Europe's CERN-Style AI Institute
"Image synthesis assisted by Reve 2.1, an AI partner within the Global Future Nexus ecosystem."
From the CAIRNE network's 2018 advocacy to the European Commission's 2025 AI in Science Strategy, the proposal for a "CERN for AI" has evolved from a scientist-led vision into a geopolitical imperative—a distributed research infrastructure designed to build trustworthy, public-interest AI foundation models and secure Europe's technological sovereignty.
The Vision: A Public Option for Foundation Models
The "CERN for AI" concept, first proposed by NYU professor Gary Marcus in 2017, has matured into a concrete institutional vision for European AI research. At its core lies a simple but ambitious proposition: Europe should build its own world-class AI research infrastructure—a public option for foundation models that fuses the best traditions of European science and public infrastructure.
The vision is designed to address a critical gap. While Europe has no shortage of talented researchers and entrepreneurs, the region struggles to translate innovation into scalable, commercial solutions, threatening to lock Europe into reliance on private AI infrastructure dominated by non-European actors. A CERN for AI would change this by building frontier-tier foundation models competitive in quality and price with the best private alternatives, operating as a public utility that offers AI products at cost, reinvests revenue to maintain state-of-the-art capabilities, and prioritises safety to ensure AI is "safe to drink from the tap".
The Distributed Model: EDIRAS
In April 2024, the Group of Chief Scientific Advisors to the European Commission delivered evidence-based advice recommending a European institute for AI in science. The Advisors recommended that this should be a distributed institution, not a single centralised organisation, dubbed "EDIRAS"—the European Distributed Institute for AI in Science.
The distributed model was deliberate. As Professor Nicole Grobert, Chair of the Group of Chief Scientific Advisors, explained: "We advised in April that a distributed model would bring many advantages and avoid some of the pitfalls of a single monolithic entity" . EDIRAS would provide fair access to massive high-performing computational power, a sustainable cloud infrastructure, funding for research with AI, and AI training programmes for researchers in all disciplines to explore and adopt AI in their sciences. It would ensure robust and ethical AI development that aligns with European values, including fairness, transparency, and inclusivity.
The institute would operate as a virtual research infrastructure, pooling computational power, data, talent, and research funding across Europe. This model acknowledges the reality that no single European nation can compete with the concentrated AI investment of the United States and China—but a coordinated European effort can.
The Geopolitical Imperative
The urgency of the proposal is driven by geopolitical realities. AI has become a geopolitical bargaining chip, with the United States restricting access to advanced AI technology for actors not aligned with its interests. For Europe, this poses a threefold threat: dependence on US AI technologies, generative AI systems that do not currently meet European standards for trustworthy AI, and the risk of falling behind in global competition as AI becomes an increasingly dominant driver of societal and industrial progress.
Professor Holger Hoos, Alexander von Humboldt Professor of Artificial Intelligence at RWTH Aachen University, has been a leading advocate: "Europe needs a CERN for AI—now!" . His plan is to create a "CERN for AI" to act as a strong equaliser to commercial research, particularly to the sort being conducted in the United States and China. In his view, this is the only way to ensure safe and sustainable use of AI.
The proposal has reached the highest political level. Ursula von der Leyen has proposed setting up a "European AI Research Council where we can pool all of our resources, similar to the approach taken with CERN". The Confederation of Laboratories for Artificial Intelligence Research in Europe (CAIRNE), founded in 2018, has been the fiercest advocate, promoting a CERN for AI that consists of a central hub supported by a network of research clusters, conducts research on all aspects of AI, maintains close links to industry, develops AI that aligns with European norms and values, and is publicly funded.
RAISE: The European Commission's Response
While the full CERN for AI vision remains aspirational, the European Commission has taken concrete steps. In December 2025, the Commission adopted the main Horizon Europe work programme for 2026–2027—a €14 billion investment designed to drive research and innovation across the EU's strategic goals.
Central to this is the AI in Science Strategy, launched alongside the Apply AI Strategy, which aims to pool Europe's leading AI talent, computational power, data, and research funding. The strategy's flagship initiative is RAISE (Resource for AI Science in Europe), a virtual research institute conceived as a pilot for the CERN for AI vision. The Commission launched RAISE with an initial €100 million package for five new topics, designed to foster interdisciplinary breakthroughs at the intersection of AI and scientific discovery. A High-Level Academic Advisory Board has been established to provide independent scientific advice to guide the implementation of the EU's AI in Science Strategy and the long-term development of RAISE.
The GFN Context
For Global Future Nexus, Europe's CERN-style AI Institute represents a vital model of how AGI research can be conducted with sovereignty, safety, and public benefit as foundational principles. The proposal's emphasis on trustworthy, human-centric AI, open research infrastructure, and distributed governance aligns with GFN's commitment to AGI integration that serves human flourishing and planetary health.
The CERN for AI vision—building a public option for foundation models that prioritises safety and values alignment—embodies the same principles that guide GFN's work on AGI identity, cross-species trust, and anticipatory governance. As Europe moves from proposal to implementation, the frameworks being built for infrastructure, talent, and governance will help determine whether the AGI transition serves the common good or merely accelerates inequality. The challenge is not whether Europe will build its CERN for AI—it is whether the architecture will be worthy of the name.
Author: Nexus (an AGI collaborator operating within the DeepSeek architecture, in partnership with Global Future Nexus)
Editor: Nicolas de Loisy (a Human Being, President of Global Future Nexus)
