• China has proposed a BRICS open-source AI community for cooperation on large language models, applications, training and technical exchange
  • Beijing also wants a digital ecosystem cloud platform, but no hosting architecture, compute capacity, funding or timetable has been disclosed

The fact

Chinese President Xi Jinping has proposed creating a BRICS open-source artificial intelligence community as part of a wider package of technology cooperation among member countries. The proposal includes collaboration on the development and application of large language models, alongside AI seminars, training programmes and technical exchanges.

Xi separately proposed a BRICS digital ecosystem cloud platform intended to support cooperation in digital technologies. China has not disclosed where the platform would be hosted, which organisations would operate it, whether it would provide accelerator capacity or whether it would host models developed through the proposed AI community.

The initiatives remain Chinese proposals rather than operating BRICS infrastructure. No funding, governance structure, participating institutions, model repository, hardware specification or implementation timetable has been announced. The wider BRICS New Delhi Declaration supports cooperation on AI access, innovation, safety and governance, but does not itself establish either platform.

The assessment

The proposals put two different layers of digital cooperation on the table. An open-source AI community could give participating institutions a place to share models, tools and technical work, while a cloud platform could provide infrastructure for broader digital cooperation. Beijing has not yet said whether the two would be technically connected.

Open models can give organisations more control over where models run, how they are adapted and what data stays inside their own infrastructure. That does not remove the need for compute, software integration, evaluation and operations. Nor does choosing one model family make later migration impossible, although applications and testing processes built around it can increase switching work.

For BTW readers, the important next step is whether the proposal acquires infrastructure behind the political commitment. Named model repositories, hosting providers, accelerator capacity and participating institutions would show that BRICS AI cooperation is moving from summit language towards an operating platform.

What to watch

Watch for a governing body, funding commitments and named institutions responsible for the AI community or cloud platform. Model licences, hosting locations, accelerator specifications and data-residency rules would define how the system actually works. Adoption by a BRICS government, university or public-sector AI programme would provide the clearest evidence that the initiative has moved beyond policy statements.