Alliances in cyberspace used to be understood as the product of deliberate political choices. This paper challenges that assumption. The EU has developed the world's most ambitious regulatory AI framework through its AI Act, Data Act, NIS2, and Gaia-X, all anchored in the language of digital sovereignty and strategic autonomy. Yet companies like Anthropic, NVIDIA, Huawei, and Alibaba continue to expand their infrastructural presence across European digital ecosystems through cloud contracts, semiconductor supply chains, and AI platform integration. This paper argues that frontier AI is reshaping geopolitical power in Europe by embedding strategic dependence on American and Chinese Big Tech infrastructures within the core digital systems on which states increasingly rely.Building on Cyber Persistence Theory and the literature on weaponised interdependence, the paper examines how continuous integration into American and Chinese AI ecosystems quietly constrains sovereign choice within Europe. Through a comparative analysis of the Netherlands, Poland, and Hungary, it analyses whether EU regulatory initiatives meaningfully reduce infrastructural dependence and why European middle powers respond differently to the same competitive pressures. Methodologically, the paper will use a triangulation approach with national AI strategies, parliamentary debates, and EU policy documents to demonstrate that sovereignty in cyberspace is becoming infrastructural and corporate-mediated, and regulation alone cannot fully reverse this dynamic, leaving states progressively "aligned by default".

‍