Suno, the AI-driven music creation app that has been getting attention for its ability to create unnervingly good music, has raised USD $125 million in a Series B funding round.
Among the investors in Suno are VC firm Lightspeed Venture Partners, which has also invested in Stability AI and Everyrealm, the metaverse firm backed by music stars like Nas and Lil Baby.
Other investors include VC fund Founder Collective, and Nat Friedman, a tech executive and investor known for his stint as CEO of GitHub. He is an advisor at Midjourney, an AI techn platform that generates images.
Joining them is Daniel Gross, a colleague of Friedman’s who led AI development efforts at Apple and served as a partner at Y-Combinator.
“We released our first product eight months ago, enabling anyone to make a song with just a simple idea. It’s very early days, but 10 million people have already made music using Suno,” Co-Founder and CEO Mikey Shulman said in a blog post announcing the funding round.
“While Grammy-winning artists use Suno, our core user base consists of everyday people making music — often for the first time.”
According to sources cited by The Information, the funding round gives the two-year-old, Massachusetts-headquartered company an implied value of $500 million.
The newly-raised funds will be used to expand the company’s staff, which as of earlier this year, amounted to just 12 employees.
Suno has been soaring in popularity ever since a Rolling Stone article earlier this year described how the AI tech’s ability to create authentic-sounding music unnerved even some of its creators. The algorithm uses OpenAI’s ChatGPT to generate lyrics, and its own proprietary algorithms to generate music and vocals.
However, speculation has grown that Suno’s model was trained on copyrighted music without license from the rights holders.
In a guest column for MBW, Ed Newton-Rex, founder of Fairly Trained, a non-profit that certifies ethically trained AI, listed off uncanny similarities between music generated by Suno and copyrighted songs such as Queen’s Bohemian Rhapsody, Abba’s Dancing Queen and Oasis’ Don’t Look Back In Anger, among others.
Suno’s founders – Shulman, Georg Kucsko, Martin Camacho, and Keenan Freyberg, all formerly of AI startup Kensho – stress that, unlike some other generative AI models, Suno doesn’t allow users to ask for music created in the style of any specific artist.
However, an early investor in the firm, Antonio Rodriguez, a partner at VC firm Matrix, told Rolling Stone he is prepared for the possibility that Suno may face lawsuits from music labels and publishers.
He described this as “the risk we had to underwrite when we invested in the company, because we’re the fat wallet that will get sued right behind these guys…”
He added that “if we had deals with labels when this company got started, I probably wouldn’t have invested in it. I think that they needed to make this product without the constraints.”
As AI has exploded in popularity over the past few years, so have lawsuits by rightsholders alleging copyright infringement by AI developers. In one closely-watched case, Universal Music Group joined several other music companies in suing Anthropic AI, claiming that its Claude chatbot rips off copyrighted song lyrics.
Other such lawsuits have involved allegations that AI copied books and newspaper articles.
However, many AI developers – including Google – and investors in AI have argued that there should be a “fair use” exemption to copyright laws for the training of AI – in much the same way that educational materials are given limited exemptions to copyright law in many countries.
“While GRAMMY-winning artists use Suno, our core user base consists of everyday people making music — often for the first time.”
Mikey Shulman, Suno
However, music industry groups, and Newton-Rex, argue that such an exemption isn’t appropriate in the case of training AI models, because those models can generate content that competes with the copyrighted material they trained on – as in the case of an AI that can create music.
Thus far, courts haven’t made any definitive rulings on whether or not AI training should be exempt from copyright.
A number of jurisdictions – most notably the European Union – have moved forward with legislation that could potentially settle the issue.
The EU’s AI Act includes provisions that require AI developers to seek authorization from rightsholders to use copyrighted materials to train their models. However, legal analysts have pointed out that the AI Act includes certain copyright exemptions as set out in EU law, for instance, for research purposes.
And the Act’s requirement for copyright to be obtained applies only to what the law calls “general-purpose AI.” Some legal experts say it’s unclear what does and doesn’t qualify under that term.Music Business Worldwide
