India’s home‑grown AI firm BharatGen has been named the anchor for the country’s participation in Project Tapestry, the AI Alliance’s new open‑source consortium aimed at creating frontier‑level models while preserving national control over data and deployments. The announcement comes amid heightened focus on AI sovereignty at the recent G7 summit in Évian and the VivaTech expo in Paris.
“AI is becoming essential infrastructure. No single company or country should determine who can build on it, adapt it, or benefit from it. Project Tapestry is a test of whether frontier AI can be built as open infrastructure: collaborative, capable, and sovereign by design.”
— Dr. Yann LeCun, Chief Science Advisor, AI Alliance and Executive Chairman, AMI Labs
Project Tapestry’s premise is to pool data, talent, compute power and funding across a broad coalition of nations, academic institutions, and industry players. By doing so, it seeks to assemble resources that exceed the capacity of any single entity while allowing each participant to retain ownership of its contributions and the resulting derivative models.
India’s commitment is framed as a bridge between the country’s “self‑reliant” AI strategy and the consortium’s collaborative model. BharatGen will co‑lead workstreams focused on distributed model training, providing a concrete pathway for Indian data and expertise to flow into the shared base model.
“India has long believed that science advances fastest when it is open and shared. Project Tapestry, which lets nations build frontier AI together while retaining sovereignty over their data and models, reflects the collaborative, self‑reliant innovation India is advancing.”
— Professor Ajay Sood, Principal Scientific Adviser to the Government of India
“As India builds frontier AI rooted in its own languages and knowledge, IIT Bombay and BharatGen supported by the India AI mission are proud to support and participate as a founding contributor in Project Tapestry, an open, global consortium for nations to advance frontier AI together.”
— Shireesh Kedare, Director, IIT Bombay and Chairman, BharatGen
French President Emmanuel Macron, who is steering the G7 AI agenda this year, has underscored the strategic importance of building AI capability across borders. Project Tapestry aligns with that vision by offering a governance structure that lets each member keep control over its data, bespoke models, and deployment pipelines.
The consortium was officially launched a month ago during a technical workshop in Paris that brought together 30 AI leaders from the United States, France, India, Vietnam, Japan, Switzerland, the UAE and other regions. French participants such as Pleias, PRAIRE, AMI Labs, Software Heritage and Current AI played a visible role, reflecting France’s push for an open yet sovereign AI ecosystem.
Vietnam’s Cultural Alignment Initiative
Vietnam has emerged as a concrete example of how the consortium can address cultural specificity. By involving the National Innovation Center and FPT Corporation, Vietnam aims to develop models that respect local language, customs and industry practices while keeping data and model ownership within national borders.
“With the National Innovation Center and FPT Corporation engaged, Project Tapestry gives Vietnam a path to help build frontier AI that reflects its language, culture, industries, and institutions while retaining control over its own data, models, and deployment.”
— Dr. Christopher Nguyen, Chief Architect of Project Tapestry and CEO, Aitomatic
Early work with Vietnamese partners has already demonstrated the feasibility of aligning model outputs to cultural nuances—a step beyond simple language translation.
Architecture and Governance Roadmap
The Paris workshop produced a reference architecture that centers on a centrally developed base model. Member nodes receive this model, fine‑tune it with locally sourced data, and can then feed improvements back into the shared repository. Crucially, the design ensures that contributors retain ownership of any enhancements derived from their proprietary datasets.
The consortium’s guiding principle is that sovereignty cannot rely on trust alone; it must be baked into the system’s technical foundations. Participants will therefore retain full rights over their data and any unique capabilities they create, even as they contribute to a collective advancement of the base model.
“Project Tapestry aims to build the most capable frontier AI possible in a way that ensures the value and control of that AI is retained by those developing and deploying it—whether they are in industry, government, a non‑profit or an individual. The knowledge and intelligence of humanity that is used to build AI does not live in any one organization. The most capable and trusted AI can’t either.”
— Dr. Anthony Annunziata, Chairman, AI Alliance and Director of AI Open Innovation, IBM
Data Diversity as a Competitive Edge
The consortium argues that the next leap in AI performance will depend on data diversity, domain‑rich and often absent from the public web—such as national scientific repositories, industry‑specific datasets, and proprietary research archives. By aggregating these sources without transferring ownership to a single central entity, Project Tapestry seeks to sidestep the concentration risks associated with traditional AI giants.
The initiative represents a tangible attempt to reconcile the demand for ever‑more capable AI with the geopolitical imperative of maintaining national control over critical technology.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
