What happened
Carnegie Endowment published a policy paper (28 September 2026) arguing that middle powers such as Germany, India, and the UK cannot build frontier AI alone yet cannot safely depend on either the US or China for access, which it says 'may soon be rationed, conditioned, and switched on or off at the discretion of a handful of American firms and the government that oversees them.' It advances a novel strategic framework — an 'AI breakout posture': constructing and maintaining the capacity to build frontier AI quickly should the need arise, while remaining closely aligned with a great-power patron, as a hedge rather than a sovereign moonshot. The paper examines why both sovereign moonshots (unaffordable, and even an alliance of middle powers 'lags behind on every relevant input') and pure adoption strategies fail, and details the policies needed to implement breakout capacity. It closes by warning US policymakers to understand that even their closest allies may pursue this option if they conclude they cannot trust Washington as a reliable supplier of frontier AI.
Why it matters
For executives and policy leads in non-frontier economies (and multinationals dependent on US/China frontier-model access), this offers the clearest emerging framework for hedging against great-power control of the AI supply chain — directly relevant to market-access, sovereignty, and contingency decisions through 2028.
Action needed
Map your organization's or jurisdiction's dependence on US/Chinese frontier AI access and evaluate the breakout-posture framework as the basis for national AI strategy and AI supply-chain contingency planning.