AI for Viksit Bharat: The Opportunity for Accelerated Economic Growth (NITI Frontier Tech Hub roadmap, September 2025)

NITI Aayog published India's first national AI strategy in 2018,⁴ and its responsible-AI approach documents followed in 2021. In 2025 its Frontier Tech Hub, led by Distinguished Fellow Debjani Ghosh as chief architect, produced a growth-focused roadmap. In May 2025 the Hub had already held a workshop with Deloitte on accelerating investment in AI-ready data centres across states.³ The roadmap, 'AI for Viksit Bharat: The Opportunity for Accelerated Economic Growth', was published in September 2025 with McKinsey & Company as the analytical partner and guidance from an Expert Council of industry, finance and legal figures, including Tata Sons chairman N. Chandrasekaran and lawyer Rahul Matthan.¹

CEO B.V.R. Subrahmanyam's foreword says that India must raise growth to 8% a year to realise Viksit Bharat and that AI "can be the decisive lever". The report calls itself the "first version" of the perspective, to be revised periodically as the technology and the economy evolve.¹

The report starts from a GDP projection. At the current rate of about 5.7% growth India's GDP would reach USD 6.6 trillion by 2035; at 8%, it would reach USD 8.3 trillion, an extra USD 1.7 trillion. It identifies three AI levers to close the gap: accelerating AI adoption across industries (30-35% of the step-up), transforming R&D through generative AI (at least 20-30%), and innovation in technology services (15-20%, to be addressed in a separate publication).¹

Lever 1: adoption. A McKinsey-based model covering over 850 occupations in 16 sectors and more than 2,100 work activities estimated that accelerated adoption could add USD 500-600 billion to GDP by 2035 beyond the baseline. Financial services and manufacturing are the most affected, with up to 20-25% of sectoral GDP attributed to AI by 2035; the report puts the banking gain at USD 50-55 billion and manufacturing at USD 85-100 billion.¹

Lever 2: R&D. The report says AI lowers the capital barrier to innovation (for example USD 1-2 billion for a new drug). It estimates USD 280-475 billion of incremental GDP from AI-led leapfrog growth in 10 of 18 growth arenas, including AI software, cloud, software-assisted vehicles, semiconductors, medical devices, aerospace and defence, cybersecurity and biopharma, and discusses pharmaceuticals and auto-components design in detail.¹

The report presents four 'potential outcomes' and sector enablers, all framed as options "for consideration":¹

Data. India could become "the data capital of the world" by building an anonymised data collection framework led by entities such as an India Data Management Office, a marketplace of certified non-personal data with privacy tags and quality certificates supported by the National Data Governance Framework, and sector platforms: a financial-services data platform (borrowers keep an opt-out), an open 'Manufacturing Data Grid', a unified national omics dataset from sequencing over 10 million genomes by 2035, and anonymised vehicle telemetry. It notes that AIKosh, the government's platform under the IndiaAI Mission, hosted over 350 curated non-personal datasets and proposes integrating sector data infrastructure with it.

Skills. AI Chairs at the top 20 technology, medical, law and business schools; upskilling 3-5% of working professionals through AI modules backed by tax deductions on employer spending; an AI Open University; certifications such as 'AI for Credit, Risk and Fraud'; AI governance literacy for boards of regulated financial entities; and a new supervisory entity to build capacity in AI oversight, model risk and audits.

Sector adoption. Banking: scaled sandbox pilots for explainable credit, fraud and anti-money-laundering models; a cross-regulatory AI innovation sandbox; a shared GPU 'landing zone' for smaller regulated entities; institution-level AI inventories feeding a sector-wide repository; mandated fairness in AI outcomes, disclosures and grievance redress; regular red-teaming; risk-based audits with independent third-party audits for high-risk systems; and 'tolerant, good-faith' incident reporting. Manufacturing: AI-ready industrial parks, an open data grid and tiered credentials. Pharmaceuticals: expanded biotech parks, a national omics dataset and training 100,000 scientists in computational biology and AI. Automotive: six to eight autonomous-vehicle testing parks, 10,000 km of 5G or early-6G smart corridors and anonymised telemetry from 20-25% of new vehicles each year.

Jobs. Job-transformation maps for 25-30 priority sectors, portable individual learning accounts modelled on Singapore's SkillsFuture, universal registration and benefits for gig and platform workers under the Code on Social Security 2020 through the e-Shram portal, wage-loss insurance pilots, and a live skills and jobs database.¹

The report also observes that AI governance frameworks, including ethics guidelines and risk controls, are needed to ensure responsible deployment, and describes the seven pillars of the IndiaAI Mission, including the 38,000-plus GPUs, as the foundation on which its levers build.¹

MediaNama's report of 17 September 2025, headlined as NITI Aayog betting on AI for 8% growth 'without a national strategy in place', noted that the roadmap is not a Cabinet-approved policy and does not carry the weight of an official national strategy, and that NITI Aayog's 2018 strategy had seen little follow-up beyond the IndiaAI Mission.² Ruchi Gupta of the Future of India Foundation said it is "very industry-focused" compared with the more public-interest framing of 2018, attributing this to an industry-dominated Expert Council, argued that the jobs analysis was thin and should have adapted international models to India's large informal workforce, and called for a Cabinet-endorsed national AI policy. The article also noted that the report offers few answers on job displacement beyond citing Singapore's reskilling models, and recorded the observation that fewer than 1% of parliamentary questions between June 2019 and April 2025 addressed AI.²

Scope note: the roadmap is concerned with economic opportunity and sector enablers, and its governance content sits mainly in sector recommendations, such as the sandboxes, audits and incident reporting proposed for banking. The government's general framework for AI governance is the separate India AI Governance Guidelines of November 2025, whose drafting committee included NITI Aayog fellows such as Debjani Ghosh.¹⁶

As of 15 October 2026 the roadmap remains a NITI Aayog publication. The sources reviewed show no Cabinet decision adopting it and no scheme created to implement it, and they did not identify other 2025 NITI Frontier Tech Hub reports beyond the May 2025 data-centre workshop; the roadmap itself says a separate publication covers the third lever, technology services.¹³

Some of its governance ideas have echoes in later government steps. The Economic Survey 2025-26 proposed an AI Economic Council, and on 13 April 2026 MeitY set up an AI Governance and Economic Group that merges that idea with the AI Governance Group recommended in the Guidelines.⁵