Anthropic opens Bengaluru hub to boost India AI growth

Anthropic has opened an office in Bengaluru and announced partnerships across India spanning enterprises, education, agriculture, and the justice system.

India is Claude.ai's second-largest market. Nearly half of usage in the country involves computer and mathematical tasks, including building applications and shipping software.

Anthropic has appointed Irina Ghose as Managing Director of India. She will lead the Bengaluru office, the company's second site in Asia after Tokyo.

"India represents one of the world's most promising opportunities to bring the benefits of responsible AI to vastly more people and enterprises," Ghose said.

Language focus

Anthropic is focusing on how its models perform in Indian languages. More than a billion people in India speak one of more than a dozen officially recognised languages, while AI models often perform better in English.

Six months ago, the company began a company-wide effort to curate higher-quality training data in 10 widely spoken Indian languages: Hindi, Bengali, Marathi, Telugu, Tamil, Punjabi, Gujarati, Kannada, Malayalam, and Urdu. It said the work has improved model performance and that it continues to refine fluency.

Anthropic is working with Karya and the Collective Intelligence Project on evaluations to test model performance on locally relevant tasks across domains such as agriculture and law. The work includes domain experts from Indian non-profits, including Digital Green and Adalat AI.

Anthropic plans to make the evaluations publicly available and said the results will inform future model improvements for Indic language speakers and India-specific use cases.

Enterprise adoption

Anthropic said its run-rate revenue in India has doubled since it announced an expansion in October 2025, driven by a mix of large enterprises, digital-native companies, and startups building products with Claude.

The India team will provide applied AI expertise to enterprise customers, digital natives, and startups. Work will include designing, building, and scaling Claude-based products.

Several large organisations have started using Claude products for software development. Air India is using Claude Code for custom software development as part of a broader agentic AI effort. CRED reported 2x faster feature delivery and 10% better test coverage with Claude Code.

Cognizant is deploying Claude to 350,000 employees globally. The deployment covers legacy system modernisation, software development, and support for clients adopting AI.

Anthropic also highlighted startups using its tools. Razorpay has integrated AI into risk systems, decision-making processes, and operations. Enterpret uses Claude for an AI assistant, and its engineering team uses Claude Code daily. Enterpret has also shipped an MCP integration that brings customer insights into Claude.

Anthropic pointed to Emergent, an AI-powered platform for building software with plain language. It said Emergent reached USD $25 million in annual recurring revenue and two million users in under five months, and was built entirely with Claude.

Education and public services

Anthropic said educational and instructional tasks account for 12% of Claude.ai usage in India, and that it is working with Pratham, one of India's largest education non-profits.

Pratham's Anytime Testing Machine uses Claude and is being piloted with 1,500 students across 20 schools. Anthropic said there are plans to expand to 100 schools by the end of 2026. The programme has also been adapted for more than 5,000 learners in Pratham's Second Chance programme for women who have dropped out of formal schooling.

Anthropic is collaborating with Central Square Foundation on work focused on EdTech and AI in underserved communities. It will provide technical expertise, mentorship, and API credits to organisations developing AI-enabled tools, including personalised tutors, teacher coaching solutions, and assessment-driven instruction.

In agriculture, Anthropic is partnering with the EkStep Foundation through the OpenAgriNet effort, including deployments of Claude intended to expand access to expert knowledge.

Anthropic also cited work with non-profits including Noora Health and Intelehealth, which are exploring uses of Claude Code and Cowork.

In the justice system, Anthropic is supporting Adalat AI on a national WhatsApp helpline that launched today. The service uses Claude for case updates, translation, document summarisation, and interactive querying of legal documents in Indian languages.

Open standards

Anthropic also highlighted its Model Context Protocol (MCP), which it describes as a universal open-source standard for connecting AI applications to external systems. It has donated MCP to the Linux Foundation.

The Indian Ministry of Statistics and Programme Implementation has launched what Anthropic called the first official Indian government MCP server, with support from the non-profit Bharat Digital. Anthropic said the server lets AI systems access and query national statistics.

In the private sector, Swiggy uses MCP for grocery ordering and dining reservations through Claude, according to Anthropic.

Anthropic said the Bengaluru office will hire local talent across a range of roles as it expands its partnerships in India.