India bets big on AI: Launches major compute facility, unveils plans for indigenous GPU chips

India has significantly strengthened its AI capabilities by launching a large-scale GPU compute facility through its IndiaAI initiative, Electronics and IT Minister Ashwini Vaishnaw announced on Wednesday. Aimed at researchers, startups, and developers, the platform provides advanced computing resources at an affordable price of Rs 67 per GPU hour, accelerating innovation across various domains.

Speaking at the launch, Vaishnaw said, “Every quarter, we will add more GPUs,” emphasising that a substantial portion of the facility will support the creation of India’s own foundational AI models. Currently, 67 applications are under review, with 22 dedicated to developing large language models.

“Every quarter, we will add more GPUs,” Vaishnaw said, adding that a significant portion of the common compute facility will be dedicated to building India’s own foundation model. He further highlighted affordability, noting, “The way we sent a mission to the moon at a fraction of the cost at which the developed world did, we’ll have our foundational model also at a fraction of the cost that some of the rich world countries have done.”

Vaishnaw also announced ambitious goals to boost India’s technology status globally, saying, “In three to five years, India will be among the top five technology nations in areas like AI, semiconductors, and deeptech.” The minister further revealed that India aims to develop its own GPU chips within the next three years, insulating itself from international supply restrictions such as the US’s export cap on GPUs.

Highlighting India’s unique approach to AI development, Vaishnaw noted the country’s strategy is techno-legal, addressing both technological advancement and regulatory frameworks simultaneously. “The way we sent a mission to the moon at a fraction of the cost at which the developed world did, we’ll have our foundational model also at a fraction of the cost that some of the rich world countries have done,” he added.

Abhishek Singh, CEO of the IndiaAI Mission, outlined additional efforts such as a startup acceleration programme in partnership with Station F, and recognised student fellowships as part of the mission’s focus on skill development. An AI Competency Framework for civil servants, delivered via the iGOT Karmayogi platform, was also introduced to strengthen AI skills within the public sector.

Emphasising India’s structured approach, Vaishnaw contrasted India’s planned governmental support with the predominantly private-sector approach in countries like the US, highlighting a targeted, strategic deployment of AI across clearly defined sectors.

India’s GPU compute facility, accessible at Rs 67 per GPU hour, aims to democratise access to advanced AI resources, inviting contributions of non-personal data from private entities as part of the larger initiative.