Mid-2026 reveals a bifurcating global AI landscape, with two competing governance ecosystems: Pax Silica, led by the US, ties frontier AI access to trusted, China-free supply chains; and WAICO, led by China, promotes open-weight models and open access for the Global South.

India should pursue a layered strategy: selectively leverage frontier models when substitutes fall short, employ open or self-hosted models for non-frontier tasks, and build a domestic software envelope (model wrapper) to ensure safe, interoperable access.

India’s stance is shaped by its Pax Silica membership and WAICO non-membership, creating urgency to develop sovereign capabilities across the full AI stack and reduce dependence on foreign intermediaries.

Incidents like Mythos and Fable 5 show frontier AI access can be restricted or revoked—initially via export-control-like measures that affected all users and partners, followed by partial restoration, underscoring AI as a strategic asset.

China’s open-weight model pathway spreads governance norms with technology, potentially exporting Beijing’s cyber sovereignty preferences, while US controls stress security, supply-chain integrity, and strategic deterrence.

France’s Albert API demonstrates a national software envelope concept to safely unify access to multiple models, providing a practical blueprint for sovereign and secure AI deployment.

WAICO’s rise centers on open-weight models, thousands of training slots for developing countries over five years, and a governance model that avoids ceding control to a single power, contrasting Pax Silica’s security-first framework.

Indigenisation remains slow and costly but is essential to avoid lock-in and enable authentic AI leadership, urging India to run parallel tracks for access management and domestic capability.