Published: February 2026
Category: Artificial Intelligence and Machine Learning · Tech Industry Developments · Government Policies and Cyber Regulations
India is taking a decisive step towards digital sovereignty as homegrown startup Sarvam AI partners with the governments of Odisha and Tamil Nadu to build large-scale, state-backed artificial intelligence infrastructure.
The collaborations aim to establish what Sarvam AI describes as “Sovereign AI” — compute capacity, language models and AI systems designed, hosted and governed within India. Together, the projects form the early foundation of a national compute grid, positioning AI as a core layer of India’s digital public infrastructure.
Odisha to host 50MW national AI compute hub
On 6 February, the Government of Odisha signed a memorandum of understanding with Sarvam AI to develop a 50-megawatt AI-optimised data centre, one of the largest such facilities announced by an Indian state to date.
The facility is designed to serve a dual role. At the state level, it will support public-sector applications such as Vision AI for mining and industrial safety, alongside Odia-to-English voice tools aimed at improving employability and skills access for young people.
At the national level, the Odisha hub is intended to anchor a production-grade compute backbone that other states and central platforms can tap into, reducing reliance on foreign hyperscalers for critical AI workloads.
Odisha’s Chief Secretary Anu Garg described the initiative as a shift from a “mine-driven” to a “mind-driven” economy, aligning the project with India’s long-term development roadmap under Viksit Bharat 2047.
Tamil Nadu launches India’s first Sovereign AI Research Park
In parallel, Tamil Nadu has announced Digital Sangam, billed as India’s first Sovereign AI Research Park, developed in collaboration with Sarvam AI and IIT Madras.
The park will house a 20MW AI-optimised data centre, alongside facilities for research, startup incubation and applied AI development. The goal is to create a physical ecosystem where compute, talent and deployment converge in one location.
Flagship projects under Digital Sangam include Vivasāya Nanban, an AI assistant designed to provide round-the-clock advisory services to nearly 8 million farming households, and a Unified Citizen Helpline that uses conversational AI to simplify access to welfare schemes and public services.
Tamil Nadu’s Industries Minister Dr TRB Rajaa emphasised the importance of voice-first AI, particularly for rural and tribal communities, noting that population-scale conversational systems can dramatically improve access to information and grievance redressal.
A broader push for AI sovereignty
At the heart of both initiatives is the concept of AI sovereignty — the idea that a nation’s data, models and intelligence infrastructure should remain under domestic control. Sarvam AI says building models for Indian languages and contexts ensures AI systems reflect local realities rather than imported assumptions.
“The choices made now about AI infrastructure will determine whether intelligence compounds within the country or accrues elsewhere,” the company said in a statement.
By embedding AI directly into governance and public services, India is attempting to create a reusable AI layer within its Digital Public Infrastructure, allowing intelligence to be shared across departments while keeping sensitive data within national boundaries.
Laying the groundwork for India’s AI future
As artificial intelligence shifts from experimentation to essential infrastructure, the Sarvam AI partnerships signal a growing willingness by Indian states to invest directly in compute, models and institutional capacity.
If successful, the Odisha and Tamil Nadu projects could serve as templates for other regions — accelerating AI adoption while reinforcing national control over one of the most strategic technologies of the decade.
Author. Adigun Adedoye.