Europe’s starting point: Scaling technology, starting to fall behind Strategic posture: Pick your battles, leapfrog The AI market is set to be one of the largest among emerging technologies, projected to reach $3.1 trillion globally by 2040.73 Generative AI (GenAI) alone could inject over $575 billion into the European economy by 2030 through productivity gains.74 However, AI’s importance extends beyond its economic growth potential; the geopolitical landscape is evolving, and it is unclear whether adversaries or even allies might place restrictions on the technology, raising strategic implications for Europe’s AI ambitions and sovereignty. Europe’s starting point and potential unlocking actions Europe has strong potential to compete in AI (comprising applied AI, GenAI and machine learning operations), which could help address its current labour shortage and reinforce its economic resilience. Yet it could have a hard time staying relevant if its own companies don’t create greater demand for the burgeoning technology. So far, European organizations have been notably slow to adopt AI in their own operations; they trail their US counterparts in that critical metric by 45% to 70%.75 Bridging Europe’s AI gap with both the US and China, whether for adoption, funding or other areas, is not an easy task given Europe’s data sharing and privacy safeguards, high energy costs and fragmentation. Some possible solutions include: –Building scale. Europe currently lacks the local computing and cloud hosting capacity to develop and scale AI across the continent. With data centre demand projected to grow 22% per year by 2030,76 European policy-makers may want to consider facilitating increased investments in data centres and cloud-hosting facilities, through targeted incentives and access to reliable and affordable (green) energy. Creating an EU-wide framework for providing “computing capital” to innovative small and medium-sized enterprises in the EU could be explored, as could opening the Euro High Performance Computing Joint Undertaking (Euro HPC JU) to a federated AI model favouring public-private cooperation to develop the relevant infrastructure. –Simplifying the regulatory and permitting environment. The EU AI Act has taken a step towards modernizing regulations specific to AI, yet 70% of European companies report they find the obligations too complex.77 Policy- makers may want to consider harmonizing national AI sandbox frameworks across all member states to facilitate the development of innovative AI applications in selected industrial sectors, while also ensuring streamlined and consistent implementation of AI regulations and the General Data Protection Regulation (GDPR). –Increasing innovation capital and investment. Europe is significantly underinvested in AI. The US funnels six times as much private capital into the technology, and European public sector funding also significantly lags both the US and China.78 To begin to bridge the gap, Europe could explore loosening pension private equity/venture capital (PE/ VC) allocation rules and setting an aspiration for member states to invest 0.1% of European GDP in GenAI infrastructure, such as data centres along several AI verticals.79 This could be supported through the allocation of public budgets for procurement of AI applications for sectors such as healthcare, defence and automotives, with potential guardrails to allocate a defined share to European innovators. –Driving commercialization. Despite some significant advancements in AI, particularly in healthcare and banking, Europe still trails the US in commercialization due to lower penetration and corporate adoption. The integrated use of open-source AI models with proprietary ones can be a way to trickle down innovation and support commercialization. Gaps in growth funding, and less sector collaboration, especially between new players and incumbents, only add to the challenges. Simplifying business regulations with measures such as a “28th tech regime” could be beneficial in addressing disparities in European companies’ scale and resources. –Strengthening research and talent. Though Europe has slightly more AI professionals than the US and produces 22% of the world’s leading AI researchers, only 14% stay in the region.80 This is driven by a large compensation gap of two to four times between the US and Europe.81 Incentives such as premiums or tax breaks for talent win-back initiatives, and support for research institutions could enhance Europe’s appeal to top-tier talent. –Cultivating ecosystems and global leaders. The scale of investment required to be globally competitive in AI is too large for any single European company or even country to manage. Building an AI start-up ecosystem in Europe could enable future development of globally relevant, scalable solutions for priority AI application verticals, such as mobility, manufacturing and defence. 4.3 Artificial intelligence (AI) Europe in the Intelligent Age: From Ideas to Action 25

Europe in the Intelligent Age 2025