AI Summit 2026: The govt’s focus is on scalable and efficient AI models: Jitin Prasada

The focus of the government is on scalable and efficient models rather than only building massive systems, minister of state for electronics and information technology Jitin Prasada said on Monday at the IndiaAI Impact Summit.

He added that GPU access is being provided at subsidised rates, and that the government is working to make other resources available as widely as possible.

"Democratisation of technology is the need of the hour," he noted while addressing a panel discussion titled, 'From AI User to Creator: The Next Leap of India’s AI Innovation'. "The government is doing everything it can to make this happen. We are already beginning to see the results."


The panelists agreed that India’s scale of data, digital public infrastructure, and rapid enterprise digitisation have made it a massive AI adoption market. The next strategic shift, however, would be to climb up the value chain by building indigenous models, platforms, and applications instead of primarily deploying global technologies.

Pratyush Kumar, cofounder of Sarvam AI, echoed Prasada's views and said that “a frugal, scale-first approach" was the path India should take.

"The immediate priority should be making AI accessible to everyone, while simultaneously investing in deeptech capabilities to compete globally,” he said, while adding that large language models still need heavy engineering work, especially around nuanced benchmarks and evaluation systems.

“AI has progressed more in the last five years than in the previous 20. The challenge now is to democratise — not gatekeep — it,” Kumar said.

Santhosh Viswanathan, managing director at Intel, highlighted that India’s education and talent landscape requires a bottom-up strategy, with AI built for grassroots adoption and diverse use cases rather than mirroring western models. The infrastructure, he noted, must reflect India’s social and economic realities.

Navrina Singh, founder, Credo AI, made a strong case for embedding governance directly into AI systems. Trust, she said, should function as `infrastructure’ rather than a box-ticking compliance exercise. With India increasingly consuming frontier models developed abroad, the country must understand the risks of third-party dependence and lead in creating trustworthy AI frameworks. This includes transparent supply chains, robust evaluation methods, and India-specific benchmarks.