Anthropic has switched on in-country inference for its Claude AI models in India through Amazon Bedrock, meaning requests sent via the company's India endpoint are now processed entirely on servers located within the country.

The move, reported by The Economic Times and a string of Indian outlets on October 5, directly addresses the data residency requirements of regulated sectors such as banking, government and healthcare, letting enterprises keep sensitive AI workloads inside India's borders. Claude Opus 5, Sonnet 5 and Haiku 4.5 are all available with in-country inference through Amazon Bedrock. For more on the company behind Claude, see our AI industry coverage.

How In-Country Inference Works

The mechanics are straightforward but consequential. In simple terms, when an Indian customer uses the India endpoint, the prompt sent to Claude and the model's response are both processed on servers physically located in India. Amazon Bedrock, AWS's unified platform for accessing and deploying generative AI models, provides the infrastructure through which customers reach Claude locally.

That differs from the standard arrangement, in which inference traffic from Indian users is handled by data centers in other regions. It gives compliance teams something they previously lacked: an auditable guarantee that AI workloads — prompts, outputs and any attached documents — never leave the jurisdiction.

Why Data Residency Matters

For global AI providers, inference location has become a competitive issue. Many regulated industries — financial institutions, government bodies, healthcare providers — operate under rules that restrict where data can be processed and stored. Until now, Indian customers sending prompts to Claude had those requests handled on infrastructure outside the country.

"Requests sent to Claude through the India endpoint are processed on servers located within India," said Irina Ghose, managing director at Anthropic India, in an interview with the Times of India. "That is what our customers in India, especially in regulated sectors, asked us for: to use Claude on their most critical work with every request processed within the country."

Ghose added that financial institutions, government bodies and large enterprises had sought local inference because of regulatory and data-handling requirements, and that deployments which had been waiting on the capability could now move ahead.

Satinder Pal Singh, director of solution architects at AWS India and South Asia, said the controls that keep processing within the country are built into the service, along with the audit trails and access controls that risk and compliance teams require before putting AI into production.

India Is Anthropic's Second-Largest Market

The technical switch carries strategic weight because of how much Anthropic has riding on India. The country is the company's second-largest market for Claude.ai, and Indian users bring it an unusually technical workload: software development accounts for 45.2 percent of work-related tasks users in the country bring to Claude — the highest share of any market, according to Ghose.

Adoption over the past year has widened across banking and payments, large enterprises, IT services, digital companies, healthcare, science and public services. Anthropic's partnerships with Indian tech giants illustrate the scale:

  • TCS is equipping 50,000 associates with Claude.
  • Infosys is combining Claude with its Topaz platform.
  • Cognizant uses Claude in systems it builds for clients.
  • LTTS is embedding Claude in its engineering platforms.
  • Swiggy has made flagship services accessible through the Model Context Protocol, with Instamart letting customers browse and buy more than 40,000 products through prompts.

A Broader India Push

The launch is part of a wider expansion plan. Anthropic opened its Bengaluru office in February — its second in Asia after Tokyo — and plans to grow its India team and local partnerships.

The company is also investing in how well Claude understands Indian users. It is working with data firm Karya to improve Claude's training data across ten Indian languages: Hindi, Bengali, Marathi, Telugu, Tamil, Punjabi, Gujarati, Kannada, Malayalam and Urdu, with the aim of improving performance across Indian languages, domains and cultural contexts.

Community building has moved fast too. India recorded among the highest registrations when Anthropic launched its Claude Partner Network in March, and around 5,000 people were certified at the first Claude partner certification event in Bengaluru in August.

What It Means for the Market

In-country inference removes one of the last structural objections regulated Indian enterprises might raise against deploying Claude for critical work. It also aligns with a broader pattern in which AI providers localize infrastructure country by country — a playbook AWS knows well from its cloud business and is now extending to generative AI through Bedrock.

For India's AI ecosystem, the launch lands alongside a government push for digital sovereignty, giving enterprises a frontier-grade model that satisfies local processing requirements without waiting for domestic alternatives to mature. Ghose said Anthropic expects the momentum to grow rapidly from here.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →