The Curious Case of the Vanishing Cost (and Tax) Base – Is AI Quietly Engineering BEPS?

Revision of chapter VII of the OECD Guidelines evokes a fundamental question on the sustainability of safe-harbour regimes and risk-insulated cost-plus entities. Transfer pricing policies framed for a labour-intensive economy in which the cost of intra-group services was, overwhelmingly, the cost of people is threatened by an impending tectonic shift. Will the integration of AI rewrite some of the foundational assumptions? This article examines what AI adoption does to the functions, assets and risks profile of cost-plus entities and asks whether the fixed markups, eligibility criteria and benchmarking foundations of safe-harbour regimes remain defensible as execution and even risk migrates from human intelligence to artificial intelligence. Anchoring on the OECD Guidelines and the safe-harbour regimes of India and Singapore, and Pillar One’s Amount B simplified and streamlined approach, the author argues that AI will not destroy safe harbours but will act as a catalyst for their recalibration and that uneven adoption across jurisdictions could create a new risk of double taxation.