European companies resist full sovereign AI shift

Anthropic’s disabling of Fable 5 and Mythos 5 for users worldwide has intensified European concerns about reliance on foreign AI providers. The move, linked to US export controls and ongoing negotiations with the US government, came shortly after the EU announced a Technological Sovereignty Package aimed at strengthening local tech suppliers and steering some public procurement toward European options.

European businesses are resisting a hard shift toward sovereign infrastructure. Volvo and Stellantis have warned that sovereignty could raise costs, while ASML, Capgemini and Ericsson have argued that protectionism would reduce investment in Europe. The EU says the bloc depends on non-EU providers for over 80% of key digital products, services, infrastructure, and intellectual property, and recent partnerships with Google Cloud by Thales and Telefónica show continued reliance on US hyperscalers.

Analysts and executives described sovereignty as important but secondary to adoption speed, performance and value-for-money for most firms. Veeam found that 83% of CEOs worldwide report pressure to accelerate rollouts of AI and data capabilities, while only 30% have visibility over the provenance and makeup of their organisations’ data resources, creating governance and compliance risks.

A more pragmatic model is emerging, with companies using a mix of European, US and Chinese AI models to reduce exposure to potential kill switches without giving up capabilities. Siemens, Renault, Orange and ChapsVision were cited as examples, while the EU’s Cloud & AI Development Act is described as focused on public-sector purchasing rather than a broad buy-European mandate.