by Sri Varshith Kumar Reddy E
As India’s AI Summit concludes at Bharat Mandapam, attention must shift from optics to structural economics. AI is now reshaping productivity, labour-capital dynamics, and fiscal federalism, yet India lacks a cohesive regulatory framework. In this context, Jammu and Kashmir’s AI Framework and proposed Centre of Excellence position it as a governance-driven, inclusion-focused testbed that could outpace richer states in policy innovation.
The India–AI Impact Summit 2026 convened at Bharat Mandapam with the familiar ceremonial energy of national ambition on display with ministerial affirmations, platform demonstrations, and the polished choreography of a country that has determined, with some justification, that it belongs at the centre of the next global technological order. What no summit agenda can choreograph, however, is the macroeconomic reality the event itself reflects. Artificial intelligence has ceased to function as a technological sector and become, instead, a condition of the economy, a factor of production that reorganises how output is generated, how income is distributed, and how fiscal capacity is allocated across an uneven federal architecture.
This distinction carries weight. When economists describe capital deepening, they mean the process by which economies raise output per worker through sustained investment in productive assets, such as more machinery, deeper infrastructure, and better tools. Artificial intelligence extends this logic into the realm of cognition. Firms invest in data, algorithms, and compute with the strategic discipline once reserved for plant and equipment; the productive asset is now embedded in software that learns and adapts rather than in steel that merely operates. Aggregate productivity can rise substantially. The distribution of those gains follows the logic of capital accumulation, not broad-based wage growth, and that difference is the crux of the matter.
The Geography of Concentrated Returns
India has encountered this asymmetry before, in attenuated form. The information technology services expansion of the 1990s concentrated extraordinary productivity gains within a narrow geography and a narrow occupational band, while the wider labour market absorbed secondary effects slowly and at considerable social cost. Artificial intelligence portends a more acute version of the same pattern, for a structural reason: unlike outsourced services, which at least required large workforces in the cities where they operated, AI-generated value demands minimal labour relative to its output. When value is created by models trained in one jurisdiction, owned in another, and consumed across a third, the architecture of taxation becomes contested in ways that existing Indian fiscal arrangements of Union List, State List, and Concurrent List were not designed to resolve.
Labour markets, meanwhile, divide along a more revealing axis than the familiar one of skilled and unskilled. The salient distinction is between those capable of supervising, auditing, and contesting algorithmic outputs and those whose work is simply mediated by such systems, without meaningful participation in their design or accountability. Wage dispersion follows, and it is already visible in Indian financial services, logistics, and the routine administration of public services as an ongoing structural shift.
When Governance Becomes Infrastructure
India’s experience with digital public infrastructure holds an instructive lesson here. Aadhaar, UPI, and the Open Network for Digital Commerce were acts of economic architecture, designed to shape market outcomes before private platforms could establish irreversible advantages. Public rails, governed under transparent rules, lowered transaction costs, extended access, and disciplined the monopolistic tendencies that unmediated platform economies invariably produce. Artificial intelligence demands the same institutional instinct at considerably greater complexity, since AI models evolve continuously, behave probabilistically, and resist audit by the static tools that sufficed for earlier digital systems.
The IndiaAI Mission’s governance guidelines represent a credible first move, addressing accountability, risk classification, and responsible deployment across sectors. Their reach, however, depends on a federal machinery whose capacity is deeply jagged. States with administrative resources and political will engage seriously; others implement centrally designed frameworks as best they can, with limited voice over their design or revenue implications. The country that built some of the world’s most inclusive digital plumbing has not yet assembled a governance architecture proportionate to the intelligence now flowing through it, a gap that markets are filling in the interim, with predictable consequences for fiscal equity and competitive concentration.
Jammu Kashmir: A Laboratory of First Principles
It is within this federal context that Jammu & Kashmir acquires its analytical interest, and why it deserves a place in any serious national AI policy conversation. The Union Territory’s 2023 Artificial Intelligence Framework addresses data exchange platforms, high-performance infrastructure, ethical protocols, privacy safeguards, and human capital development with a coherence that many larger states have not yet achieved. A Centre of Excellence in AI, GIS, and emerging technologies, now under development in partnership with BISAG-N, brings institutional specificity to goals of digital governance and sustainable regional development. These constitute a policy architecture assembled in a region whose distinctive constraints are fiscal dependence on Union transfers, geographic severity, and security imperatives that sharpen the value of deliberate design rather than compromise it.
The economic logic of AI-enabled public service delivery in J&K rests on prosaic grounds. Applications in land record management, grievance redressal, and predictive analytics for power procurement and water allocation reduce administrative friction across multiple governance layers. In a Union Territory structurally reliant on central grants, any durable improvement in fiscal precision and public sector productivity carries a multiplier that budget transfers alone cannot replicate. The spillovers radiate outward with stabilising local procurement markets, improving planning cycles, and giving administrators the informational basis for decisions that presently depend on incomplete and delayed data.
Shared computing infrastructure of publicly governed clusters accessible to universities, line departments, and nascent enterprises without negotiating proprietary arrangements addresses the structural inequality that market provision will not resolve voluntarily. The governance of such infrastructure proves as consequential as its capacity: pricing principles, data ownership rules, and model portability standards determine whether a compute commons genuinely distributes capability or merely subsidises familiar concentrations of advantage under a different institutional name.
In agriculture, where J&K’s mountainous terrain and meteorological volatility produce severe and recurring disruptions, AI systems trained on region-specific topographic and precipitation data could reduce income variance for farming households in ways that national models, calibrated for different ecologies, are structurally ill-equipped to achieve. Tourism presents a complementary case through real-time analytics on visitor flows and environmental carrying capacities, which can enforce sustainable limits without sacrificing the revenue on which the regional economy meaningfully depends, making AI a tool of ecological governance as much as of economic optimisation.
Human capital policy requires equivalent deliberation. The temptation is always to chase generic credentials; the actual need is to build interpretive capacity, the ability of frontline officials, agronomists, and health workers to contest algorithmic outputs, identify failure modes, and sustain accountability in systems that increasingly mediate access to land, welfare, and justice. Governance sandboxes, designating specific districts or administrative domains as controlled pilots under transparent audit and accessible grievance mechanisms, convert regulatory uncertainty into iterative institutional learning. Errors made at a small scale, with public accountability, are the raw material of durable policy.
The Choice the Summit Cannot Make
The summit closed, as summits do, with communiqués and commitments, the measured satisfaction of a country persuaded of its own global significance. The more consequential choices lie in the institutional details that no plenary session can resolve. Whether India constructs a governance architecture capable of distributing AI’s productive gains equitably, or permits them to consolidate in the familiar enclaves of capital and connectivity, will not be settled by any single declaration from any single podium.
Regions like Jammu & Kashmir do not merely wait upon that determination. They can prototype its resolution by demonstrating at a scale where institutional choices retain their legibility that AI governance can be made to serve economic resilience rather than simply accelerate efficiency for those already positioned to capture it. The summit is a declaration of intent. Whether that intent hardens into design will be visible, in time, in places far quieter than Bharat Mandapam and, in the long run of Indian economic history, considerably more instructive.
(The author is a pracademic working on government policy and public institutions. Ideas are personal.)
