India Bets $100 Billion on AI as Global Tech Leaders Converge at Summit

Anthropic has formally launched its first India office in Bengaluru, timed alongside the country's AI Impact Summit, after revealing that India has grown into the second-largest market for its Claude AI platform globally — behind only the United States. Run-rate revenue in India has doubled since October 2025, when the company first announced its expansion plans.

Nearly half of Claude's usage in India is technical in nature — coding, mathematical problem-solving, application development, and modernizing legacy software systems — reflecting a developer-heavy user base. Anthropic's Bengaluru hub is its second office in Asia, following Tokyo, and signals the company's intent to build local presence in markets where adoption is accelerating fastest.

The opening came with a slate of enterprise and education partnerships. Air India is using Claude Code to accelerate software development; CRED and Cognizant are deploying Claude tools for engineering productivity and legacy system modernization; and education nonprofit Pratham is piloting an AI-powered testing tool with 1,500 students across 20 schools, with plans to expand to 100 schools by year's end. Multilingual support for India's officially recognized languages was highlighted as a core development priority.

Neuromorphic computers — chips designed to mimic the architecture of the human brain — have demonstrated an unexpected ability to solve partial differential equations (PDEs), the complex math underlying physics simulations of fluid dynamics, electromagnetic fields, and structural mechanics. The finding, published in Nature Machine Intelligence by Sandia National Laboratories researchers, challenges the long-held assumption that such computations require energy-hungry conventional supercomputers.

The team ran their NeuroFEM algorithm on Intel's Loihi 2 neuromorphic chip and found that doubling the number of processor cores nearly halved the time needed to solve problems — while consuming significantly less energy than a standard CPU running identical calculations. The chip's parallel, event-driven design proved naturally suited for the iterative solvers that PDEs demand.

The implications extend well beyond efficiency metrics. Sandia sees a credible path toward the first neuromorphic supercomputer — one capable of running large-scale simulations at a fraction of the energy cost of today's systems. Practical applications in the near term include national security workloads managed by the National Nuclear Security Administration, with broader potential across scientific computing fields ranging from climate modeling to aerospace engineering.