IndiaAI Mission Powers Ahead with Subsidised GPUs, Indigenous Models & Skill Push
- 34,381 GPUs onboarded at low hourly rates to boost AI development.
- 30 AI apps approved; 4 startups selected to build India-specific foundation models.
- 200+ AI fellowships granted; AI labs planned in Tier 2/3 cities.
In a significant push towards becoming a global leader in artificial intelligence, India has onboarded 34,381 GPUs through 14 empanelled service providers under its national AI strategy, Union Minister Ashwini Vaishnaw informed the Lok Sabha. These GPUs are being provided at subsidised rates, with the average price at Rs 65 per GPU per hour. Notably, the H100 GPUs, crucial for foundational model training, are offered at Rs 92 per hour significantly cheaper than rates charged by commercial cloud hyperscalers.
India's rapidly growing IT ecosystem, which generates over $250 billion annually and employs more than 6 million people, provides a strong foundation for AI advancements. According to global indices like the Stanford AI rankings, India ranks among the top nations in AI skills, policy frameworks, and capabilities. The country is also the second-largest contributor to AI-related projects on GitHub, reflecting a robust and vibrant developer community.
To harness this potential, the government launched the IndiaAI Mission in March 2024. This comprehensive initiative is designed to build a robust, inclusive, and innovation-driven AI ecosystem aligned with national development goals. One of the key pillars of this mission is the IndiaAI Application Development Initiative, which supports the development of AI solutions tailored to Indian challenges in areas like climate change, agriculture, healthcare, governance, disaster management, and assistive learning. To date, 30 AI applications have been approved under this initiative.
The government has also launched AIKosh, a unified data platform that integrates both governmental and non-governmental datasets. The beta version, launched in March 2025, already hosts more than 890 datasets, 208 AI models, and over 13 development toolkits to help developers focus on core AI functionality.
In a bid to ensure sovereign capability and global competitiveness in generative AI, the IndiaAI Foundation Models initiative aims to build indigenous Large Multimodal Models (LMMs) trained on Indian datasets and languages. From over 500 received proposals, four start-ups Sarvam AI, Soket AI, Gnani AI, and Gan AI have been selected in the first phase.
To further strengthen the talent pipeline, the IndiaAI FutureSkills initiative has provided fellowships to over 200 students in its first year. It has partnered with 26 institutes to support PhD students and plans to establish AI and Data Labs in Tier 2 and Tier 3 cities, fostering innovation across the country.
