From Physics Lab to Music Studio
Mikey Shulman’s journey seems unlikely at first glance. A physicist with a Harvard PhD in quantum computing doesn’t typically end up revolutionizing music creation through AI music. But dig deeper, and the path makes perfect sense.
Since age four, Shulman played piano. Through high school and college, he rocked bass guitar in bands, performing at intimate New York clubs. He loved making music but felt frustrated by the barriers—years to master an instrument, more years to learn production software.
The gap bothered him: billions listen to music, yet only a tiny fraction create it. What if AI could bridge that divide?
The Kensho Connection
At Kensho Technologies, a financial AI startup, Shulman served as the first machine learning engineer and later ML lead. Working on natural language processing, his team transcribed corporate earnings calls, analyzing sentiment to predict stock movements. Audio processing opened his eyes to new possibilities.
All four Suno co-founders—Georg Kucsko, Martin Camacho, and Keenan Freyberg—came from Kensho. Music-loving engineers who saw the future clearly. In 2022, they left together to launch Suno.
Kensho had been acquired by S&P Global for over $500 million in 2018. Shulman spent seven years there mastering NLP systems and time-series prediction models, learning how to commercialize AI at scale. His pivot to music AI wasn’t chance—it was inevitable.
Kensho Technologies analyzing corporate earnings calls using natural language processing (source: Gemini)
Making Music from Words
Suno’s first product, Bark, launched in April 2023. This open-source text-to-audio model generated sound effects like laughter and crying, earning 19,000 GitHub stars.
The real breakthrough came in December 2023. Users could type prompts like “80s-style rap about grocery shopping” and Suno would create complete songs. GPT-4 generates lyrics while Suno’s proprietary model handles vocals, harmonies, instrumentation, and song structure.
The technical secret? Transformer architecture with autoregressive generation. “Autoregressive methods create more interesting music than diffusion models,” Shulman explains. Building music piece by piece produces beautiful melodies, even if recording quality occasionally suffers.
“Audio is far more fascinating than text,” he notes. “But the data is massive, unstructured, and lacks common crawl datasets. Our love for music let us overcome these challenges.”
Text to audio (source: Gemini)
Explosive Growth, Billion-Dollar Valuation
The numbers tell the story. In May 2024, Suno raised $125 million in Series B at a $500 million valuation, led by Lightspeed Venture Partners with participation from Matrix, Founder Collective, Nat Friedman, and Daniel Gross.
Six months later, everything changed. By October 2025, Bloomberg reported Suno was raising over $100 million at a $2 billion valuation—4x the previous round.
Financial performance backs the hype. According to Sacra, a market intelligence platform tracking tech investments, Suno hit $147 million ARR by September 2025, up from $45 million at year-end 2024. Nearly 50% of users who create the free tier’s 10-song limit convert to paid subscriptions.
The pricing is straightforward: 10 songs daily for free, $10 monthly for 500 songs, $30 monthly for 2,000 songs. Microsoft integrated Suno directly into its AI-powered Copilot.
Suno AI interface (source: Suno AI)
Copyright Battle Lines
Success brought scrutiny. In June 2024, Sony Music, Universal Music Group, and Warner Music Group sued Suno for allegedly training AI on copyrighted music without permission. Led by the RIAA, the lawsuit seeks up to $150,000 per infringement—potentially billions in damages.
Germany’s GEMA copyright organization and independent artists filed separate suits. In October 2024, major labels added accusations of “stream ripping” from YouTube videos.
Shulman pushed back hard. “Suno’s technology is designed to generate completely new outputs. It doesn’t memorize or reproduce existing content. That’s why we don’t allow prompts referencing specific artists.”
Ironically, Bloomberg and Financial Times reported Universal and Warner are negotiating licensing deals with Suno. Discussions cover licensing fees and equity stakes. Investors want legal issues resolved before committing more capital. Suno needs investment to fund settlements. Can both happen simultaneously?
Democratizing Music Creation
Shulman’s vision is clear: music must evolve from passive consumption to active creation. Today we stream songs on Spotify. Tomorrow, everyone makes their own.
“I didn’t want to build a company that makes 10% of creators 10% faster,” he says. “To impact a billion people’s music experience, you need to build for a billion people.”
User behavior splits two ways. Power users spend hours refining tracks, pursuing perfection. Casual users create songs about lactose intolerance with their kids or immortalize the moment Starbucks misspelled their name. Suno calls this “turning life into a soundtrack.”
In June 2025, Suno acquired WavTool, an AI-powered digital audio workstation, extending its reach to professional musicians.
September 2025 brought the V5 model with major improvements in audio quality, lyrics, and song structure. The cover feature proved popular—reimagining familiar songs in different genres lowers creative barriers. Genre mixing and detailed instrument-level tuning became possible. Suno can now effectively replace traditional DAW tools.
Suno V5 (source: Suno AI)
Will AI Democratize Music?
Shulman and Suno pose a fundamental question: Should music remain the domain of specialized professionals, or become everyone’s expressive medium?
Twenty years ago, photography required professional cameras and studios. Now everyone shoots selfies on smartphones and shares instantly. Suno believes music will follow the same path.
The $2 billion valuation signals investors are betting on this vision. $100 million ARR proves real demand exists. Copyright lawsuits echo conflicts that arise during every industrial transformation. Record labels eventually came to the negotiating table.
One question remains: When AI-generated music becomes indistinguishable from human-created music, what will we choose? Shulman argues “good taste matters more than good technology.” AI provides tools; human taste and choices determine value.
Suno’s experiment continues. A physics PhD who once taught at MIT now dreams of gifting music to a billion people. That dream is halfway to reality.
The Road Ahead
How far will Suno’s experiment go? That question starts now. OpenAI is expected to add similar functionality to ChatGPT in 2026. Will dedicated Suno users help the company differentiate further? Or will it fade like countless startups swept away by OpenAI’s feature updates?
This is the startup’s fate—and fortune—to discover.
See more insightful news!
- “7 Simple Habits That Will Make Your Adult Friendships Last”
- “Europe Awakens: 7 Investment Opportunities from a 30-Year Breakthrough”
- “China’s Global AI Ambition: The Shocking Strategy Behind Its New Action Plan”
- “Decoding the Fed’s Warning: How $1.1 Trillion in Margin Debt Threatens Your Portfolio”
- “How to Raise Your Child as an AI Super-Talent Earning $10 Billion?”
