Anthropic has expanded its Claude AI presence in India by introducing in-country inference through Amazon Bedrock, allowing requests sent through an India-based endpoint to be processed on servers located within the country. Most notably, it is targeting businesses and organisations which have strict data residency, security and compliance requirements.
With the new setup, Indian organizations can access Anthropic's Claude models through Amazon Bedrock and process their prompts and AI-generated responses within India. The development could be particularly significant for financial institutions, government organisations, insurers and large enterprises that may have restrictions around where sensitive information can be processed.
Anthropic has made Claude Opus 5, Claude Sonnet 5 and Claude Haiku 4.5 available in India through Amazon Bedrock. The company has also highlighted the involvement of major Indian businesses and technology partners in validating the capability for enterprise use.
What Anthropic's In-Country Inference Means
In-country inference refers to processing AI requests within a specific geographical jurisdiction rather than sending data to infrastructure located in another country. For companies dealing with sensitive information, the location where AI processing takes place can play a significant role in their compliance and risk-management strategy.
With Anthropic's India endpoint, requests are sent to Claude on servers based in India. This enables organisations to use Claude for business applications while still having control over where their data is being distributed.
The capability is particularly relevant as Indian companies increasingly integrate generative AI into production systems. Although organisations have been experimenting with AI assistants and models for some time, privacy, data governance and regulatory compliance can prevent highly sensitive workloads from moving beyond the pilot stage.
By providing local inference, Anthropic is attempting to remove one of those barriers.
Amazon Bedrock Provides Infrastructure Across Mumbai And Hyderabad
Amazon Bedrock is playing a major role in Anthropic's India deployment. The service employs AWS infrastructure in Mumbai and Hyderabad to process Claude requests while keeping the relevant workloads within India's geographical boundaries.
The infrastructure is designed to distribute workloads between the two Indian AWS regions. This can lead to high performance and availability in times of high demand without sending requests overseas.
For enterprise customers, the infrastructure also has security and governance features like access controls, audit trails and encryption. These are important for organisations whose compliance and security staff would need visibility into the use of AI systems.
The ability to operate in multiple Indian regions can also help businesses be more resilient while still meeting the data-residency requirement.
Data Residency Could Accelerate Enterprise AI Adoption
The biggest impact on regulated industries would be the move. Banks, insurers, government departments and large corporations often have sensitive customer, financial or operational information.
Such organisations would need to comply with specific data-handling requirements before allowing information to be processed by third-party AI systems. Keeping inference in India may help companies to evaluate Claude for workloads that involve sensitive information, but individual organisations will have to determine their own legal and regulatory obligations.
The development also comes as India enhances its data protection and responsible use of artificial intelligence. The Digital Personal Data Protection framework has raised the bar for how organisations collect, process and protect personal information.
In-country inference does not automatically make every AI deployment compliant with every legal requirement. But local processing can address an important part of an organisation's data-governance framework.
Reliance, CRED And Godrej Industries Among Early Users
Anthropic says that several Indian enterprises and partners have already helped validate its in-country inference capability. Reliance, CRED, and Godrej Industries are among the companies testing the local deployment.
Godrej Industries has highlighted the combination of regulatory alignment, performance and resilience as important factors in moving Claude into production workflows.
TCS, which has a significant enterprise AI presence, has also welcomed the development. The company has highlighted high demand for enterprise AI deployments in regulated industries and said local inference can help in addressing evolving data residency and governance needs.
These early enterprise relationships could be key for Anthropic to expand Claude beyond individual developers and technology enthusiasts and into large corporate workflows.
Claude's India Expansion Gains Momentum
Anthropic's decision to introduce in-country inference follows its wider expansion in the Indian market. In August, the company announced local Claude inference in the context of strong demand from Indian organisations.
The company has mentioned India as one of Claude's largest markets globally and has also pointed to the country's large and technically sophisticated developer community.
Anthropic opened its first India office in Bengaluru in February 2026 after market expansion in 2025. In the future, the company will continue expanding its Indian team and work on local partnerships.
This new local inference step supports that strategy and addresses one of the biggest requirements for enterprise AI adoption in the country.
Privacy And Zero Data Retention
Amazon Bedrock's security architecture is another important part of the deployment. Data transferred between the Mumbai and Hyderabad AWS regions is protected through encryption over AWS's private network.
The setup also operates under a Zero Data Retention policy for the relevant inference service, meaning customer prompts and generated responses are not intended to be permanently stored in the destination regions.
But organisations should still check the exact terms, security configuration and applicable policies before using AI systems for sensitive workloads. Data residency is only one part of a broader security and compliance framework.
Businesses also need to consider access permissions, employee usage, authentication, audit requirements and the nature of the information being submitted to an AI model.
A More Competitive Indian AI Market
Anthropic’s move comes as competition among large AI companies is increasing in India. Local data-processing capabilities have become increasingly important as businesses move from experimenting with generative AI to deploying models in real-world applications.
For Anthropic, offering Claude through Indian infrastructure could help it compete more directly for enterprise workloads that require stronger governance controls. It also gives Indian businesses another option when evaluating AI models for regulated or sensitive applications.
The introduction of in-country inference does not guarantee that every organisation will immediately move its AI workloads to Claude. Companies will still need to assess model performance, pricing, security, compliance and integration requirements. However, processing requests in India removes a critical infrastructure barrier. As more Indian companies shift from AI to customer service, software development, research, analytics and internal operations, the ability to keep sensitive processing within national borders becomes an increasingly important factor in deciding which AI platforms make it into production.
