India and AI: Time to build national capacity

This is where the conversation gets serious. In recent years, the Indian government has moved away from symbolic intent toward putting real architecture behind that intent. IndiaAI mission approved in March 2024 with an outlay of Rs 200 crore ₹10,371.92 crore in five years, is not a symbolic announcement. It is an effort to create the rails on which India’s AI future will run. Its design spans compute capacity, foundation models, datasets, application development, skilling, startup financing, and secure and reliable AI. That width matters. This shows that the government understands the basic reality of this moment. AI leadership is not created through isolated innovation. It is created through the ecosystem.

That policy direction is in line with the Prime Minister’s broader vision of democratizing technology. In the Indian context, this phrase holds strategic importance. Democratization cannot just mean access to digital tools. It should also mean access to opportunity, access to potential and access to preparation for the future. India does not need an AI economy that benefits a narrow layer of elite institutions and advanced firms while leaving the larger workforce behind. It needs an AI path that speaks to India’s scale, India’s linguistic diversity, India’s developmental challenges and India’s public interest priorities. That is why this mission matters beyond the technology sector. It is in many ways a nation building project.

One of the most encouraging aspects of India’s approach is the recognition that AI literacy should become a fundamental skill. The Youth AI for All campaign, associated with the IndiaAI Mission, reflects this shift by treating AI awareness not as an advanced expertise but as a basic requirement for the future citizen and future worker. The numbers already indicate the scale. Official figures show that 4,69,951 learners have registered for the foundational AI course and 1,31,785 have completed it. This is a promising start as it expands the base of participation and signals that AI cannot remain confined to policy forums, engineering campuses or high-level laboratories.

But this is where India should avoid complacency. Literacy is essential, but literacy alone is not readiness. A country does not become AI capable just because people can take a short course, recognize common terminology or experiment with public tools. True readiness begins when organizations can integrate AI into how they work, how they train, how they make decisions, and how they deliver results. The next phase of India’s AI strategy must therefore move decisively from awareness to application.

This shift is particularly urgent in sectors like health care, education, agriculture, and public administration, where the promise of AI is not abstract. This is practical. For example, in healthcare, India does not need AI just for advanced diagnostics or future research. AI is needed to strengthen workforce readiness, expand quality training, reduce variability in learning risk, and support a system under constant pressure. In such a situation, AI should not be discussed only as a software layer. This should be understood as the capability layer.

This is where the policy debate should be more mature. India has often been strong in announcing technology, celebrating pilots and showcasing innovation. The more difficult task has always been institutionalization. Now this is the challenge before us. AI will not change India because it is enthusiastically talked about. It will transform India only when it is integrated into the operating logic of institutions. The measure of progress will not be the number of events, platforms or headlines around AI. It will be whether AI meaningfully improves how our classrooms train, how our hospitals learn, how our startups are built and how our public systems respond.

There are reasons for optimism. Under the IndiaAI compute capability pillar, more than 38,000 GPUs have already been onboarded through listed service providers, with access being offered at discounted rates to eligible users. This is a major intervention as computation remains one of the most significant barriers to entry in AI development. When cost-effectiveness is expanded, innovation is no longer limited to only the players with the most capital. It becomes possible for startups, researchers, institutions and emerging developers to participate in the ecosystem with more seriousness.

There is movement on the model development front as well. The official update in March 2026 confirmed that twelve teams were shortlisted in the first phase for indigenous foundational AI models, and models developed by Sarvam AI, BharatGen, Gyani and Socket were launched during the IndiaAI Impact Summit 2026. This is an important policy signal. India is not positioning itself as just a consumer market for global AI systems. It seeks to build domestic capacity in core technologies with relevance to Indian languages ​​and Indian use cases. This is not just a matter of prestige. This is a matter of strategic necessity.

The linguistic complexity of India makes this even more important. A country with continental diversity cannot rely indefinitely on systems primarily trained for other societies, other contexts, and other perceptions. If AI is to create broader developmental value in India, it must understand Indian languages, Indian data realities, and Indian public service conditions. This is where the connection between AI policy and digital inclusion becomes particularly strong. Indigenous models and open innovation platforms are not just technological achievements. They are means of access.

The skill dimension is equally important. The government update has highlighted support targets for 500 PhD fellows, 5,000 postgraduates and 8,000 graduates under the IndiaAI FutureSkills pillar. He also mentioned the setting up of 27 data and AI labs in tier 2 and tier 3 cities and approval for 543 more labs in ITIs and polytechnics. This is exactly the kind of distributed talent strategy that India needs. The future of Indian AI cannot be concentrated in a few metro clusters. If the country is serious about democratization, it will have to be broadened socially and geographically.

And yet, policy ambition must now face the reality of implementation. This is the point where many national missions lose momentum. The initial vision is compelling. The institutional language is strong. The ecosystem is active. But adoption remains patchy, standards are underdeveloped, regional pathways are unclear and public institutions struggle to integrate new capabilities into old structures. India must not let AI become another domain where ambition exceeds absorption.

There is now a need for a more performance based framework. The country needs sector specific adoption roadmaps, institutional incentives, curriculum level integration, implementation standards and credible public private collaboration mechanisms that go beyond event-based enthusiasm. In health care and medical education, this is especially important. India has a huge opportunity to combine AI with simulation-based training, competency building and adaptive learning systems so that the technology not only strengthens diagnosis and administration, but also preparedness and professional judgment. If we want the health care workforce of the future to be truly future ready, AI must become part of our training, rehearsal, assessment, and improvement.

This is a big national point. India’s AI moment will ultimately not be decided by the scale of its announcements. This will be determined by the depth of its institutional adoption. The countries that will lead in AI in the next decade will not just be those building powerful models. They will be the ones who will build an efficient society around those models. They will be those who combine innovation with skills, infrastructure with inclusion and technology with trust.

India has the material to do so. It has the demographic depth, public digital experience, entrepreneurial energy, linguistic diversity that can drive model innovation, and now a more serious policy architecture than before. But it should move rapidly from aspiration to system building. The future of Indian AI will not be secured by code alone. This will be secured by how effectively the country links computers to classrooms, research to relevance, policy to implementation and intelligence to public purpose.

This is the challenge before India. This is also an opportunity. If the country gets this right, AI will not become just another technological success story. This will become a new chapter in India’s development story.

This article is written by Anil Agarwal, former Rajya Sabha MP, and Aadith Chinnaswamy, COO of Medisim VR.