Reliance’s AI push highlights India’s reliance on foreign tech

In short: India wants to be a top global AI builder, but Reliance’s AI plans show the country still depends heavily on foreign technology.

What's going on

India’s government has set an ambitious goal for the country to become one of the world’s top powers in artificial intelligence. At an AI summit in New Delhi in February, Prime Minister Narendra Modi spoke about building AI systems in India and using them around the world.

Mukesh Ambani, the billionaire chairman of Reliance Industries, backed that message in his recent annual address to shareholders. He said Reliance aims to be “India’s sovereign AI backbone”, meaning a home-controlled foundation of AI services and computing power that Indian businesses can build on.

But the Financial Times reports that much of Reliance’s current strategy relies on technology made outside India. Reliance is investing in data centres (large buildings filled with computers) that use Nvidia hardware, and it is integrating large language models, which are AI systems trained on huge amounts of text (like an autocomplete that can write full paragraphs), from Google and Meta into its Jio telecom and digital services.

The article argues this reflects a wider pattern. Big global tech firms often treat India more as a fast-growing market than as a place where core AI technology is built. India also spends about 0.6 to 0.7 percent of its GDP on research and development, compared with roughly 2.5 percent in China and 3 to 3.5 percent in the US, and it lacks large, well-organised datasets needed to train advanced AI.

What to watch

A key question is whether India will focus on building useful AI applications for local needs, or invest more in the harder parts, like chips, research, and the data needed to train top models. Limits on access to the newest AI models, such as recent US export restrictions mentioned in the report, could make dependence on foreign AI feel more like renting than owning (like running your business from a leased office you might lose).

Source: Financial Times