MAIN CONTENT
An industrial race measured in computing power
Europe's dilemma begins with the expensive machinery behind modern AI. Advanced processors, data centres, cloud systems and specialised talent enable the development and use of powerful models. US companies dominate much of that infrastructure, while China has made AI a strategic industrial priority. European firms possess strong researchers and industrial customers, but many still rely on computing capacity and technologies developed elsewhere. The Dublin discussion therefore put technological sovereignty alongside competitiveness rather than treating it merely as a slogan.
Arthur Mensch, chief executive of France-based Mistral AI, was invited to discuss how the EU might reduce critical dependencies and widen the adoption of AI throughout manufacturing, services and public administration. Success would not require every company to design its own frontier model. It could mean reliable access to computing resources, opportunities for European suppliers, and tools suited to the languages and needs of the single market. A strategy centred exclusively on spectacular models could overlook the quieter productivity gains that come from practical use.

















