As the BRICS summit gets underway at Bharat Mandapam, the temptation will be to reach for the bloc's familiar script: More trade in local currencies, another call to reform the UN Security Council, a fresh round of promises on climate finance. India, holding the chair for 2026, could recite those lines comfortably. It has done so before, in 2012, 2016 and 2021.
But the ground under BRICS has moved. Ten years ago, the grouping's centre of gravity was investment flows and growth rates. Today, the currency of power is measured in compute clusters, chip fabs, and who gets to write the rules for the algorithms running through them. Washington and Beijing worked this out first. Brussels is fast building a regulatory wall around AI. The question this summit actually needs to answer is not how BRICS talks about growth, but whether it can agree on anything real about the technology stack -- compute, AI infrastructure, semiconductors, data localisation, governance -- that increasingly decides who grows and how.
That is the dialogue India has been building toward all year, ever since it hosted the AI Impact Summit in February. BRICS is the next test of whether India can convert domestic credibility into something the rest of the Global South will actually follow.
India's pitch to the developing world has always rested on a simple, defensible claim: It built Aadhaar, Jan Dhan and UPI as public infrastructure, largely without ceding control to private platforms, and then let more than 50 countries study or adopt pieces of that stack. That history buys India a hearing in Jakarta, Lagos and São Paulo that neither Washington's venture-funded AI giants nor Beijing's State-backed champions can easily match.
{{/usCountry}}India's pitch to the developing world has always rested on a simple, defensible claim: It built Aadhaar, Jan Dhan and UPI as public infrastructure, largely without ceding control to private platforms, and then let more than 50 countries study or adopt pieces of that stack. That history buys India a hearing in Jakarta, Lagos and São Paulo that neither Washington's venture-funded AI giants nor Beijing's State-backed champions can easily match.
{{/usCountry}}The trouble is that public digital infrastructure was last decade's proof point. This decade's test is AI, and BRICS members are already regulating it in strikingly different directions. Brazil has moved toward a rules-heavy AI law that echoes Europe's. China pairs frontier-model development with licensing requirements and active content moderation. Russia has built AI policy around a security-first framework. The Gulf members of the bloc -- Saudi Arabia and the UAE -- are less interested in writing rules than in buying compute at scale, positioning themselves as the capital and hardware bridge between Western chipmakers and everyone else.
India's own answer, the AI Governance Guidelines that MeitY released in November, sits closer to Washington's instinct than Beijing's or Brasília's: Light-touch, techno-legal, built on existing law rather than a standalone AI statute. Guidelines believe that embedding safeguards into the technology itself will do more work than a new regulator.
At Rio de Janeiro last year, BRICS leaders signed a statement declaring that access to AI must be fair, equitable and inclusive, regardless of a country's stage of development. It was the right sentiment and, like most leaders' statements, thin on mechanics. New Delhi's job is to give that sentence a spine.
Start with the most basic constraint every non-frontier economy faces: compute is expensive, and renting it from a handful of western clouds or Chinese providers is not sovereignty, it is dependency with better branding. BRICS already has an institution built to solve exactly this kind of collective-action problem -- the New Development Bank, founded in 2015 precisely to fund shared infrastructure across the bloc. A dedicated compute-financing window through the NDB might be an option. It’ll let members pool capital for shared training and inference capacity rather than each country negotiating alone with Nvidia or Huawei, turning the Rio declaration from rhetoric into a budget line.
Language is the second lever, and it happens to be one where India has already done the hard work at home. BRICS nations collectively speak a striking share of the world's languages, most of them poorly served by the large models trained overwhelmingly on English and Mandarin data. India's own Bhashini programme has built open datasets and benchmarks for 22+ Indian languages that commercial labs had no incentive to touch. Offering that playbook like shared low-resource-language datasets, with common benchmarking standards as a BRICS-wide public good AI will be a concrete, exportable win.
Third, BRICS governments are collectively the largest employers of citizens' data on earth, simply by virtue of population. A minimum-standard approach to AI in the public sector i.e shared baselines for procurement and audit of government AI systems, without forcing any member to harmonise its domestic AI law. This will give smaller members a floor to build on while leaving sovereignty intact. India's own strength here is not technical alone; a functioning democracy with real regulatory institutions has a factor of trust that a one-party State or a monarchy does not.
There is also a harder, more political reason for India to move now. In July, China convened 29 countries in Shanghai to launch the World AI Cooperation Organization: WAICO, an explicitly Global South-facing body that counts Russia, Brazil, Indonesia, South Africa and Pakistan among its founding members. India did not sign on, and neither did the US or the EU. Beijing's pitch was nearly identical to what Rio promised -- fair, equitable, people-centred AI governance -- except administered from Shanghai, on China's terms.
That is the live alternative India is competing against, whether or not anyone at Bharat Mandapam says so out loud. If the Global South's AI governance conversation ends up run entirely out of one capital, "fair and equitable" will mean whatever that capital decides it means. India's argument for leadership was never going to be that its own initiatives should replace China's. It is that no single country's platform should be allowed to become the only one and that a Global South agenda credible enough to attract Gulf money, Latin American markets and African infrastructure needs a convenor whose interests visibly stop short of controlling the outcome.
None of this happens in a vacuum. India has its own bilateral technology arrangement with Washington, part of what analysts have taken to calling Pax Silica: The American-led effort to keep the semiconductor and AI supply chain anchored to trusted partners. That relationship is not a contradiction to India's BRICS ambitions so much as the reason India, more than any other member, can plausibly sit at both tables without being owned by either.
That is precisely the case for hosting this conversation in New Delhi rather than leaving it to Rio's next chair or Shanghai's newest institution. India does not need BRICS to declare it the leader of the Global South on AI. It needs to leave this summit with one or two mechanisms -- a compute financing window, a language-data commons, a public-sector minimum standard -- that outlast the theme banners outside Bharat Mandapam.
(The views expressed are personal)
This article is authored by Sagar Vishnoi, co-founder and director, Future Shift Labs.