More than a year after the Union Cabinet approved the IndiaAI Mission with a budget of ₹10,371.92 crore, the programme has moved well beyond its opening targets on infrastructure while its workforce and skilling arms are still in early stages.
The mission's compute pillar has grown faster than planned. An initial target of 10,000 GPUs has been surpassed, with over 38,000 GPUs now onboarded through a subsidised national facility and made available to startups, researchers and students at ₹65 per hour. The AIKosh national dataset platform has expanded significantly, now hosting over 9,500 datasets and 273 sectoral models.
On model development, the IndiaAI Innovation Centre received 67 proposals in its first month and has since identified 20 sovereign AI model projects for support, spanning large language models and small language models. Sarvam AI, Gnani.AI and BharatGen are among the teams building indigenous foundation models, with domain-specific work under way in healthcare, language technologies and agentic AI.
The FutureSkills pillar aims to support 500 PhD fellows, 5,000 postgraduates and 8,000 undergraduates. Thirty Data and AI Labs have been launched as the first wave of a planned 570-lab network, with the labs set up in partnership with NIELIT and Intel to deliver foundational AI and data training, including in Tier 2 and Tier 3 cities.
On the governance side, India released AI governance guidelines in November 2025 built around a risk-based, light-touch framework. The IndiaAI Safety Institute, announced in January 2025, is overseeing safety research and has approved 13 projects targeting deepfakes, bias and misinformation. Thirty sector-specific AI applications had been approved by July 2025 in areas including healthcare, agriculture and cybersecurity.
India hosted the AI Impact Summit 2026 in February, positioning itself as the first Global South nation to chair such a gathering and drawing commitments on responsible frontier AI from leading developers worldwide.
