Why India may need a middle-power AI alliance to secure digital sovereignty

At the recent AI Safety Connect (AISC) strategic briefing on the sidelines of the ongoing India AI Impact Summit 2026 , a critical question was raised: can India truly compete in the global artificial intelligence (AI) arms race by going it alone? While the IndiaAI Mission is a massive leap forward, with its focus on GPU procurement, sovereign infrastructure, and data localisation, experts are beginning to wonder if a change in equation is required to beat the silicon giants of the West and the East.

In an exclusive interaction with Dataquest, Mark Brakel, Global Director of Policy at the Future of Life Institute (FLI), offered a provocative alternative. He suggests that for India to secure its digital future, it must look beyond its borders, not to Silicon Valley, but to “middle power” peers like Brazil, South Korea, and France.

THE BILLION-DOLLAR GAP

The current Indian strategy is centred on building domestic compute capacity. However, the sheer scale of investment from US corporations, often exceeding the entire AI budget of medium-sized nations, creates a daunting barrier.

Mark argues that while India’s mission is ambitious, it may still be under-resourced compared to the sheer capital flight of companies like OpenAI or Anthropic. The solution? A “CERN for AI”. Just as the European Organization for Nuclear Research (CERN) allowed nations to pool resources for physics breakthroughs no single country could afford, a collective AI model could provide the “sovereign intelligence” India needs.

THE CASE FOR A COLLECTIVE MODEL

During the briefing, Mark elaborated on why the “solo” approach to foundation models might be a strategic trap for middle powers.

“I think it would be good for the Indian government to do this in partnership with other middle powers. Because I think India alone can't compete with the billions that the corporations in the US and in China are investing. Even though it's a big country, if you look at the numbers, it is still a much smaller amount than the India AI mission has, compared to other countries.
Hence, I think the recommendation is; like you have in Geneva for physics research, where India also participates, is maybe a better way forward. This is where the Indian money gets pooled with the Brazilian money, the South Korean money, the French money. Because currently, all countries are trying the same thing, and I don't think they will win out.
It would be a better strategy for these countries to come together, build an AI model, and utilise them in a sovereign manner to protect their data. We need to think about: what do we want to use it for? What is important to our population? We need it for healthcare; we need it for personal tuition. We don't know what OpenAI wants it for, but presumably, they are mainly interested in replacing workers because they need to earn back what they have invested. And those might not be in India's or Brazil's national interest.”

ALIGNING AI WITH NATIONAL INTEREST

The crux of the argument is not just about compute power, but intent. Silicon Valley models are built on a venture capital mandate: high returns, often achieved through automation and labour replacement.

For India, the priorities are fundamentally different. The “national interest” involves:

• Democratising education.

• Healthcare access with diagnostic tools for rural populations.

• Sovereignty to ensure that the “brain” of the country’s digital infrastructure is not owned by a foreign corporation.

THE PATH FORWARD

If India pivots towards leading a middle-power alliance, it could set the standard for “public good AI”. By pooling data and funds with nations that share similar democratic and social goals, India might ensure that it is not just a consumer of AI applications, but a co-owner of the underlying intelligence.

The IndiaAI Mission has laid the foundation. The next step may be to build the bridge.