Entertainment

Saregama Turns to Generative AI to Build Music Videos for Catalogue Songs It Never Owned Visuals For

Kiran MachadoPublished 3d ago4 min readBased on 6 sources
Reading level
Saregama Turns to Generative AI to Build Music Videos for Catalogue Songs It Never Owned Visuals For
Photo by Saregama / Public domain

Saregama, India's oldest record label, is using generative AI to create music videos for film songs in its catalogue dating back to the 1960s and 1970s — songs for which the label has always owned the audio rights but never the original film footage.

On Saregama's fiscal first-quarter earnings call on Tuesday, August 4, Managing Director Vikram Mehra explained the gap. The label owns "all the rights connected to the song, except the right of the original music video," he said, because at the time the films were made, "there was no market for it" and the visuals were not sold to music labels. Mehra called the missing videos one of the "weaknesses" in Saregama's older catalogue. Music Business Worldwide

The cost of filling that gap with AI is striking. Mehra said a music video made with generative AI "can be INR 70,000 also" — roughly $740. That is a fraction of what a conventional music-video shoot would cost, and it allows Saregama to produce visuals for hundreds of older songs that have sat on YouTube and streaming platforms with only static images or lyric cards.

The strategy splits by audience. For Generation X, Saregama will keep the original audio and attach "more contemporary looking videos." For Generation Z, the label will "modify the audio a bit" with a "change in the instrumentation" while maintaining the underlying composition — essentially, updated arrangements wrapped around the same melodies, paired with AI-generated visuals.

Saregama's flagship YouTube channel, Saregama Music, has roughly 64 million subscribers, giving the label a direct distribution pipeline to test how these refreshed videos perform.

Beyond music videos, the label's content team is using third-party generative AI tools to create what Mehra described as "neighbouring audio content" — podcasts built around Saregama's songs, which the label then licenses to third-party podcast companies. The approach turns catalogue recordings into raw material for new formats without recording new episodes from scratch.

The AI push extends to acquisition. Mehra said Saregama's predictive AI — running for four years — helps decide which rights to buy, and he claimed the label's "hit-to-flop ratio is better than any competitor in the market." Saregama's content-acquisition budget for the year is INR 300 crore to INR 350 crore ($31.6 million to $36.9 million), fed by those predictive models.

On copyright, Mehra drew a line between licensed and unlicensed AI use. "Saregama's position on AI-based music remains that we support licensed innovation, but will continue our fight against unlicensed exploitation of our music," he said. The label, an RP-Sanjiv Goenka Group company, has entrusted Novex Communications with exclusive on-ground public performance rights for its sound recordings and music videos, with unauthorized use potentially triggering legal action under copyright laws. Saregama

Saregama is also reshaping its film strategy. The label is winding down its own film business and channeling film-music investment through Bhansali Productions, in which it invested $36 million in December 2025 for the music rights to its future films.

For a label whose catalogue spans decades of Indian film music, the AI videos are a practical fix to a rights gap that predates streaming, YouTube and even the compact disc. The songs were always there. The pictures were not.