Anthropic Economic Index: India Country Brief
TL;DR
India is the second-largest user base for Claude.ai globally, characterized by a high concentration of use within the professional IT sector and a tendency to delegate more autonomy to AI for complex tasks. While absolute usage is high, per-capita adoption remains low, indicating a significant opportunity for expansion beyond the current technology-centric hubs.
Global Adoption and Per-Capita Disparity
India ranks second globally in the share of total Claude.ai use, accounting for 5.8% of total conversations observed in November 2025. However, this high ranking is a result of India's large population rather than broad adoption. When adjusted for the working-age population, India ranks 101st out of 116 countries, trailing other Asian nations such as Singapore and Malaysia.
Geographic and Occupational Concentration
AI adoption in India is currently concentrated among the established technology workforce and specific economic hubs:
Geographic Concentration
Over half of India's Claude.ai use is concentrated in four states: Maharashtra (15.5%), Tamil Nadu (13.2%), Karnataka (12.7%), and Delhi (10.5%). This distribution mirrors the geography of India's IT sector and urban economic output, specifically centering on cities like Bangalore, Hyderabad, Chennai, Mumbai, and Delhi NCR.
Occupational Concentration
India ranks first globally in the share of AI use devoted to software-related tasks, with 45.2% of all O*NET-mapped tasks falling into this category. This exceeds the shares seen in Vietnam (42.1%) and Egypt (39.2%). While software development and engineering dominate, educational tasks also represent a significant portion of use cases.
Economic Primitives: Distinctive Usage Patterns
Analysis of "economic primitives"—metrics measuring human-AI collaboration—reveals that Indian users utilize AI differently than the global average:
- Productivity Speedup: Indian users experience a 15x speedup on tasks, reducing work that would take 3.8 hours (without AI) to approximately 14.8 minutes. This exceeds the global average speedup of 12x (3.1 hours reduced to 15.4 minutes).
- Work Orientation: 51.3% of Indian use is work-related, compared to 46% globally. Personal use is lower in India (27.8%) than the global average (34.7%).
- AI Autonomy: Indian users delegate more decision-making autonomy to AI, scoring 3.60 on a 1-5 scale compared to the global average of 3.38.
- Human-Only Ability: Indian users more frequently use AI for tasks they could not complete alone. Only 84.6% of tasks were completed by humans alone, compared to 87.9% globally.
- Prompt Sophistication: India ranks in the top 10% globally for the education level of AI responses, reflecting a correlation between high-quality user prompts and sophisticated outputs.
Strategic Implications for AI Growth
Expanding Beyond IT Services
Because 45.2% of tasks are software-related and use is concentrated in four states, broadening AI's economic impact requires moving beyond the existing IT services framework and professional strengths.
Productivity Gains as a Value Proposition
The 15x speedup on complex tasks demonstrates that India is extracting significant value from AI by compressing the time required for harder tasks more effectively than the global average.
Addressing Structural Barriers
The gap between India's 2nd place ranking in total use and 101st place in per-capita use is consistent with global correlations between per-capita AI adoption and per-capita income. Increasing adoption will require addressing barriers related to income, digital infrastructure, and awareness outside the IT sector.
The Role of AI Skills
Given the global correlation between prompt sophistication and response quality, targeted training programs for workers outside the current IT-heavy user base could significantly increase the returns on wider AI adoption.
Methodology
This analysis is based on privacy-preserving data from Claude.ai consumer use (Free, Pro, and Max) from November 13–20, 2025. Geographic assignment was conducted via IP-based geolocation, and occupational classification used O*NET task taxonomy and SOC occupation groups. Only countries with at least 200 observations were included in the rankings to ensure measurement certainty.
Sources
Related
- Dispatch
- Dispatch
- Dispatch
- Dispatch
- Dispatch