Anthropic shutdown sparks sovereignty debate across Europe

Editor's take

Anthropic was compelled to disable two of its large language models, Fable 5 and Mythos 5, globally due to a U.S. government order. This action highlights the inherent tension between U.S. national security interests and the desire for European technological independence.

The shutdown directly impacts European researchers and businesses reliant on Anthropic's models, potentially hindering their AI development and innovation. It intensifies the ongoing debate about whether Europe should prioritize building indigenous foundation models, like Mistral AI's Mixtral 8x7B, or focus on ensuring continued access to existing models, even those developed outside the continent.

Future developments will focus on the EU's concrete policy responses and investment strategies for fostering domestic AI capabilities. The extent to which European entities can secure reliable, long-term access to cutting-edge LLMs from U.S. providers, or successfully scale their own, will be critical indicators of their AI sovereignty.