Once empanelled, Jio aims to offer GPU-as-a-service to startups and researchers at the most competitive price in the world, a senior company executive told ET, detailing RIL’s overall strategy around next-generation technologies.
“There are three factors to accessible and affordable AI – the device, the data (on cloud), and high-speed broadband networks and Jio is working on all these so that every individual can have a personal AI agency based on his needs and likes,” the executive said on condition of anonymity.
Jio Platforms is working with Nvidia and others to develop AI use cases in areas like retail, healthcare, agriculture and education.
“The way Jio has democratised data usage, we want to do the same with AI,” the executive said, referring to how Reliance Jio brought mobile data prices in India crashing down to now being among the cheapest in the world.
Discover the stories of your interest
The Nvidia GPUs are also being used for Reliance industries’ captive use and building Jio-branded products and service on large language models (LLMs).Besides the AI Mission, Jio eyes monetisation opportunities with Jio Brain – an enterprise AI suite.
“The AI Mission is a government initiative to offer GPUs at a subsidised cost. For instance, assuming if the price is Rs 25 per GPU hour, the government will subsidise Rs 5-10 out of that. We (Jio and others) as vendors will compete to offer lower prices to Rs 25,” the executive explained.
The India AI Mission is a Rs 10,300-crore project to support responsible AI innovation in the country. Of that, Rs 5,000 crore has been earmarked for providing AI compute hardware by empanelling cloud providers, data centre companies, etc.
Jio Platforms and 18 other bidders have submitted bids to qualify for empanelment under the Mission in December, which are undergoing technical evaluation.
RIL is building a 1 GW AI data centre in Jamnagar both for captive use and offering GPU-as-a-service to commercial projects like the India AI Mission. As the usage scales, Jio will expand more AI inferencing facilities in its own premises across India, bringing AI closer to the edge, the executive said.
When asked about the high cost of AI infrastructure and running LLMs, the executive said, “Green energy will play a big role in bringing down operational costs of AI going forward. Global research has also suggested that there has been a significant decline in cost of AI hardware. India needs affordability. Without that, AI is just a fancy thing here… It will never scale up.”
For AI to be accessible, it must be available on low-cost devices through data hosted on the cloud, the person said.
Jio recently announced free cloud storage up to 100 GB for users to store photos, videos, documents and other digital content.
Internally, Reliance is working on embedding AI in corporate functions across its companies such as oil and gas refinery operations, telecom network management and retail store security through the Jio Brain suite.
Jio Brain is housed under Jio Platforms.
“We are creating end-to-end workflows with real-time data, which helps us in smarter and quicker decision making,” the executive said. “Think of it as automatic, very simple manual tasks. But at scale, they create so much value. Once Jio Brain as a platform is perfected within Reliance, we can offer it to other enterprises as well.”
