India’s stance on Artificial Intelligence (AI) has moved from the margins of policy debate to the centre of economic decision-making. The previous economic surveys largely framed AI as a futuristic possibility, while the Economic Survey 2025-26 discusses its presence across banking, logistics, healthcare, education, and governance. Yet the real challenge before India is no longer adoption alone, but its alignment with economic structure, labour realities, and developmental priorities.
In 2025, global AI adoption reached nearly 88% of organisations, while only 7% had fully integrated it across operations. High-income countries account for nearly 58% of global AI deployment, while low- and middle-income countries together account for just over 40%. This skewed usage reflects a deeper asymmetry where, while AI use is spreading, the capacity to build and control advanced AI systems remains highly concentrated.
At the frontier of AI development lie large foundational models that demand enormous computing power, energy, data, and capital. Training a single frontier model can cost hundreds of millions of dollars, require town-scale electricity consumption, and place significant pressure on water resources through large data centres. Access to advanced chips and high-bandwidth memory is increasingly shaped by global supply constraints and export controls. For most countries, including India, attempting to replicate such capital-intensive pathways would impose prohibitive fiscal and environmental costs.
The Economic Survey 2025-26, therefore, advances the clear proposition that India’s AI strategy must be grounded in economic realism rather than technological imitation. The budgetary allocation for India’s AI mission underscores the practical challenges of implementation. The revised expenditure for the IndiaAI Mission has been scaled down from the budgeted Rs 2,000 crore to Rs 800 crore in 2025-26, with a similar allocation of Rs 1,000 crore for 2026-27, against the originally envisaged Rs 10,300 crore outlay over five years.
India ranks among the top global contributors to AI research publications, hosts one of the world’s largest pools of technical talent, and ranks second only to the United States in AI workforce literacy. Yet paradoxically, India accounts for only about 2% of global start-ups focused on training data curation, pointing to an underdeveloped domestic data value chain. This gap matters, as access to high-quality, context-specific data is becoming a critical bottleneck in AI development.
Given these conditions, the Economic Survey makes a strong case for a bottom-up, application-led AI ecosystem. India stands to gain more by developing AI models that are cheaper to train and easier to adapt to different sectors. The model requiring limited hardware will be best suited to India.
Locally grounded AI solutions are already emerging in India. In agriculture, AI-enabled platforms have improved price discovery and logistics for nearly 1.8 million farmers across 12 states. In healthcare, low-cost AI-based screening tools are expanding early cancer detection in resource-poor settings. Municipal bodies are deploying AI systems to monitor water usage, classroom learning outcomes, and disaster risks. Language technologies, particularly voice-based systems, are extending digital access to non-English speakers, addressing long-standing inclusion gaps.
The economic implications of AI adoption are nuanced. While the early fears of mass job displacement have not yet materialised, evidence from advanced economies suggests a more subtle shift. AI adoption has weakened the traditional relationship between output growth and employment creation, where productivity continues to rise, but employment growth has become less responsive to economic expansion.
In the policy context, India faces the dual challenge of AI adoption. On the one hand, delaying AI adoption to protect jobs could lock firms into low-productivity trajectories. On the other hand, rapid and uncalibrated automation may boost efficiency but risks displacing workers into low-value service roles faster than the economy can reabsorb them. As global firms increasingly automate routine coding, testing, and support functions, India’s traditional role as the world’s back office is under strain. Hence, India’s success in AI for growth will depend on the sequencing and speed of AI diffusion so that labour augmentation, reskilling, and institutional adjustment can occur in parallel.
Regulation without overreach
Internationally, AI regulation ranges from stringent risk-based frameworks to largely voluntary guidelines. India’s innovation ecosystem is still fragmented, resource-constrained, and heavily start-up-driven, and cannot bear the burden of stringent compliance requirements without stifling experimentation. At the same time, regulatory absence risks undermining trust, especially as AI systems are deployed in education, finance, and governance.
The latest Economic Survey report, therefore, argues for regulation that is light, incentive-based, and risk-weighted, emphasising transparency, product registration, and targeted oversight in high-impact applications. Governance of AI in India not only demands careful calibration but also the protection of human capability. Emerging evidence cited in the Survey suggests that excessive reliance on generative AI for cognitive tasks may weaken critical thinking and learning outcomes, particularly among students.
With over 100 crore broadband subscribers, India represents one of the world’s largest sources of human-generated data. Rather than imposing rigid data localisation, policy is shifting towards retaining economic value from domestic data while remaining open to cross-border flows. Hence, even with the emphasis on accountability and auditability in domestic AI data processing, a balance between strategic autonomy and global integration requires a clear set of rules. Complete self-sufficiency is neither feasible nor desirable, but excessive dependence on foreign AI systems, particularly in critical sectors, carries long-term risks.
AI is not an inevitable destiny but a strategic choice. Therefore, India needs to shape an AI ecosystem that is economically grounded, socially responsive, and institutionally aligned rather than chasing the technological frontier. In the AI era, restraint and realism may prove as important as speed.
(Prashant is an assistant professor at the Department of Public Policy, Manipal Academy of Higher Education; Alok is a senior research fellow at the Centre for Economic Studies and Policy, ISEC)
Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.