Europe’s AI expansion risks deepening foreign dependence, EESC warns

The EESC argues that the Commission’s Cloud and AI development plan relies on funding that has not been secured, does not determine whether Europe can generate enough electricity to power new data centres and gives affected communities too little say

The EU’s drive to expand its cloud and artificial intelligence capacity could leave Europe even more dependent on foreign technology companies unless funding and public procurement are used to support European providers, the European Economic and Social Committee (EESC) has warned.

In an opinion adopted on 23 September, the EESC backed the ambition of the European Commission’s proposed Cloud and AI Development Act, which aims to strengthen Europe’s digital infrastructure and reduce reliance on non-European companies. However, the EESC was critical of the proposal in its current form, saying it lacked secure financing, a plan to ensure that electricity generation could keep pace with the planned expansion, and the social and environmental safeguards needed to deliver on that promise.

At the heart of the Commission’s plan is a target to at least triple the EU’s data-centre capacity by 2030. The EESC warned that building more infrastructure in Europe would not automatically deliver greater technological sovereignty. Without public contracts for sovereign European-owned and operated services, new investment could end up benefiting the large non-EU providers that already dominate the market.

The EESC called for the ‘Union added value’ criterion in public procurement to be strengthened, increasing its proposed weighting from 15 to 40 points out of 120. It said the Commission’s non-binding 15-point weighting was too weak to influence purchasing decisions, reverse dependence on non-EU hyperscalers or generate demand for European-owned and operated cloud and AI services.

It also called for a proposed target for awarding at least 25% of relevant public contracts to innovative small and medium-sized companies to become a shared requirement rather than an aspiration. Large contracts should be divided into smaller lots where possible, while administrative requirements should be simplified to make it easier for smaller firms to bid.

Financing should be secured rather than assumed

Financing is another major concern. According to the EESC, the proposal sets ambitious targets without guaranteeing the investment needed to meet them. Much of the funding would depend on existing EU programmes or decisions still to be taken under the next EU long-term budget.

‘We welcome the Commission’s aim to strengthen Europe’s cloud and AI capacity and reduce our dependence on non-European suppliers. But it’s not enough to talk about ambition. If we want to expand Europe’s data-centre capacity substantially, we need to know how this will be financed,’ rapporteur of the opinion Miroslav Hajnoš said.

The EESC is calling for cloud and AI infrastructure to be made a clear funding priority and for the Commission to publish an overview of the EU financing available before Member States finalise their national strategies. The European Investment Bank should help attract more private European investment into data centres and cloud infrastructure.

Faster connections are no substitute for adequate power generation

The EESC also questioned whether Europe could generate enough electricity to support a threefold increase in data-centre capacity by 2030. Faster grid connections would not solve a shortage of power, it warned. The EESC called on the Commission to assess whether electricity generation can keep pace and proposed requiring grid operators to provide developers with information on available capacity, connection costs and schedules within six months of a request.

Local communities should have a meaningful say

The EESC also stated that the communities hosting data centres need stronger protection. Such facilities can place pressure on local electricity and water supplies and generate noise and heat, yet the proposal does not require Member States to assess or mitigate the health risks before designating data-centre acceleration zones (areas where new facilities could be approved more quickly). National authorities should carry out environmental, social and health impact assessments before approving such areas, the EESC said.

Local residents should also have the right to take part in a genuine consultation process. The EESC criticised the Commission’s proposal for treating public involvement largely as a way of increasing acceptance rather than giving affected communities a meaningful say. It suggested that Member States consider compensating communities that face lasting local impacts.

To improve environmental performance, the EESC proposed a European sustainability certification for cloud and AI services. This would allow public buyers to compare providers based on factors including their use of energy, water and renewable resources.

Social dialogue and consumer organisations are absent from the proposal

The proposal makes no reference to social dialogue, despite the effects of cloud and AI adoption on employment and the organisation of work. The EESC said workers, civil society and consumer organisations should be involved in shaping national cloud and AI strategies. It called for training programmes to be developed with employers and trade unions and for specific measures to attract and retain young people and young professionals.

Erosion of human capabilities

One of the opinion’s strongest warnings concerns the gradual loss of human expertise as organisations become more reliant on AI. The EESC said companies and public authorities risk losing the ability to perform the tasks themselves or independently check the work carried out by automated systems.

This decline may go unnoticed because productivity can appear stable even while human knowledge is being lost. National authorities should monitor this risk and assess whether essential public functions could still be performed, verified or restored if an AI or cloud service failed.

‘As we roll out AI, we need to think about which skills people must retain. Education will have a key role to play. As Europe builds its technological capabilities, we must not lose our human and institutional autonomy,’ Mr Hajnoš concluded.