PIB EXPLAINER:
India’s technology future is being shaped by a powerful principle: the democratisation of Artificial Intelligence. The national vision of AI for Humanity emphasises that advanced technologies must serve every citizen, strengthen public welfare, and support inclusive growth. Turning this vision into reality requires AI to work reliably at population scale across sectors such as healthcare, agriculture, education, finance, and governance. India’s emerging AI Stack provides exactly this integrated foundation, bringing together applications, models, compute, infrastructure, and energy into a cohesive ecosystem capable of delivering real-world impact.
An AI stack is the full set of tools and systems that enable artificial intelligence to function end-to-end. From user-facing applications to the computing power and energy running behind the scenes, each layer plays a critical role in ensuring AI solutions are scalable, reliable, and accessible. India’s strategy focuses on strengthening every layer simultaneously so that innovation is not confined to isolated pilots but embedded into everyday service delivery.
At the top sits the application layer, where AI directly touches citizens’ lives. Across the country, AI applications are helping farmers make informed decisions, assisting doctors in early disease detection, and improving the efficiency of public services. India’s “AI diffusion” approach prioritises widespread adoption of real-world use cases so that technology translates into measurable social and economic gains.
Beneath applications lies the AI model layer, the intelligence engine of the ecosystem. Under the IndiaAI Mission, 12 indigenous AI models are being developed to address India-specific needs. Startups building these sovereign models receive subsidised compute support, with up to 25 percent of costs covered through grants and equity, lowering entry barriers. BharatGen is advancing India-centric foundation and multimodal models ranging from billions to trillions of parameters, while IndiaAIKosh has emerged as a national repository hosting 5,722 datasets and 251 AI models contributed by 54 entities across 20 sectors as of December 2025. Indian innovators are also pushing the frontier: Sarvam AI is developing large language and speech models for Indian languages, and Bhashini under the National Language Translation Mission already hosts more than 350 AI models enabling multilingual digital access.
Powering these models is the compute layer, the muscle of AI. India has committed over ₹10,300 crore to the IndiaAI Mission and is expanding affordable access through the IndiaAI Compute Portal, which provides shared cloud access to more than 38,000 GPUs and 1,050 TPUs at subsidised rates below ₹100 per hour. A secure national GPU cluster with 3,000 next-generation processors is also being established for strategic applications. Complementing this effort, the ₹76,000-crore India Semiconductor Mission has approved 10 semiconductor projects, while indigenous chip initiatives such as SHAKTI and VEGA processors are strengthening domestic hardware capability. The National Supercomputing Mission has already deployed over 40 petaflops of capacity, supported by flagship systems like PARAM Siddhi-AI and AIRAWAT.
Equally critical is the data centres and network infrastructure layer, the home and highways of AI. India’s nationwide optical fibre backbone and rapid 5G rollout, now covering 99.9 percent of districts and 85 percent of the population, ensure high-speed data movement. The country currently accounts for about 3 percent of global data centre capacity with roughly 960 MW installed, projected to expand sharply to 9.2 GW by 2030. Mumbai–Navi Mumbai leads as the largest hub, alongside Bengaluru, Hyderabad, Chennai, Delhi NCR, Pune, and Kolkata. Global technology leaders are reinforcing this momentum with major commitments, including Microsoft’s ₹1.5 lakh crore investment, Amazon’s ₹2.9 lakh crore cloud expansion, and Google’s ₹1.25 lakh crore AI hub in Visakhapatnam.
Supporting the entire ecosystem is the energy layer. India met a record peak demand of 242.49 GW in FY 2025–26 while keeping shortages to just 0.03 percent, ensuring reliable power for AI workloads. Total installed capacity has reached 509.7 GW, with non-fossil sources contributing over 51 percent. Plans to develop 57 GW of pumped storage and 43,220 MWh of battery storage will further stabilise the grid. The SHANTI Act is also positioning nuclear energy, including Small Modular Reactors, as a dependable clean power source for energy-intensive AI infrastructure.
Together, these coordinated investments are building a resilient, sovereign, and inclusive AI ecosystem. By strengthening applications, models, compute, infrastructure, and clean energy in parallel, India is ensuring that artificial intelligence scales responsibly and equitably. Anchored in the vision of AI for Humanity, the India AI Stack is transforming data and compute into real-world public value, advancing innovation, self-reliance, and welfare for all in the digital era.
Courtesy: Press Information Bureau, Srinagar
