India previews AI stack progress ahead of AI Impact Summit 2026: Details

The government has detailed advances across applications, models, compute, infrastructure and energy as New Delhi prepares to host the India AI Impact Summit from February 16

The India AI Impact Summit, scheduled from February 16 to February 20 in New Delhi, will focus on the democratisation of artificial intelligence as one of the topics under the themes of People, Planet and Progress. Ahead of the event, the government has shared updates on the India AI Stack, outlining how the country has been preparing to deploy AI at scale.

 

The India AI Stack is structured across five layers: applications, AI models, compute, data centres and network infrastructure, and energy. By strengthening each layer of the AI stack, India aims to expand access to AI and support its use across public and private sectors at population scale.

 

Here is a detailed snapshot:

AI adoption through high-impact applications

Indian startups are developing AI applications designed for local languages, regional contexts and sector-specific requirements, supporting wider adoption across the economy.

 

Agriculture: According to the note shared by the government, AI-based advisory tools are being used to improve sowing decisions, crop yields and input use. State-level deployments in Andhra Pradesh and Maharashtra have reported productivity gains of up to 30–50 per cent.

 

Healthcare: The government noted that AI applications are supporting early detection of tuberculosis, cancer, neurological disorders and other conditions, strengthening preventive and diagnostic services.

 

Education: The National Education Policy 2020 incorporates AI learning through CBSE curricula, the DIKSHA platform and initiatives such as YUVAi, with the aim of equipping students with applied AI skills.

 

Justice delivery: The government said that e-Courts Phase III uses AI and machine learning for translation, case management, scheduling and citizen-facing services, with a focus on improving access through vernacular languages.

 

Weather and disaster management: The India Meteorological Department uses AI for forecasting rainfall, cyclones, fog, lightning and fires. Tools such as Mausam GPT are being used to support farmers and disaster response agencies.

Development of the AI model layer

Under the IndiaAI Mission, the government has extended support to develop 12 indigenous AI models to address India-specific use cases.

 

To support sovereign model development, startups are provided subsidised access to compute resources, with up to 25 per cent of compute costs supported through a combination of grants and equity.

 

BharatGen is developing India-focused foundation and multimodal models, ranging from billions to trillions of parameters, to support research, startups and public-sector use.

 

IndiaAIKosh functions as a national repository for datasets, models and tools. As of December 2025, it hosts 5,722 datasets and 251 AI models, contributed by 54 entities across 20 sectors.

 

Startups are also building full-stack and domain-specific models aligned with Indian languages, healthcare needs and public service delivery. Sarvam AI is developing large language and speech models for Indian languages, while Bhashini, under the National Language Translation Mission, hosts more than 350 AI models covering speech recognition, translation, text-to-speech, optical character recognition and language detection.

Compute capacity and AI infrastructure

The IndiaAI Mission has been allocated more than ₹10,300 crore over five years.

 

The IndiaAI Compute Portal operates on a compute-as-a-service model, offering shared cloud-based access to 38,000 GPUs and 1,050 TPUs at subsidised rates of under ₹100 per unit, aimed at supporting startups and smaller organisations.

 

A secure national GPU cluster with 3,000 next-generation GPUs is being developed for sovereign and strategic AI use cases.

 

Under the India Semiconductor Mission, which has an outlay of ₹76,000 crore, 10 semiconductor projects have been approved, covering chip fabrication and packaging. Indigenous chip initiatives such as the SHAKTI and VEGA processors are contributing to domestic AI hardware development.

 

The National Supercomputing Mission has deployed more than 40 petaflops of computing capacity across IITs, IISERs and national research institutions. Systems such as PARAM Siddhi-AI and AIRAWAT support AI workloads including natural language processing, weather modelling and drug discovery.

Data centres and network infrastructure

A nationwide optical fibre network supports high-speed data movement for cloud and AI services. Fifth-generation mobile services have been rolled out across all States and Union Territories and are available in 99.9 per cent of districts, covering around 85 per cent of the population.

 

India accounts for about 3 per cent of global data centre capacity, with an installed capacity of around 960 MW. This is projected to increase to 9.2 GW by 2030, driven by growth in AI and cloud workloads.

 

Mumbai–Navi Mumbai remains the largest data centre hub, accounting for over a quarter of total capacity, followed by Bengaluru, Hyderabad, Chennai, Delhi NCR, Pune and Kolkata.

 

Global technology firms are investing in India’s AI and digital infrastructure. Announced commitments include Microsoft’s ₹1.5 lakh crore for data centres and AI training, Amazon’s ₹2.9 lakh crore for cloud infrastructure and AI-led digitisation by 2030, and Google’s ₹1.25 lakh crore investment for a 1 GW AI hub in Visakhapatnam.  ALSO READ | Claude Cowork explained: Anthropic's AI tool, plugins that spooked markets

Energy to support AI growth

India met a peak power demand of 242.49 GW in FY 2025–26, with energy shortages reduced to 0.03 per cent, supporting uninterrupted power supply for data centres and high-performance computing facilities.

 

Total installed power capacity stood at 509.7 GW as of November 2025. Non-fossil fuel sources account for 256.09 GW, or more than 51 per cent of total capacity.

 

The government plans to achieve 57 GW of pumped storage projects by 2031–32 and deploy 43,220 MWh of battery energy storage systems to improve grid stability and support data centres operating alongside renewable energy.

 

In addition, the Sustainable Harnessing and Advancement of Nuclear Energy for Transforming India (SHANTI) Act positions nuclear power as a continuous source of clean energy for AI and data centres. The legislation enables private sector participation and supports the deployment of small modular reactors and micro-reactors.