European firms spread AI workloads as U.S. access limits bite

European companies are weighing AI providers on more than performance as U.S. access limits and sovereignty concerns make single-vendor dependence harder to defend. At VivaTech in Paris, Siemens, Renault Group, Orange and ChapsVision said they were combining U.S., Chinese and European models to preserve flexibility if access conditions change.

The pressure increased after the U.S. government ordered Anthropic to suspend access to its Fable 5 and Mythos 5 models for foreign nationals on national security grounds. Siemens said it uses DeepSeek, Alibaba’s Qwen and Nvidia’s Nemotron alongside other U.S. and European models, while Orange said its infrastructure can run all open-source models, including Chinese ones, when self-operated on European infrastructure.

Europe’s homegrown general-purpose AI market remains limited, with France’s Mistral leading the field and DeepL stronger in narrower use cases. Large companies are increasingly treating AI procurement as a resilience exercise focused on vendor concentration, model portability, cross-border access, continuity, data controls and cyber oversight.

The European Commission unveiled its European Technological Sovereignty Package on June 3, including the Chips Act 2.0, the Cloud and AI Development Act, the EU Open Source Strategy and a roadmap for digitalisation and AI in energy. The Commission says the bloc relies on non-EU countries for over 80% of key digital products, services, infrastructure and intellectual property.