NITI Aayog's 'AI for Viksit Bharat' Report Says AI could add $1Trn to India’s GDP by 2035

NITI Aayog’s recent report on AI, AI for Viksit Bharat, argues that artificial intelligence can be the decisive lever to help India reach an aspirational “Viksit Bharat” growth path (roughly 8% annual GDP growth).

The report says this can only happen if the government and industry move fast on practical, sector-focused enablers. It lays out three channels for value, faster adoption across industries, AI-led R&D and a technology-services boom, and finds that the first two alone could add between $500–600 billion and $280–475 billion respectively by 2035.

Combined, that implies an incremental GDP uplift of about $780 billion to $1.075 trillion if adoption and policy assumptions hold.

Report Suggestions

The report narrows its focus to two near-term levers: accelerating adoption across priority sectors to lift productivity, and transforming R&D through generative AI to unlock leapfrog innovation.

It stresses that these must be delivered alongside four strategic enablers, infrastructure (compute and cloud), governance (ethics and model-risk controls), industry mechanisms (sandboxes and marketplaces) and workforce development (skilling and academic chairs), and recommends a phased, KPI-tracked rollout timed to India’s 2035 goals.

Central to the plan is scaling the IndiaAI Mission and an expanded AI Kosh: a federated GPU pool (the report cites about 38,000 GPUs as a benchmark), India-tuned large-language models and a consented, anonymised national dataset platform.

The authors recommend transforming AI Kosh into a trusted national data marketplace and building sector data platforms, a Manufacturing Data Grid, financial cross-industry alternative data, automotive telemetry sharing and a national omics dataset targeting sequencing of about 10 million genomes by 2035, to make India a global “data capital”.

Sector Priorities & Impact

The report highlights four priority sectors where measurable gains can be captured quickly: manufacturing, financial services, pharmaceuticals and automotive.

In banking it forecasts roughly $50–55 billion incremental value by 2035 from “bionic” deployments (virtual RMs, explainable underwriting, automated compliance and fraud controls), while manufacturing could deliver $85–100 billion through predictive maintenance, digital twins and intelligent quality control.

For R&D-led leapfrogs, pharma and automotive are singled out: AI-driven drug design and virtual trials could unlock tens of billions in licensing value, while software-assisted vehicles and deep-learning design surrogates could deliver import substitution and exports worth $20–25 billion for autos if telemetry sharing and 5G corridors are in place.

The roadmap recommends an aggressive skilling stack: AI chairs at elite institutes, an AI Open University, industry incentives to upskill 3–5% of the workforce, and sector credentials (for example “AI for Financial Services”). It also calls for governance-literacy programmes and the creation of supervisory capacity for model risk, audits and certification to ensure safe deployment at scale.

Implementation Mechanics

To move from pilots to scale the authors propose industry sandboxes for explainable credit and fraud models, AI-ready industrial parks with shared HPC/edge compute, biotech parks with national omics and fast-track regulatory pathways, and physical-digital test corridors (10,000 km of 5G for automotive SAV testing).

These market-creation measures are paired with recommendations to prioritise inferencing-efficient hardware in future GPU procurements to relieve near-term supply crunches.

The report flags major constraints: legacy IT and data fragmentation in enterprises, privacy and consent fatigue, talent shortages, patent and IP bottlenecks in life sciences, chip shortages for autos and exposure to cyber risk as digital-physical systems scale.