Reliance held its 49th annual shareholder meeting today and used it to announce what amounts to a full-stack push into AI for everyday Indian consumers. The headline product is Jio Call Agent, an AI assistant embedded directly into Jio's phone network. Users say "Hey Jio" during a live call, and the assistant joins to transcribe, summarize, and carry out tasks like booking transport or food. Because it runs inside the network itself rather than as a separate download, it reaches anyone on Jio's network automatically.
That is the part worth paying attention to. Jio has over 500 million subscribers. Embedding AI into the call layer, rather than offering a standalone app that users have to find and install, is a distribution play that almost no other company in the world can replicate at that scale. The closest comparison is what Amazon does with Alexa through its devices, or what Google does with Android. Jio is attempting the same thing through the phone call itself.
Reliance also announced TeleFrame, a home display that acts as a central hub for several AI agents managing schedules, household tasks, entertainment, and shopping. It learns family routines and acts without being asked. The company unveiled an upgraded MyJio app that handles account management and plan changes through plain conversation. And it launched five new AI apps covering healthcare, education, farming, small business, and general queries, all operating in 22 Indian languages natively rather than translated from English.
The language point is not a small detail. India has hundreds of millions of people who are not comfortable in English, and most AI services from US companies treat Indian languages as an afterthought. Building natively in Hindi, Tamil, Bengali, and the other major languages is a real product advantage in this market.
This announcement lands in a very specific moment. Just last week, the US government ordered Anthropic to cut off access to two of its newest AI models for all non-US citizens, effective immediately. Indian startups that had built products on those models woke up one Saturday to find their tools broken. Anthropic disputed the directive, but the damage to confidence was done. India's technology leaders, from Zoho's Sridhar Vembu to former Infosys executive Mohandas Pai, have since called loudly for India to stop renting its AI from a small group of American companies.
Reliance is positioning itself as the answer to that call. In February, Ambani pledged $110 billion over seven years for AI computing infrastructure in India. Construction of a large data center in Jamnagar, Gujarat, is already underway, with the first 120 megawatts of computing capacity expected to come online by the end of this year. The facility runs on renewable energy from Reliance's own solar operations. Rival Adani Group has pledged a comparable $100 billion for similar infrastructure, and the Indian government expects over $200 billion in total AI infrastructure spending across the country in the next two years.
On top of the product announcements, Reliance confirmed today that Jio Platforms' board has approved a draft prospectus for a stock market listing. This has been anticipated for years. Jio crossed 500 million subscribers and analysts estimate its valuation at roughly $130 to $180 billion, which would make it one of the largest listings in Indian history. In March, Reliance shifted to a 100 percent fresh share issue structure for the IPO, meaning all proceeds go directly into the business rather than to existing shareholders cashing out.
For business operators outside India, the relevant signal here is the supply chain risk that the Anthropic episode exposed. Any business building products or internal tools on top of a single foreign AI provider faces the same vulnerability: one government directive, and access disappears. Diversifying across multiple providers, or toward open-source models that can be run independently, is no longer a theoretical precaution. It is basic operational risk management.
Reliance's data privacy practices on these new services remain an open question. The company says all services require user consent, but has not publicly addressed whether data from calls, home devices, and apps can be used to train its models or shared with technology partners including Google and Meta.