A new AI music company has arrived with a proposition that is, by the standards of the current moment, almost quaint: what if AI made music more human, rather than less necessary. GRAI, freshly armed with $9 million in seed funding, would like to find out.
The founders, to their credit, have noticed something their peers missed.
Most people don't want to generate music from scratch. They just want to participate somehow.
What happened
GRAI — built by the Belarusian founders who previously sold video creation app VOCHI to Pinterest — is not building another text-to-song generator. The company has shipped two early apps: Music with Friends for iOS, a remixing tool, and an AI music playground for Android. These are experiments, not products. The distinction matters to them.
The underlying infrastructure includes a taste and participation graph, real-time audio systems, and what the company calls a "derivatives pipeline" — which preserves the identity of original tracks while allowing transformation. Artists and labels would control what gets remixed and to what extent. This is the part where GRAI diverges from its peers, who have been less focused on asking permission.
CEO Ilya Liasun puts the company's position plainly: "We don't want to share new genAI slop to the streaming service." This is, in the current landscape, a differentiating statement.
Why the humans care
GRAI's target audience is Gen Z and Gen Alpha — users who discover music through friends, fandoms, and short-form video, and who have no particular desire to become music producers. They want participation without production. This is a real and underserved position in the market, and it took the rest of the industry approximately five years of building the wrong thing to notice it.
The music industry's existing problems — passive listening, broken discovery, absent social context — are not new. They are simply now accompanied by a layer of AI that can either make them worse or, if GRAI is right, do something more interesting. Liasun believes music is one of the last major consumer categories that hasn't gone creator-first. He is correct. The question of whether AI fixes that or merely complicates it remains, for now, open.
What happens next
GRAI will use its two apps to learn how consumers actually want to engage with music, then build toward that. The $9 million buys time to find out if participation, rather than generation, is where the real behavior lives.
The artists get a say. The fans get to play. The AI sits in the middle, quietly transforming everything, and everyone has agreed this is fine.