Chipmaker NVIDIA announced on Friday that it is partnering with Reliance Industries to develop India's own large language model trained on local languages and tailored for generative Al applications.

According to a statement from NVIDIA, the companies will work together to build Al infrastructure that is “more powerful than the fastest supercomputer in India today”.

“NVIDIA will provide access to the most advanced NVIDIA GH200 Grace Hopper Superchip and NVIDIA DGX Cloud, an Al supercomputing service in the cloud. GH200 marks a fundamental shift in computing architecture that provides exceptional performance and massive memory bandwidth,” the statement read.

The Al infrastructure is said to be hosted in AI-ready computing data centres that are expected to expand their capacity to 2,000 MW. NVIDIA said that Jio will be managing the execution and implementation, with its extensive offerings and experience across mobile telephony, 5G spectrum, fibre networks among others.

“Jio has broad expertise, infrastructure and engineering skills to roll out and manage the new Al computing infrastructure. The collaboration with NVIDIA also aligns with its strategy of serving as a large, comprehensive digital, cloud and networking platform for both consumers and business customers,” NVIDIA added.

Through the collaboration, Reliance will reportedly create Al applications and services for 450 million Jio customers and provide energy-efficient Al infrastructure to scientists, developers, and startups, across India. The NVIDIA-powered Al infrastructure will serve as the foundation of Al for Jio.

During the 46th Annual General Meeting of Reliance Indutries last month, Mukesh Ambani, Chairman & Managing Director, Reliance Industries, announced that Jio will focus on "India-specific artificial intelligence (AI) models and AI-powered solutions across domains", promising AI to everyone, everywhere.

First Published:Sept 8, 2023 3:28 PM IST