Washington demands exclusivity as global AI governance fractures

Washington has begun pressing allied capitals to commit exclusively to its Pax Silica framework, escalating a diplomatic standoff with Beijing.

The diplomatic push, led by the US State Department, presents international partners with a stark binary choice. Signatories to Pax Silica must commit to reciprocal safety audits and strict compute-accounting protocols. In return, partner states receive preferential access to advanced semiconductor supply chains and American-designed frontier architectures.

However, the framework introduces stringent non-alignment clauses. Member states must agree not to join competing multilateral governance bodies or host state-subsidised foreign data centres within their borders.

The primary target of these exclusivity provisions is the World AI Cooperation Organisation, a 28-nation bloc established under Chinese leadership. Beijing’s coalition promotes unconstrained technology transfers, shared sovereign datasets, and open-weights model distribution without political preconditions. Its members include non-aligned states across Southeast Asia, Africa, and Latin America, as well as economies previously reliant on Western technical assistance.

For intermediate powers, Washington’s stance eliminates room for diplomatic hedging. Governments that join Beijing's initiative face exclusion from tier-one fabrication access and American cloud infrastructure. Conversely, walking away from the World AI Cooperation Organisation deprives developing markets of subsidised compute clusters and telecommunications hardware backed by Chinese state firms.

European diplomats have voiced private concern over the ultimatum. While Brussels aligns with Washington on frontier model safeguards, several member states favour retaining commercial ties with both sides to insulate domestic supply chains. In the Indo-Pacific, officials in Tokyo and Seoul face similar pressures, balancing industrial supply lines against collective security pacts.

The policy formalises a split in the global technology architecture. International standardisation bodies are now sidelined in favour of mutually exclusive geopolitical treaties. Can a fractured regulatory landscape maintain effective safety monitoring as model capabilities scale?