Europe and Canada establish an AI computing power cooperation alliance, sharing 14 supercomputers, 10 quantum computers, and 19 AI factories

Under the cooperation arrangement, Canada will become an active participant in Europe's public-private high-performance computing network. Canadian companies will be able to gain priority access to European supercomputing resources, and its startups and research institutions will also have opportunities to participate in related projects. For the EU, this cooperation helps bring in Canada's competitive technology companies and technical resources, expanding the scale of Europe's artificial intelligence ecosystem.

EuroHPC currently has a considerable computing infrastructure network, including 14 world-class supercomputers, 10 quantum computers and 19 AI factories. The entire network covers 38 participating countries, supports 67 research projects, and has 36 national competence centers.

The program's current total budget is approximately 8.2 billion euros. In addition, Europe plans to invest an additional approximately 30 billion euros to build up to 7 new AI data centers, with construction of the related facilities expected to begin in 2027. These data centers will further strengthen Europe's computing capacity in areas such as AI training, scientific computing and complex engineering simulation.

This cooperation is not simply about opening up computing resources. As the scale of AI models continues to expand, the computing power, energy and data center investment needed to train and run advanced models have become an important component of national technological competitiveness. For Canada, joining EuroHPC means its research institutions and companies can access a broader range of European computing resources without having to rely entirely on its own limited supercomputing facilities.

Europe, for its part, hopes to leverage Canada's technological industry advantages to expand the scope of cooperation in its own AI industry. The two sides can share resources in high-performance computing, quantum computing, AI research and related industrial applications, forming a transatlantic technology cooperation network.

This cooperation also has a clear geopolitical backdrop. In recent years, relations between Canada and the United States have gradually become tense, with growing disagreements between the two sides on issues such as trade, industrial policy and national strategy. Europe clearly hopes to use this opportunity to strengthen its relationship with Canada and develop it into a closer technology partner.

Canadian Prime Minister Mark Carney and von der Leyen have previously jointly launched a cooperation initiative called the "Alliance for the Future," providing a new channel for cooperation among like-minded countries. Von der Leyen has also suggested considering making Canada the EU's first quasi-member state. Sharing supercomputing resources is only part of this broader cooperation framework.

However, whether this alliance can truly change the balance of power in the global high-performance computing field remains highly questionable.

Anton Leicht, an AI expert at the Carnegie Endowment for International Peace, believes that strengthening cooperation between "middle powers" such as Canada and the EU is one of the most direct short-term measures for building shared AI infrastructure. But he also points out that without large-scale new computing facilities, such cooperation will be hard-pressed to have a substantive impact on the United States' dominance in the supercomputing field.

The United States is currently investing enormous sums in building a new generation of high-performance computing clusters and AI data centers. Leicht specifically emphasizes that as long as the United States continues to build large-scale computing clusters based on gigawatt-level energy scale, other countries will find it difficult to change the competitive landscape through cooperation alone, even if they strengthen resource sharing.

This in fact points to the core problem facing Europe and Canada: sharing existing computing power can certainly improve resource utilization efficiency, but AI competition ultimately still depends on the absolute scale of computing facilities, the ability to supply advanced chips, energy security and sustained funding.

Therefore, this cooperation between Europe and Canada is more like establishing a more autonomous transnational cooperation network outside the US-dominated AI infrastructure system, rather than immediately forming an independent technology bloc capable of rivaling the United States. Whether the two sides can translate existing cooperation into truly competitive computing capacity will also depend on whether the planned tens of billions of euros in investment can be implemented as scheduled.