Europe Launches RAISE Initiative to Give Researchers Better Access to Artificial Intelligence

Europe wants to make artificial intelligence a more powerful scientific tool

The European Commission is taking another step towards integrating artificial intelligence into scientific research through the RAISE initiative, short for Resource for AI Science in Europe.

The initiative is designed to help researchers access resources that are increasingly important for AI-enabled discovery, including computing capacity, advanced models, data, software and specialist expertise.

A related pledging initiative was launched at the European AI in Science Summit in Dublin on 15 October 2026. It aims to bring together industry and philanthropic organisations willing to contribute resources and capabilities to support researchers across Europe.

The programme addresses a practical problem in modern science. Researchers may have promising ideas and valuable datasets but lack the computing resources, technical staff or specialist AI tools needed to investigate them.

By connecting research teams with shared resources and partnerships, Europe hopes to accelerate scientific work while strengthening its own position in AI research.

What is RAISE?

RAISE is intended to function as a virtual research institute supporting excellent European research on and with AI.

According to the European Commission, its approach involves pooling computational resources, data, talent and research funding to support researchers and encourage collaboration between disciplines.

The idea is not simply to give scientists access to a chatbot. Scientific AI can involve complex workflows that combine large datasets, simulation software, specialised models and high-performance computing.

A research team investigating a new material, for example, might use AI to identify promising molecular structures, computational simulations to evaluate their properties and laboratory experiments to test the most promising candidates.

Such work can require substantial resources and expertise. Shared infrastructure may help researchers who cannot afford to build these capabilities independently.

The initiative also reflects a broader shift in scientific practice: AI is increasingly being considered part of the research infrastructure itself, rather than merely an administrative productivity tool.

How AI could accelerate scientific discovery

1. Drug discovery and medical research

Developing new medicines is a long and expensive process. Researchers must identify promising biological targets, investigate potential molecules, assess safety and efficacy, and conduct extensive testing.

AI can help analyse biological data, identify patterns and prioritise candidates for further investigation.

In a well-designed workflow, a model may narrow thousands of possibilities to a smaller set that researchers can examine experimentally. This does not mean that AI can independently certify a treatment as safe or effective. Laboratory work, clinical trials and regulatory review remain essential.

Nevertheless, better computational screening could help researchers allocate time and resources more efficiently.

2. Materials science

Researchers working on batteries, semiconductors, catalysts and advanced industrial materials often need to evaluate a large number of possible chemical structures and physical properties.

AI models can help suggest promising candidates or approximate certain calculations, while physics-based simulations and experiments test whether the predictions hold up.

If these methods reduce the time required to identify useful materials, they could have implications for energy storage, electronics, transport and manufacturing.

3. Climate and environmental research

AI can process large volumes of satellite imagery, sensor readings and climate-related data. Researchers may use it to improve forecasts, identify environmental changes and model complex systems.

Potential applications include monitoring deforestation, tracking pollution, analysing extreme weather patterns and improving estimates of energy demand.

The quality of the results depends on the underlying data, the scientific assumptions and the ability to validate predictions against observations.

4. Physics and astronomy

Modern physics experiments and astronomical surveys can generate enormous quantities of data.

AI may help identify unusual signals, classify observations and prioritise data for further investigation. In some cases, machine-learning techniques can make parts of a computational workflow more efficient.

Scientists still need to establish whether a detected pattern represents a meaningful phenomenon or a statistical artefact.

Across all these fields, the main opportunity is to help researchers investigate more possibilities and focus human attention on the questions that matter most.

Why access to computing power matters

Scientific AI can require significant computing resources, particularly when researchers work with large models, extensive datasets or complex simulations.

High-performance computing infrastructure is expensive to establish and operate. It requires processors, storage, networking, cooling, electricity and technical expertise.

If access to these resources is concentrated among a small number of well-funded institutions, research opportunities may become unevenly distributed.

Pooling resources could help universities, laboratories and research teams access capabilities that would otherwise be beyond their budgets.

RAISE’s initial pilot is funded through Horizon Europe, with €33 million allocated in the 2025 work programme and €107 million in the 2026–2027 work programme, according to the Commission’s published information. These figures refer to dedicated funding for the pilot and should not be interpreted as the total value of all possible contributions to the broader initiative.

The additional pledging initiative is intended to encourage private-sector and philanthropic support alongside public funding.

The important test will be whether these commitments translate into accessible resources, useful research partnerships and measurable scientific progress.

Can Europe use AI research to strengthen its technological independence?

Europe’s AI strategy is also connected to the question of technological sovereignty: how much control the region has over the infrastructure and technologies on which its economy depends.

Researchers may rely on computing platforms, AI models and cloud services supplied by organisations outside Europe. These services can be valuable, but dependence on a narrow set of providers may create risks related to cost, availability, data governance and strategic priorities.

RAISE could contribute to a more resilient research ecosystem if it helps European institutions build expertise, collaborate across borders and gain access to suitable computing resources.

However, technological independence does not necessarily require every component to be developed domestically. Scientific collaboration is international, and many research projects benefit from using the best available tools regardless of where they originate.

The more practical objective is to ensure that European researchers have meaningful choices and the capacity to pursue important work without being constrained by a lack of access.

The risks: AI-generated results are not automatically scientific discoveries

AI can produce persuasive answers and predictions that are nevertheless incorrect.

In scientific research, this problem is particularly serious because an erroneous result may influence subsequent experiments, publications or investment decisions.

Research teams therefore need transparent methods, reproducible workflows and independent validation. Where AI systems generate hypotheses or analyse evidence, researchers must understand the limitations of the tools and document how conclusions were reached.

Other issues include data protection, intellectual property, bias in training data and the risk that widely available AI tools could be used in harmful ways.

The European Commission’s stated principles for its AI-in-science initiative include scientific excellence, responsible and trustworthy AI, collaboration and compliance with applicable legal and ethical standards.

These principles will matter in practice. The value of AI in science should be measured not simply by the volume of model-generated output, but by whether it helps produce reliable, reproducible knowledge.

What researchers and universities should watch

For universities, laboratories and research-intensive companies, the next steps are likely to involve understanding which resources will become available, how access will be allocated and what application requirements will apply.

Researchers should look for official calls, eligibility criteria and opportunities to participate in cross-disciplinary projects.

Institutions may also benefit from investing in staff who can bridge AI engineering and domain-specific science. A powerful model is most useful when researchers understand both the technology and the scientific problem it is being applied to.

Industry partnerships may offer access to tools and expertise, but agreements should clearly address publication rights, data access, intellectual property and research independence.

The bottom line: Europe’s RAISE initiative seeks to make AI-enabled scientific research more accessible and collaborative. Its long-term success will depend on practical access to computing resources, high-quality research partnerships and rigorous scientific validation—not merely on announcing new programmes.