Indian start-ups are now actively building foundational AI systems within domestic compute and data ecosystems. has introduced large-scale foundation models, including a trained and hosted on India-based data centre infrastructure, alongside multilingual speech, document, and enterprise AI agent layers. Its partnerships with state governments to build AI-optimised data centres further demonstrate efforts to develop indigenous compute capacity—an important element of technological sovereignty. Similarly, BharatGen, a government-backed initiative, has launched , a multimodal foundation model supporting 22 Indian languages and designed as public digital infrastructure accessible to government services and research institutions. Other firms, such as Tech Mahindra and Gnani.ai, are developing domain-specific and multilingual AI systems focused on education and voice-based public interfaces, reinforcing India’s emphasis on locally relevant AI ecosystems.
This technological direction aligns with India’s policy vision articulated in the ), which identified healthcare, agriculture, education, smart mobility, and infrastructure as priority sectors. From the outset, the strategy framed AI not merely as a commercial technology but as a tool for social development and public service delivery, shaping subsequent investments in sector-specific AI applications.
Healthcare illustrates how this public-good direction is faring in practice. Indian firms are building specialised medical AI systems grounded in domestic datasets and clinical needs. Start-ups such as JiviAI’s have developed medical LLMs designed for clinical decision support, while Eka Care’s provides multilingual AI documentation tools compliant with Ayushman Bharat Digital Mission standards, reducing administrative burdens on doctors. Public-facing platforms such as Fractal Analytics’s further expand access by offering multilingual health information and report evaluation, helping reduce informational barriers in India’s access-oriented healthcare system. Additionally, indigenous AI-powered surgical technologies a growing ambition to integrate AI into advanced medical hardware.
In agriculture, AI deployments are even more explicitly oriented toward the public good due to state involvement and scale. Initiatives such as the AI-enabled under PM-KISAN, which handles millions of farmer grievances in multiple Indian languages, demonstrate how LLMs and Natural Language Processing (NLP) are being embedded in grievance redressal and welfare delivery systems to reduce bureaucratic delays. , combining open models such as Google’s NeuralGCM and the European Centre for Medium-Range Weather Forecasts (ECMWF)’s AI systems, has enabled the government to send early monsoon advisories to nearly 38 million farmers, materially influencing planting decisions and incomes. Domain-specific models such as BharatGen’s further highlight efforts to build India-centric foundational models trained on local data and agrarian contexts, delivering contextual farmer advisories, policy information, and research insights, among others, in response to agricultural queries.
Such application-layer successes are reinforced by institutional efforts under the , including public compute, datasets via , language infrastructure through , and sectoral Centres of Excellence. Taken together, India’s AI-for-public-good trajectory reveals a pattern of pragmatic, use-case-driven development that is capitalising on the country’s growing innovation ecosystem and tech-adaptive populace.
The Indian AI landscape today is characterised by strong technological capabilities, a commitment to responsible innovation, and growing structural readiness. India has the highest skill penetration level in the world. According to the Stanford AI Index Report 2025, India ranks among the top countries globally in , with of Indian enterprises already having several Generative AI use cases live, while another are in pilot stages, indicating strong experimentation. Surveys show that of Indian business leaders believe GenAI will have a significant impact, and a majority feel operationally ready to deploy it.
However, countries with high rates of AI diffusion tend to have a larger share of internet users or near-complete internet penetration. In countries such as the United States (US), the UAE, and Singapore, AI is diffusing rapidly due to high internet penetration—something India has yet to achieve. India’s AI infrastructure is still maturing, leaving its level of AI adoption behind that of leading AI-first societies. Countries such as the have achieved significantly higher AI usage penetration among their working populations, exceeding 60 percent, despite having far smaller populations. , too, despite its smaller demographic scale, shows higher proportional adoption. The also benefits from near-universal internet access, mature digital infrastructure, and high mobile internet usage, enabling AI tools to diffuse faster and more evenly. India’s challenge, therefore, is not demand or openness to AI, but incomplete digital inclusion and uneven institutional capacity. At present, only about of India’s internet users—approximately 429 million people—interact with AI-enabled features, while roughly have not yet done so.
In terms of AI-led industrial transformation, some have begun investing heavily in AI development and deployment, but over of organisations still allocate less than of their IT budgets to AI, creating a mismatch between ambition and execution. Specialised talent availability is another major constraint. Despite having the largest pool of AI learners globally, India ranks , resulting in a significant skills gap that directly affects the design, deployment, and governance of public-sector AI systems.
Thus, to more effectively realise its ambitions, India could move towards deeper public AI capacity building. Sustained investments in computing infrastructure for public compute and open foundational models trained on Indian data need to be actualised, while reducing reliance on external platforms. While India’s domestic GPU manufacturing plans are still at a relatively nascent stage, its current approach of making GPUs and compute infrastructure available at is proving transformational.
Finally, robust and enforceable institutional frameworks for AI governance, such as the recently released , should set the template for institutional norm-setting and implementation. With transparency and periodic evaluation, AI systems can be successfully deployed for welfare and public services, making them more trusted, accountable, and inclusive.
