CAPITAL AI WASH: THE MARIO GAMBIT

This paper introduces the concept of Capital AI Wash: the structural presentation of AI capability as nationally sovereign when the underlying ownership, financial dependency, and executive control remain anchored in a foreign jurisdiction. Using Google DeepMind as its primary case study, and the events of 17 June 2026 as its datestamp, the paper argues that the United Kingdom's claim to frontier AI standing is a sovereignty-by-proxy arrangement contingent on the operational goodwill of a US corporation subject to US executive authority. The paper examines Project Mario, the failed DeepMind independence bid of 2015 to 2017, as forensic evidence of the mechanism by which this dependency was consolidated. It connects the Mario failure to the concurrent erosion of DeepMind's founding ethics charter, the removal of Google's weapons and surveillance pledge in February 2025, and the Anthropic Mythos access restriction of June 2026 as successive demonstrations that the dependency is not theoretical. The paper argues that Capital AI Wash functions through genuine excellence: Hassabis's Nobel Prize, his deliberate London residency, and his knighthood all provide authentic symbolic cover for a governance structure that places the sovereignty claim's single point of failure in Mountain View, not Whitehall. This is named, following the collection's established vocabulary, as the Single Lane Paradox operating at the level of national infrastructure.