OpenAI's GPT-5.6 Terra and Luna Go Live on Amazon Bedrock in India

AWS announced that OpenAI's flagship GPT-5.6 Terra and Luna models are now available on Amazon Bedrock in India with in-country inferencing. The data-residency angle is the point: because inference runs on AWS infrastructure within India, organizations in regulated domains such as financial services can use frontier models without data leaving the country — a decisive factor for compliance-bound enterprises in a market with tightening data-localization expectations.

This extends the week's cross-region inference theme, in which AWS made the GPT-5.6 family (Sol, Terra, Luna) available across 25+ regions with routing profiles for higher throughput. India-specific in-country inferencing takes that a step further, treating data sovereignty as a first-class product feature — echoing, from a different angle, France's sovereign-AI procurement of Mistral.

Alongside the model availability, AWS committed more than $500 million to student cloud and AI training in the region and partnered with Novo Nordisk on an AI drug-discovery hub built on Amazon Bedrock AgentCore. Those are genuinely separate initiatives bundled into the same regional announcement; the common thread is AWS investing to entrench Bedrock as the default enterprise AI platform in a fast-growing market.

Strategically, localized frontier-model access is how the hyperscalers compete for sovereign and regulated workloads globally — the same playbook now running in Europe, the Middle East, and Asia. The caveat is that in-country inferencing can carry latency and cost tradeoffs versus centralized serving, and 'available in India' does not guarantee price parity. Watch whether Indian financial-services and public-sector customers adopt at scale, and whether Google and Azure match the in-country inferencing offer.