Rethinking Europe’s AI Strategy
As the global race to dominate artificial intelligence intensifies, Europe may be headed in the wrong direction, according to Christian Klein, CEO of SAP SE, one of the region’s largest software firms. Speaking to Bloomberg, Klein questioned the wisdom of investing billions into AI supercomputing infrastructure, such as data centers and chips, when Europe lacks a solid foundation in applied AI use cases.
“Do we really need to build five data centers filled with high-powered chips? Is that what Europe needs? I doubt it,” Klein said.
He emphasized that large language models (LLMs) — like GPT — are becoming increasingly commoditized, and the real opportunity lies in leveraging AI to solve business problems, not just in building expensive hardware.
The U.S. Is Winning the Infrastructure Race
While Klein urges focus on practical AI applications, the United States is doubling down on infrastructure. In January, several U.S. tech giants launched Project Stargate, with plans to invest $500 billion into AI hardware, including data centers. One major agreement included OpenAI leasing 4.5 gigawatts of power from Oracle data centers—equivalent to four nuclear power plants.
In response, the European Union launched InvestAI, a €200 billion initiative aimed at boosting the region’s AI ecosystem. The plan includes a €20 billion fund for “AI gigafactories”, which would presumably compete with U.S.-based data hubs.
However, Klein believes chasing the U.S. playbook is a strategic misstep. He argues Europe should concentrate on integrating AI into supply chains, healthcare systems, manufacturing, and other real-world domains where AI has clear ROI.
EU AI Regulation Sparks Industry Backlash
Further complicating Europe’s AI ambitions is the region’s aggressive regulatory framework. The EU Artificial Intelligence Act officially came into force on August 1, 2024, with full implementation set for August 2, 2026.
The law introduces sweeping restrictions, including:
- February 2025: Ban on AI systems with “unacceptable risk”, such as social scoring or manipulation tools
- August 2025: Requirements for general-purpose AI models
- August 2026: Rules for “high-risk” AI applications, including those tied to biometrics, critical infrastructure, and education
In protest, over 45 leading European firms — including Airbus, ASML, Mistral AI, and Mercedes-Benz — signed an open letter asking for a two-year delay in enforcement. They argue the lack of implementation guidelines could stifle innovation and discourage investment in the region.
Interestingly, SAP and Spotify are not among the signatories, signaling potential divergence in industry opinion.
Conclusion: A Fork in the Road for European AI
Europe stands at a critical crossroads. While the U.S. accelerates AI development through massive infrastructure and looser regulation, Europe is grappling with how to balance innovation with control. Christian Klein’s remarks shine a light on a different path—one focused on applied AI that delivers tangible value to businesses and society.
Whether Europe chooses to double down on hardware or pivot toward practical integration will shape its global competitiveness in the AI era.
