India began with an AI strategy (2018) but lacked the compute to compete. The IndiaAI Mission, approved in 2024 with a ₹10,372 crore outlay, changed that — pooling GPUs into a national facility offered to startups and researchers at subsidised rates.
Artificial Intelligence · IndiaAI
Artificial Intelligence · IndiaAI
India is racing to build sovereign AI. The IndiaAI Mission (2024), backed by ₹10,372 crore, set out to give startups and researchers cheap access to serious computing power — and has already onboarded over 38,000 GPUs (from an initial target of 10,000), offered at subsidised rates. It also funds Indian large language models trained on Indian languages, an open datasets platform (AIKosh), an AI Safety Institute, and AI skilling. India ranks among the top few nations for AI talent. The counter shows GPUs onboarded for common compute.
38,231
GPUs onboarded, against an initial target of 10,000
₹10,372 cr
IndiaAI Mission outlay
Top 4
Global AI-talent rank
AI compute (GPUs)
| Year | AI compute (GPUs) |
|---|---|
| 2023 | 0 GPUs |
| 2024 | 10,000 GPUs |
| 2025 | 38,000 GPUs |
| 2026 | 38,231 GPUs |
2023
0GPUs
AI compute (GPUs)
20232026
Since 2023
Added
+38,231 GPUs
2023
0 GPUs
2026
38,231 GPUs
Why it matters AI could add hundreds of billions of dollars to India's economy; owning the compute, data and models means the benefits and the guardrails stay in Indian hands.
- IndiaAI Mission (2024): ₹10,372 cr; 38,000+ GPUs onboarded
- Sovereign LLMs, AIKosh datasets, AI Safety Institute
- Source: MeitY / IndiaAI
History
Sovereign AI
From a target of 10,000 GPUs, India onboarded 38,231 by March 2026. The mission also funds Indian large language models trained on Indian languages (Sarvam, BharatGen and others), the AIKosh open datasets platform, an AI Safety Institute, and AI skilling — aiming for 'AI in India, and AI that works for India'.
How it works
India's AI strategy rests on shared public infrastructure rather than leaving it all to big private labs. The IndiaAI Mission pools tens of thousands of GPUs into a common compute facility that startups and researchers can rent cheaply, funds India-specific foundation models (like the multilingual BharatGen), and builds open datasets. The aim is 'sovereign AI' — models trained on Indian languages and data — plus using AI in governance, health and farming.
Outlook
The IndiaAI Mission, backed by about ₹10,300 crore, has already onboarded over 38,000 GPUs and picked startups to build Indian large language models across 22 languages. The next stage is sovereign capability: more AI compute and chips built at home under the India Semiconductor Mission, AI Centres of Excellence growing research and talent, and light-touch guardrails — so India shapes AI to its own languages and needs rather than importing someone else's models.
By the numbers
IndiaAI Mission (2024) ₹10,372 cr; 38,000+ GPUs; ≈$11 bn private AI investment (2013–24); top-4 AI talent; 2nd on GitHub AI. Source: MeitY / IndiaAI.
Data current to: As stated in the Lok Sabha on 25 March 2026
Source: MeitY / IndiaAI Mission — GPUs onboarded for common compute, 2024–2026. · link