India’s Ministry of Finance, in its annual Economic Survey for the financial year 2025-26, said that India’s strategy for artificial intelligence (AI) must be sequenced carefully to avoid premature lock-ins or regulatory overreach; hence, the nation should focus on a guarded stance amid heightened uncertainty and resource constraints.
The country’s objective is to build coordination first, capacity next, and binding policy leverage last, allowing institutions and markets to co-evolve, according to the Economic Survey released on Thursday, 29 January 2026.
“Artificial Intelligence does not confront India with a single policy question, but a series of choices that must be made under conditions of heightened uncertainty and resource constraints,” as per the survey.
AI Roadmap for India
The Economic Survey charted that the first phase should focus on operationalising already announced institutions and aligning incentives to enable experimentation. Policy should enable bottom-up innovation by expanding the reach of the existing shared infrastructure under the IndiaAI Mission.
This includes a government-hosted community-curated code repository and pooled access to public datasets, facilitated by initiatives already underway to enable shared access to computing infrastructure. A clear focus on application- or sector-specific, small and open-weight models will enable efficient resource utilisation.
Once coordination mechanisms are functional and early experimentation has generated evidence, policy can shift toward selective scaling in the medium-term. Shared and certified domestic computing infrastructure should expand, with voluntary participation by large and resourceful firms linked to regulatory facilitation and access to public datasets. At the same time, AI regulation should be formalised on a risk-based and proportionate basis.
Graduated obligations for AI firms should be codified according to scale and sector of use. Oversight must be embedded within existing sectoral regulators rather than through a single omnibus AI law. The AI Safety Institute’s role should deepen from analysis to structured scenario testing, red-teaming, and international cooperation, with clearly articulated non-negotiable boundaries for high-risk applications.
Our long-term goals must encompass two main objectives. First, India’s focus should shift towards resilience. Access to advanced computing hardware will require strategic partnerships and diplomacy. The objective must be to reduce India’s vulnerability to external shocks. Second, sustained adaptation of labour markets and education systems will be essential. Primary education must prioritise foundational cognitive and socio-emotional skills, while skilling systems must align themselves with both AI- and human-centric sectoral requirements.
Given the constraints related to capital, computing capacity, energy, and infrastructure, pursuing scale for its own sake is neither efficient nor necessary. Instead, the chapter makes the case that a bottom up, multiple sector-specific approaches under a single vision has the potential to pay dividends and turn into a source of dignified employment for India’s youth. India’s development of AI must be grounded in open and interoperable systems to promote collaboration and shared innovation. This pathway aligns more closely with India’s strengths in human capital, data diversity and institutional coordination.
The proposed framework for data governance strikes a balance between openness to cross border flows and strengthening accountability and regulatory visibility. It is rooted in the objective of ensuring that the value accruing from India’s domestic data is retained within the country for the benefit of the people. The government’s role is framed as that of an enabler and coordinator, helping markets and institutions adjust in step with technological change.
Overall, the chapter treats AI as a strategic choice. The central message is that India’s opportunity lies in deploying AI in a way that is economically grounded and socially responsive.
AI is no longer a distant or speculative technology. It is increasingly being adopted, even if in an experimental capacity, in organisations around the world. Based on a survey of 1993 firms by McKinsey, 88% of organisations surveyed in 2025 reported that they are utilising AI in at least one of their business functions. Of those using it, 31% are in the process of scaling its application across the organisation, while 7% have already fully deployed and integrated AI.
Innovations and continuous improvement in AI capabilities are driving firms and new start-ups to develop ways in which AI can be applied to solve real-world problems.
At the same time, greater visibility into AI adoption has also brought greater clarity on the nature of the technology itself. Over the past year, it has become evident that while the use of AI tools can be widespread, the frontier of AI remains highly concentrated. The development and training of advanced foundational models is increasingly capital-, compute-, data- and energy-intensive, favouring a small set of firms with access and the political capital to secure large-scale infrastructure projects, specialised hardware, and deep pools of technical talent.
Early evidence has also begun to temper some of the more extreme predictions surrounding AI’s near-term labour impact. For instance, a study conducted by Yale’s Budget Lab indicates that the broader labour market in the United States has not experienced a discernible disruption due to AI.
This does not invite complacency, especially from a policymaker’s perspective. While labour may be complemented in the near term as organisations work to incorporate AI into their tasks, productivity gains from augmentation have a ceiling. All in all, caution is still warranted as India attempts to solve the puzzle of AI and labour. This represents one of the most considerable looming uncertainties about the technology.
