dario amodei: New index shows India leads total Claude.ai use but lags per-capita
India ranks second worldwide in total use of Claude. ai, but a closer read of a new economic index brief reveals that adoption is heavily concentrated among a small set of tech-forward states and occupations. The dataset, covering roughly 1 million Claude. ai conversations from November 2025 (Eastern Time), highlights striking productivity boosts for Indian users alongside clear opportunities to broaden access beyond established IT hubs.
Concentration in a few states and software roles
The index finds India accounts for 5. 8% of global Claude. ai use, trailing only one other market by total volume. That headline figure, however, masks a lopsided geography: Maharashtra, Tamil Nadu, Karnataka and Delhi together generate more than half of the country’s Claude. ai activity. Those four states house the country’s largest technology centres and urban economic nodes, suggesting current uptake is driven primarily by established tech workforces rather than mass consumer adoption.
Occupation mapping in the brief reinforces this pattern. Nearly half of India’s AI-mapped tasks — 45. 2% — are software-related, the highest share globally. The task mix skews heavily toward software development and engineering functions, while educational and coursework tasks also form a visible share of activity. This occupational concentration aligns with India’s position as the world’s largest exporter of IT services and points to specialized, high-value use rather than broad-based daily use by the general population.
Big productivity gains, frontier use
One of the most notable findings is the scale of productivity speedups Indian users are realizing. On average, tasks that would take 3. 8 hours without AI are completed in 14. 8 minutes with Claude. ai — roughly a 15x speedup. By contrast, the global average in the index shows a 12x speedup for comparable tasks. This gap reflects the more complex, time-consuming problems Indian users bring to the assistant and indicates that many Indian users are operating at the frontier of human–AI collaboration.
The brief introduces a framework of “economic primitives” — core measurements that capture how humans and AI collaborate. Within this framework, India stands out for two connected patterns: a stronger orientation toward work use and a higher share of tasks that would be difficult or impossible to complete without AI. More than half (51. 3%) of Claude. ai interactions from India are work-related, compared with 46% globally. Coursework and learning-related interactions also register above the global average, reinforcing the dual role of Claude. ai as a productivity and instructional tool in India’s economy.
Per-capita gap and policy implications
Despite ranking second by total usage, India’s per-capita metrics tell a different story. When adjusted for working-age population, India falls near the bottom of the index sample, ranking 101st out of 116 countries with sufficient data. That disparity signals that high aggregate volume largely reflects population scale rather than widespread individual adoption. It also exposes a sizable upside: broadening access beyond the concentrated IT workforce could unlock considerable economic and social benefits.
Targeted interventions could focus on expanding access outside metropolitan tech hubs, strengthening digital skills among non-IT workers, and tailoring products to local languages and everyday workflows. The concentration of use in a handful of states implies that infrastructure and training investments in secondary cities and rural areas could materially shift the per-capita picture over time.
In sum, the index paints India as a market where AI is already delivering outsized gains for a relatively narrow slice of users. The combination of strong productivity effects and concentrated adoption creates both an immediate payoff for existing skilled workers and a clear potential for growth if access and capabilities expand more broadly across the population.
