Last week, a US export-control order forced Anthropic to switch off its two most capable AI models for users worldwide (Fable and Mythos). It showed, in a single weekend, that access to frontier AI can be withdrawn by a government that is not your own.
Our new discussion paper, The Sovereignty Paradox, argues that real AI sovereignty was never about owning that model. It is about strategic optionality: deciding, layer by layer across the AI value chain, what to own, what to diversify, and where to trust an external partner. The paper maps the global value chain, the power and compute bottleneck now constraining AI, and a layer-by-layer posture for Middle East governments and local industry.
