" and specifies "AI servers powered by semiconductors and quantum technologies designed and manufactured in the Union." The challenge is holistic, aiming to secure the entire vertical from the silicon up to the AI management layers.

Misconception 2: This replaces the AI Act's requirements. Grand Challenge 2 is an industrial and capacity-building measure, not a regulatory compliance framework. It does not replace the safety, transparency, or fundamental rights obligations of the EU AI Act (Regulation (EU) 2024/1689). Instead, it provides the underlying sovereign infrastructure on which compliant AI systems can be deployed. The AI Act regulates the behavior and risk of AI systems, while CADA's Grand Challenge 2 seeks to build the sovereign infrastructure that hosts them.

Misconception 3: Only hyperscalers can participate. While large players may have the resources to build full stacks, the proposal's focus on "open-source software foundations" and "open cloud computing stacks" (Article 4(2)) is designed to allow modular participation. SMEs and startups can contribute specific layers, such as edge connectivity tools or open-source middleware, without needing to own data centers. The establishment of a "catalogue of European open cloud computing solutions" (Article 4(2)(e)) is intended to highlight these diverse contributions.

Misconception 4: The EU will mandate the use of these stacks immediately. CADA is a proposal, and these measures are primarily about fostering development and deployment through incentives, research, and strategic projects. While public procurement rules (Article 32) may favor solutions with EU added value, there is no immediate mandate for all private sector entities to switch to these specific stacks. The goal is to make them viable, competitive, and available by 2035, not to force an overnight migration.

Official sources

Related

This is general information about a draft EU regulation, not legal advice.