Summary The proposed Cloud and AI Development Act (CADA) establishes a dedicated framework to accelerate autonomous driving in the EU by explicitly prioritizing the development of AI models for automotive applications, including software-defined vehicles. As proposed, the regulation mandates Member States to facilitate the testing and deployment of these systems through a network of Experience and Acceleration Centres for AI (Centres for AI). Recital 19 specifically highlights the need for Member States to cooperate with the automotive industry, suppliers, cities, and regions to enable the "safe and trustworthy deployment of AI-enabled connected and autonomous mobility solutions" across diverse European environments. For technical leaders, this creates a structured pathway to access compute resources, validate physical AI in real-world settings, and leverage public-sector support for large-scale deployment.

Detail

The Cloud and AI Development Act (COM(2026) 502 final) is a legislative proposal designed to strengthen the EU's cloud and AI ecosystem. While it addresses broad infrastructure and sovereignty goals, it contains specific, high-impact provisions for the automotive sector. The proposal recognizes that AI is no longer merely a software layer but a critical infrastructure component for next-generation mobility, requiring a coordinated approach to data, compute, and testing environments.

Strategic Prioritization of Automotive AI

Under the proposed framework, the development of AI for the automotive sector is integrated directly into the EU's strategic industrial policy rather than being left to market forces alone. Article 4(7)(e) of the proposal explicitly lists "facilitating the development, testing and deployment of AI models and tools in the automotive sector, including for autonomous driving" as a specific operational objective of the Cloud and AI Leadership Initiatives.

This strategic focus is further elaborated in the explanatory recitals. Recital 19 states that advancements in the automotive sector should support the development, testing, and deployment of innovative software platforms contributing to Union industrial leadership in "software defined vehicles and autonomous driving." The recital emphasizes that the Cloud and AI Leadership Initiatives should reduce obstacles to testing and deploying AI models, particularly within cities and regions, to foster Union leadership in these areas.

Crucially, Recital 19 outlines a specific cooperation model: "Member States should facilitate the development, testing and deployment of AI systems for autonomous driving, including through cooperation with the Centres for AI, the automotive industry, suppliers, cities and regions, with a view to enabling the safe and trustworthy deployment of AI-enabled connected and autonomous mobility solutions across diverse European environments." This clause creates a legal expectation for Member States to actively coordinate between public authorities (cities/regions) and private actors (industry/suppliers) to create the necessary testbeds.

The Role of Experience and Acceleration Centres for AI

A key mechanism for delivering this support is the establishment of Experience and Acceleration Centres for AI (Centres for AI). Article 5 requires each Member State to establish these centres, which build on the existing network of European Digital Innovation Hubs. These centres act as regional and local accelerators for the uptake of AI and cloud computing technologies.

For the automotive industry, Recital 19 specifies that Member States should facilitate the development, testing, and deployment of AI systems for autonomous driving through cooperation with these Centres for AI, the automotive industry, suppliers, cities, and regions. The goal is to enable the "safe and trustworthy deployment of AI-enabled connected and autonomous mobility solutions across diverse European environments."

This means that CTOs and technical architects can expect a decentralized network of support structures. These centres are tasked with:

  • Helping organizations accelerate digital transformation through access to AI technologies.
  • Connecting organizations with European providers of cloud and AI technologies.
  • Ensuring access to upskilling and reskilling schemes.
  • Facilitating the transfer of expertise across regions.

By mandating cooperation with "cities and regions," the proposal acknowledges that autonomous driving cannot be tested in isolation; it requires real-world urban and regional infrastructure. The Centres for AI would serve as the interface between automotive developers and local authorities to unlock these testing environments.

Support for Physical AI and Industrial AI

Autonomous driving relies heavily on "Physical AI," which refers to AI systems capable of perceiving the physical environment and executing complex actions within it. Article 4(4) of the proposal outlines operational objectives to advance Union capabilities in physical AI models and systems. This includes:

  • Accelerating the development of a European physical AI stack.
  • Facilitating access to specific datasets for physical AI.
  • Supporting the development, testing, and validation of physical AI models and systems in real-world environments.

Furthermore, Article 4(5) focuses on industrial AI, aiming to accelerate the development and uptake of sectoral AI models across strategic industrial sectors. For automotive companies, this implies targeted support for accessing the necessary computing resources and AI tools required to develop and operationalize AI models tailored to industrial sector needs, such as manufacturing and logistics.

Compute Resource Allocation

The proposal also addresses the critical bottleneck of computational capacity. Article 9 establishes that the Union and Member States shall ensure that sufficient AI computing resources are allocated to support the development of frontier AI priority projects, as well as AI industrial innovation, physical AI, and public sector AI projects. This ensures that European companies working on autonomous driving technologies have access to the high-performance computing (HPC) capacity necessary for training complex models.

Sovereignty and Supply Chain Resilience

While the primary focus here is on development, CADA also addresses the broader context of technological sovereignty. The proposal aims to reduce dependencies on third-country providers for cloud computing services. For the automotive sector, this means encouraging the use of European cloud and AI stacks. Article 4(2) supports the development of cloud computing stacks supporting the Union's technological autonomy, including AI-optimized servers and software based on processors designed and manufactured in the Union.

What this means for you

For CTOs, architects, and SMEs in the automotive sector evaluating the practical impact of CADA, the proposal offers several concrete opportunities and implications:

  1. Access to Testing Infrastructure: If you are developing autonomous driving algorithms, you will have access to a formalized network of Centres for AI. These centres are designed to help you navigate regulatory hurdles and connect with local suppliers and cities willing to participate in testing. This is particularly valuable for SMEs that may lack the resources to build their own extensive testing grounds.
  2. Compute Availability: The commitment to allocate sufficient AI computing resources for physical AI and industrial AI projects means that access to HPC clusters may become more streamlined for projects aligned with the EU's strategic objectives. You should monitor calls for expressions of interest for "frontier AI priority projects" or industrial AI initiatives.
  3. Collaboration Opportunities: The emphasis on cooperation between industry, cities, and regions suggests that public-private partnerships will be encouraged. Engaging with local Centres for AI can help you identify pilot projects for connected and autonomous mobility in diverse European environments, which is crucial for validating the robustness of your AI systems.
  4. Supply Chain Considerations: As the EU pushes for technological autonomy, there may be incentives or requirements to use European cloud services and hardware. Architects should begin evaluating the feasibility of migrating workloads to EU-based sovereign cloud providers and assessing the compatibility of their software stacks with European hardware initiatives.
  5. Data and Skills: The proposal highlights the need for high-quality data and skilled personnel. Centres for AI will facilitate access to upskilling schemes, which can help address talent shortages in AI and autonomous driving development.

Common misconceptions

Misconception 1: CADA directly regulates autonomous vehicles. CADA does not set technical safety standards for autonomous vehicles or regulate the deployment of self-driving cars on public roads. That falls under other EU legislation, such as the type-approval regulations for motor vehicles. CADA focuses on the underlying cloud, AI, and data infrastructure that enables the development of these technologies.

Misconception 2: Only large hyperscalers will benefit. While large companies have significant resources, CADA explicitly aims to support SMEs and small mid-caps. The establishment of Centres for AI and the focus on reducing barriers for SMEs in public procurement and innovation initiatives are designed to level the playing field.

Misconception 3: CADA replaces the AI Act. CADA complements the AI Act. The AI Act sets out harmonized rules for the placing on the market and use of AI systems, including high-risk systems. CADA focuses on the supply side: boosting domestic capabilities, increasing compute capacity, and fostering innovation. They are distinct but complementary instruments.

Misconception 4: Autonomous driving development is solely a software issue. CADA recognizes the importance of the hardware-software stack. The proposal supports the development of AI-optimized servers and processors designed in the Union, acknowledging that the performance and sovereignty of autonomous driving systems depend on the entire technology stack, not just the algorithms.

Official sources

Related

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