The AI Alliance has announced that India’s BharatGen will anchor its initial participation in Project Tapestry, a global open consortium focused on building frontier AI through distributed model development. The announcement comes amid rising global discussions on AI sovereignty, highlighted at recent events such as the G7 Summit in Évian and VivaTech in Paris.
Project Tapestry aims to enable collaborative AI development across nations and institutions while ensuring participants retain control over their own data, models, and deployment.
Project Tapestry is designed to redefine how frontier AI systems are built by distributing model development across countries, research institutions, and industry partners. The initiative allows participants to collaborate on large-scale AI systems while maintaining sovereignty over their datasets, models, and deployment environments.
The approach is built on the principle that AI infrastructure should function as shared global infrastructure without centralizing control in any single organization or nation.
India’s BharatGen, a full-stack sovereign AI initiative, will anchor India’s role in Project Tapestry and co-lead workstreams focused on distributed model training. This positions India as a key contributor to the development of globally collaborative yet locally governed AI systems.
Indian leadership emphasized that the initiative aligns with the country’s broader focus on self-reliant, open, and inclusive AI development rooted in national languages and knowledge systems.
Project Tapestry introduces a collaborative AI development framework in which a central base model is trained and then distributed to participating nodes. These nodes refine and extend the model using local datasets while retaining ownership of their improvements and derivatives.
Participants contribute enhancements back to the shared system, enabling continuous improvement without compromising data sovereignty or institutional control.
The initiative already includes participation from organizations and institutions across multiple regions, including the United States, France, India, Vietnam, Japan, Switzerland, and the UAE.
Vietnam has been highlighted as a key early use case, focusing on culturally aligned AI systems that reflect national language, customs, and industrial context while maintaining full local ownership of data and deployment.
A core principle of Project Tapestry is that AI sovereignty must be embedded in system architecture rather than relying solely on trust agreements. This ensures that each participating country or organization retains control over how AI models are trained, adapted, and deployed.
The model emphasizes distributed compute, federated learning principles, and shared innovation without centralized data consolidation.
Project Tapestry operates under the AI Alliance’s nonprofit structure, bringing together researchers, governments, enterprises, and open source communities. The goal is to pool global data, talent, compute, and funding while avoiding dependency on any single commercial provider.
This structure is designed to expand access to frontier AI capabilities while preserving autonomy for all contributors.
Project Tapestry represents a major step toward rethinking how frontier AI systems are built, governed, and deployed. By combining global collaboration with strict data and model sovereignty, the initiative aims to create a new framework for AI development that balances innovation with national and institutional control.
India’s BharatGen anchoring the project highlights the growing importance of sovereign AI strategies in shaping the next phase of global artificial intelligence infrastructure.
The AI Alliance is a global nonprofit organization focused on advancing open, safe, and responsible AI through collaboration across industry, academia, startups, and government. It operates across multiple countries and supports initiatives spanning AI models, data, safety, and governance.
