Anthropic Brings Claude Inference To India As Sovereign

What happened

Anthropic is bringing in-country inference for its AI platform Claude to India via Amazon Bedrock, AWS’ managed service for building GenAI applications, removing one of the biggest barriers that has prevented banks, financial institutions and government agencies from deploying international frontier AI models in production.

IT services giants TCS and Infosys are embedding Claude into enterprise AI offerings, with TCS equipping 50,000 associates with the model and Infosys integrating it with its Topaz platform while setting up an Anthropic Centre of Excellence. The announcement comes months after Anthropic opened its first India office in Bengaluru and appointed Ghose to lead its local operations, enterprise partnerships and hiring efforts.

2% of Claude’s global usage, making it Anthropic’s second-largest market after the US. As part of its localisation strategy, Anthropic also introduced rupee-denominated pricing for Claude subscriptions in India earlier in July, allowing users to pay in INR instead of USD as it deepens its presence in the country.

Why it matters

The capability, which will be available in the coming weeks, marks the latest step in the US-based AI company’s India expansion and reflects a broader shift in how global AI players are treading the Indian market. For context, in-country inference allows AI models to process user prompts and generate responses on servers physically located within India.

This ensures enterprise data remains within national borders while reducing latency and helping organisations comply with data residency requirements. This capability is emerging as a critical enabler for enterprise AI adoption in India. Sectoral regulations and India’s data protection regime require keeping sensitive information within national borders, prompting several banks and public agencies to evaluate frontier AI models without deploying them in production.

Anthropic is not the first AI company to pursue this strategy. 5 Flash for enterprise customers in India. Homegrown foundation model startup Sarvam AI has also built its sovereign AI strategy around locally hosted inference. Claude deployments through Amazon Bedrock also bring enterprise-grade audit trails, access controls and governance capabilities that risk and compliance teams typically require before deploying AI applications.

Anthropic also used the announcement to showcase Claude’s growing adoption across India Inc. Financial institutions including Axis Bank and IndusInd Bank are deploying the model for engineering, productivity and enterprise knowledge management, while NPCI is building its agentic AI platform, AiNxt, using Claude.

What to watch

Global AI companies are increasingly optimising their offerings for India, where adoption of GenAI has accelerated across enterprises, startups, developers, and students. This is also raising the competitive stakes for Indian foundation model startups. For companies such as Sarvam AI, locally hosted inference has been one of the strongest differentiators against global frontier models.

As Anthropic and Google begin offering similar capabilities, domestic AI labs may increasingly have to compete on model performance, language capabilities, pricing, enterprise integrations and domain-specific applications rather than infrastructure alone.