What started out as a hobby project to test the abilities of AI algorithms to generate music has now turned into a multi-million-dollar business.
Friends and musicians Seth Forsgren and Hayk Martiros released AI music generator Riffusion last December, as an experiment with the use of “latent diffusion” AI technology to generate music.
Riffusion’s technology uses sonograms – visual representations of music – to train itself on musical genres, and create music using verbal prompts. The initial demo released by Forsgren and Martiros went viral, and Forsgren has told news media that millions of people have tried it out.
Earlier this year, the two formed a company that has now secured USD $4 million in a seed funding round, led by Greycroft Partners, South Park Commons and Sky9.
The company has also picked up a number of new advisors, including musicians Alex Pall and Drew Taggart of The Chainsmokers.
At the same time, Riffusion has announced it’s going live with a new app that enables anyone to create music via short, shareable audio clips that the company has dubbed “riffs.” App users can simply describe the musical style and lyrics, and Riffusion’s AI will generate riffs, along with artwork to go with them, in a matter of seconds.
“We started Riffusion as a hobby project last year and were blown away by the creativity it unleashed with casual users and professional musicians,” Forsgren said in a statement. “We see Riffusion as a new instrument – one that anyone can play. We all love music, but most of us just listen passively. Riffusion enables everyone to be an active participant in it.”
The company is clear that its technology is not intended to be used for copyright infringement, noting that the algorithm is “not built to produce deepfakes” and “doesn’t recognize famous artist names in its prompts. Instead, users describe musical qualities through natural language or even recording their own voice, and the team’s foundational model generates unique outputs each time.”
Riffusion’s investors are jumping into a potentially enormous new market. A report earlier this year from Market Research Future projected that the generative AI market could be worth $405 billion by 2032. That report named Riffusion as an “eminent player” in the burgeoning new market.
The number of startups aiming to take a share of this pie is growing rapidly, and now includes Stability.ai, the AI company that developed Stable Diffusion, on which Riffusion’s technology was originally developed. Stability.ai has now launched its own music generator.
Large tech and social media companies are also jumping into the game, with Google’s MusicLM and Meta’s AudioCraft being two prime examples.
“We see Riffusion as a new instrument – one that anyone can play. We all love music, but most of us just listen passively. Riffusion enables everyone to be an active participant in it.”
Seth Forsgren, Riffusion
And of course, the major recording companies are increasingly involved as well, often in the form of partnerships with generative AI companies.
Among the most recent examples is Warner Music Group-owned Spinnin Records signing a 50-album deal that will see AI music company Endel create functional music for sleep, concentration and relaxation out of Spinnin’s music catalog.
Warner also invested into AI music tech company LifeScore, which earlier this month announced it’s launching its own record label. The startup says the label will feature music composed and performed by artists, “then amplified with LifeScore’s generative AI technology”.Music Business Worldwide
