Gaurav Sharma helped get music to a hundred million people, then decided pressing play was not nearly enough for them

The standard route into music technology is to love music, learn to code, and treat the industry as a distribution problem to be solved by somebody who has never had to clear a sample. Gaurav Sharma took the better one. He studied Computer Music Composition at the Peabody Institute, which is a conservatory, which means he spent his formative years surrounded by people for whom music was something you made rather than something you queued.

Then he went and built one of the great music build-outs of the decade. He joined Saavn as chief operating officer in 2010 and helped carry it past a hundred million monthly listeners and through the 2018 Reliance Jio merger that made it JioSaavn. Getting a subcontinent onto licensed music at that scale is genuinely hard and genuinely good, and most people who pull it off spend the rest of their career doing a slightly different version of the same thing.

What he saw from up there

He did not do that. He looked at a hundred million people holding every song ever recorded and got interested in the one thing they still could not do, which was touch any of it. His line is that fans do not want less connection to artists, they want more, and that is exactly right.

The proof is that they went and took it anyway. A decade of remixes, edits, sped-up versions and mashups built the entire culture of short-form video, almost none of it cleared and almost none of it paying the people who wrote the songs. Hook is the version where all of that is legal and everybody gets paid, which is a far more ambitious thing to attempt than simply building another place to listen.

The line he draws

Sharma founded an AI music company and spends a good part of his public voice insisting that AI companies pay musicians. Those two things only look opposed if you skip what he is actually objecting to. His complaint is not with the technology, it is with training on musicians' work and never paying for it. Hook licenses more than twenty million songs, pays out every time one is used, and lets artists claim what fans make with it wherever it travels.

That is a far harder company to build than one that scrapes first and negotiates later. Every song in that catalogue is a deal somebody had to go and sign. His own framing is the sharpest version of the argument: the real fight is not AI against artists, it is whether the innovation and the people who own the work end up on the same side of the ledger.

Plenty of people say something in that neighbourhood. Fewer say it from inside a company that has to carry the cost of meaning it, and fewer still are saying it after running the operations of a service with a hundred million users, which is exactly where you learn what this industry will and will not sign.

The receipts

The app landed on iOS in September 2024 and grew 45x in the twelve months before its Series A, which Khosla Ventures led at $10 million in February 2026, taking the total raised to $16 million. Downtown, Primary Wave, Avex, Too Lost and Glassnote are on the licensing side, along with Universal.

The part that is hardest to fake is the artist side. The Weeknd, Lil Wayne, Metro Boomin and Empire of the Sun have run remix collaborations on Hook directly, and artists at that level do not hand their catalogue to something they think is a threat.

The bet underneath all of it is that the licensed version can also be the fun one. If that lands, a decade of remix culture stops being something the industry argues about in court and becomes something it gets paid for, and the person who made that happen will have been the guy who built the streaming habit in the first place. It is the most interesting bet anybody in music technology is making right now.

The real debate isn't AI versus artists, it's whether innovation and rightsholder participation can coexist.

Gaurav, writing in Music Business Worldwide, August 2026