AI Report from MeitY

[ABS News Service/07.03.2025]

AI Compute and Semiconductor Infrastructure

India is rapidly building a strong AI computing and semiconductor infrastructure to support its growing digital economy. With the approval of the IndiaAI Mission in 2024, the government allocated ₹10,300 crore over five years to strengthen AI capabilities. A key focus of this mission is the development of a high-end common computing facility equipped with 18,693 Graphics Processing Units (GPUs), making it one of the most extensive AI compute infrastructures globally. This capacity is nearly nine times that of the open-source AI model DeepSeek and about two-thirds of what ChatGPT operates on.

Here are the key developments:

· Scaling AI Compute Infrastructure: The initial phase of the mission has already made 10,000 GPUs available, with the remaining units to be added soon. This will enable the creation of indigenous AI solutions tailored to Indian languages and contexts.

· Opening Access to High-Performance Computing: India has also pioneered the launch of an open GPU marketplace, making high-performance computing accessible to startups, researchers, and students. Unlike many countries where AI infrastructure is controlled by large corporations, this initiative ensures that small players have an opportunity to innovate.

· Robust GPU Supply Chain: The government has selected 10 companies to supply the GPUs, ensuring a robust and diversified supply chain.

· Indigenous GPU Capabilities: To further strengthen domestic capabilities, India aims to develop its own GPU within the next three to five years, reducing reliance on imported technology.

· Affordable Compute Access: A new common compute facility will soon be launched, allowing researchers and startups to access GPU power at a highly subsidised rate of ₹100 per hour, compared to the global cost of $2.5 to $3 per hour.

· Strengthening Semiconductor Manufacturing: In parallel, India is advancing semiconductor manufacturing, with five semiconductor plants under construction. These developments will not only support AI innovation but also reinforce India’s position in the global electronics sector.

Advancing AI with Open Data and Centres of Excellence (CoE)

Recognising the importance of data in AI development, the government has launched the IndiaAI Dataset Platform to provide seamless access to high-quality, non-personal datasets. This platform will house the largest collection of anonymised data, empowering Indian startups and researchers to develop advanced AI applications. By ensuring diverse and abundant datasets, this initiative will drive AI-driven solutions across key sectors, enhancing innovation and accuracy.

· Centres of Excellence: The government has established three AI Centres of Excellence (CoE) in Healthcare, Agriculture, and Sustainable Cities in New Delhi. The Budget 2025 further announced a new CoE for AI in education with an outlay of ₹500 crore, making it the fourth such centre.

India’s AI Models & Language Technologies

The government is facilitating the development of India’s own foundational models, including Large Language Models (LLMs) and problem-specific AI solutions tailored to Indian needs. To foster AI research, multiple Centres of Excellence have also been set up.

· India’s Foundational Large Language Models: IndiaAI has launched an initiative to develop indigenous foundational AI models, including LLMs and Small Language Models (SLMs), through a call for proposals.

· Digital India BHASHINI: An AI-led language translation platform designed to enable easy access to the internet and digital services in Indian languages, including voice-based access, and support content creation in Indian languages.

· BharatGen: The world’s first government-funded multimodal LLM initiative, BharatGen was launched in 2024 in Delhi. It aims to enhance public service delivery and citizen engagement through foundational models in language, speech, and computer vision. BharatGen involves a consortium of AI researchers from premier academic institutions in India.

· Sarvam-1 AI Model: A large language model optimised for Indian languages, Sarvam-1 has 2 billion parameters and supports ten major Indian languages. It is designed for applications such as language translation, text summarisation, and content generation.

· Chitralekha: An open-source video transcreation platform developed by AI4Bhārat, Chitralekha enables users to generate and edit audio transcripts in various Indic languages.

· Hanooman’s Everest 1.0: A multilingual AI system developed by SML, Everest 1.0 supports 35 Indian languages, with plans to expand to 90.

· AI Talent Hubs: The India Skills Report 2024 by Wheebox forecasts that India’s AI industry will reach USD 28.8 billion by 2025, with a CAGR of 45%. The AI-skilled workforce has seen a 14-fold increase from 2016 to 2023, making India one of the top five fastest-growing AI talent hubs, alongside Singapore, Finland, Ireland, and Canada. The demand for AI professionals in India is projected to reach 1 million by 2026.

· AI Startup Support Ecosystem: India hosts 520+ tech incubators and accelerators, ranking third globally in active programs. 42% of these were established in the past five years, catering to the evolving needs of Indian startups. AI-focused accelerators like T-Hub MATH provide crucial mentorship in product development, business strategy, and scaling. In early 2024, MATH supported over 60 startups, with five actively discussing funding, highlighting India's growing AI startup landscape.