Epic scale goes digital as Mahabharat gets an AI-powered global retelling

MUMBAI: When ancient epics meet modern code, even the Mahabharat can travel faster than time. At Inter BEE 2025 in Japan, one of the country’s most influential broadcast and technology exhibitions, Microsoft offered audiences a glimpse of how Indian storytelling is being re-engineered for the global stage with artificial intelligence doing the heavy lifting.

During a Microsoft keynote on November 19, the tech giant unveiled the first look trailer of Mahabharat: Ek Dharmayudh, an AI-powered cultural project by Collective Artists Network, created in collaboration with its AI division Galleri5. Built using Microsoft Azure AI, the project blends creative IP development with advanced AI-driven production workflows.

The showcase went beyond visual spectacle. Galleri5’s AI systems were deployed across world-building, visual development and pre-visualisation, while Collective anchored the project with detailed research, writing and cultural direction ensuring the epic’s narrative depth survived its technological makeover.

A standout moment came when Microsoft presented a Japanese-language trailer, dubbed from Hindi using Microsoft’s AI tools. The move marked one of the earliest instances of an Indian production house using AI to enable multi-language, cross-border distribution at scale, signalling how local mythology can now be packaged for international audiences without losing its cultural core.

By placing Mahabharat: Ek Dharmayudh on its keynote stage, Microsoft positioned the project as a live case study in how AI can reshape content pipelines not just by speeding up production, but by expanding cultural reach. From Hastinapur to Tokyo, the message was clear: Indian epics are no longer bound by geography, language or format.

As broadcasters and tech leaders gathered at Inter BEE 2025 to debate the future of media, this AI-led retelling quietly made its point, the next global content wave may well be powered by ancient stories, retold through algorithms.