Sarvam AI Selected to Build India´s First Sovereign Large Language Model

Sarvam AI will develop India’s first sovereign large language model, integrating advanced reasoning and voice support for multiple Indian languages under the IndiaAI Mission.

Bengaluru-based Sarvam AI has been selected by the Government of India, as part of the IndiaAI Mission, to develop the country’s inaugural sovereign large language model (LLM). This initiative will supply Sarvam AI with dedicated computing resources to craft a foundational model from the ground up, focusing on reasoning and voice capabilities, as well as fluency across various Indian languages.

The sovereign LLM project is structured to ensure that all development, deployment, and optimization occur entirely within India, using local infrastructure and emphasizing the cultivation of domestic talent. The new model is intended for secure, large-scale implementation across a broad spectrum of applications, reflecting a significant strategic push to secure India’s technological autonomy in the fast-evolving language technology landscape.

This project aligns with the overarching goals of the IndiaAI Mission, which aims to enhance strategic autonomy and foster innovation within the domestic artificial intelligence sector. Sarvam AI has demonstrated expertise in creating foundational models tailored for Indian linguistic and contextual needs—initially as a research endeavor, but now as a full-fledged sovereign Artificial Intelligence platform serving governments, nonprofits, and businesses. Their solutions are recognized for being cost-effective, scalable, and optimized for Indian conditions, positioning the company as a significant player in the nation’s emerging artificial intelligence ecosystem.

81

Impact Score

HMS researchers design Artificial Intelligence tool to quicken drug discovery

Harvard Medical School researchers unveiled PDGrapher, an Artificial Intelligence tool that identifies gene target combinations to reverse disease states up to 25 times faster than current methods. The Nature-published study outlines a shift from single-target screening to multi-gene intervention design.

How hackers poison Artificial Intelligence business tools and defences

Researchers report attackers are now planting hidden prompts in emails to hijack enterprise Artificial Intelligence tools and even tamper with Artificial Intelligence-powered security features. With most organisations adopting Artificial Intelligence, email must be treated as an execution environment with stricter controls.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.