Quick answer
Economic Survey 2025-26 treats indigenous AI as a strategic priority; India has only 2% in Chart XIV.6’s distribution of training-data curation startups.
India’s own AI solution means domestic capability to develop, adapt and govern AI for Indian economic and social needs, not complete technological isolation. The Survey links it to dignified employment, critical-sector resilience, cultural representation and greater freedom in foreign-policy choices.
Concept diagram
Indian AI starts with curated Indian data
A curation funnel turns varied local inputs into a trusted basis for relevant AI systems.
- Raw data must be cleaned, labelled and evaluated before reliable use.
- Strategic autonomy is the ability to build and adapt, not simply to isolate.
Basics: data becomes useful through curation
Curation is productive work
- Data curation selects, cleans, labels, documents and checks datasets so they can support reliable model development. Raw data volume alone is not a usable training asset.
- LLMs trained on poorly representative language data can misread local contexts; evaluation must test performance on the intended population and task.
Training, adaptation and inference
- Training learns model parameters; fine-tuning adapts a trained model to additional examples; inference applies the resulting system.
- Domestic capability can be built across this chain without training every model from scratch or owning every GPU design.
Strategic autonomy
- Autonomy is the ability to preserve critical choices and manage external disruption. It differs from autarky, which seeks economic isolation.
- Foreign chips and models can remain useful while diversified access, domestic expertise and interoperable systems reduce dependence.
Data chart
Chart XIV.6: Distribution of training-data curation startups
- The United States accounts for 40% in the printed distribution, compared with India’s 2%.
- India’s diverse raw data does not automatically imply a scaled data-curation industry.
Source: Economic Survey 2025-26, Chart XIV.6 (World Bank) · Reproduces the categories printed in the Survey chart; they total 94%. No residual category is invented. This is not a ranking of all AI startups.
Show the data (Per cent)
| Period / item | Share |
|---|---|
| United States | 40 |
| European Union | 21 |
| Other high-income | 17 |
| United Kingdom | 9 |
| China | 5 |
| India | 2 |
Why domestic capability matters — paras 14.28–14.30
| Channel | Development opportunity | Dependence risk |
|---|---|---|
| Employment | Move beyond a low-cost labour proposition | Routine outsourced work may lose competitiveness |
| Critical sectors | Adapt AI for services, manufacturing, defence and power | External restrictions can disrupt essential functions |
| Foreign policy | Preserve room for diplomatic choices | AI inputs can become bargaining instruments |
| Culture and context | Reflect domestic languages and institutions | Imported systems may embed unsuitable assumptions |
| Domestic data | Build useful sector-specific datasets | Value can accrue elsewhere if capability is absent |
From domestic data to domestic value
Curate responsibly
Build usable, documented and representative datasets with appropriate permissions and safeguards.
Develop and validate
Combine domain knowledge, suitable models and Indian-language or sector-specific evaluation.
Retain capability
Grow domestic expertise and enterprises while maintaining access to global innovation.
Common exam traps
Statements examiners like to twist — learn the correct version.
India’s own solution means every input must be indigenous.
The Survey rejects complete self-sufficiency as infeasible and inefficient; it advocates autonomy with integration.
India’s 2% refers to all AI firms or global AI usage.
Chart XIV.6 concerns startups focused on curating training data.
More raw data automatically produces a strong model.
Representativeness, curation, lawful access, compute and evaluation determine whether data becomes useful.
Prelims focus
- Chart XIV.6: United States 40%, EU 21%, India 2%.
- Distinguish strategic autonomy, autarky, data curation and inference.
Mains angle
- Explain AI as a strategic priority connecting employment, critical infrastructure, culture and diplomacy.
- Discuss how data diversity can become a comparative advantage only through domestic capability building.
Flip & recall
Answer in your head first, then tap the card to check.
Practice MCQs
Original Prelims-style questions with explanations.
0 / 3 correct · 3 left
Q1.Chart XIV.6 concerns:
Q2.Which best describes strategic autonomy?
Q3.Which converts raw data into a more usable training resource?
