India Commits Billions to Sovereign AI Infrastructure
India is positioning itself as a major AI power with plans for massive infrastructure investment. At the AI Impact Summit, officials highlighted $200 billion in AI investments and VC commitments, while business leader Jeet Adani stated a $100 billion commitment for sovereign AI infrastructure. This investment is seen as an accelerator for agentic systems and persistent AI workflows, with India aiming to shape global AI policy.
- The government's investment is anchored by the IndiaAI Mission, a five-year program with an approved outlay of ₹10,372 crore (over $1.2 billion) aimed at building a comprehensive AI ecosystem. - A core component of the strategy is a public-private partnership to expand compute capacity, with the government providing Viability Gap Funding to subsidize up to 50% of the cost for private entities and startups to access GPU infrastructure. - India is significantly scaling its physical hardware, with 38,000 GPUs currently operational under its AI mission and plans to add another 24,000 units within the next six months to support developers and researchers. - Technology partnerships are central to the infrastructure build-out, with NVIDIA collaborating with cloud providers like Yotta, L&T, and E2E Networks to create "AI factories," including Yotta's "Shakti Cloud" which will be powered by over 20,000 NVIDIA Blackwell Ultra GPUs. - A key objective is the development of indigenous foundational models trained on local data; one such initiative, BharatGen, has already produced a 17-billion-parameter Mixture-of-Experts (MoE) model designed to support India's 22 recognized languages. - To foster the startup ecosystem, Qualcomm has launched a $150 million Strategic AI Venture Fund for Indian startups, focusing on on-device AI applications in sectors like automotive and IoT. - The Adani Group's $100 billion commitment through 2035 is specifically targeted at building green-energy-powered hyperscale data centers, aiming to leverage India's renewable energy potential for the high energy demands of AI workloads. - International collaboration is also scaling up, with Abu Dhabi's G42 and Cerebras partnering with India's Centre for Development of Advanced Computing (C-DAC) to establish an 8 exaflop AI supercomputer that will operate under Indian data governance frameworks.
