May 13, 2026, 8:00 AM ETInterview, Fireside Chat

Suno's Mikey Shulman: Everyone Can Make Music Now

Physics-to-AI Transition: Founder Mikey Shulman (PhD in Quantum Computing) leveraged the insight that playing at the nexus of disparate fields (e.g., physics and audio) represents a massive opportunity across domains.
Initial Miscalculation: Early "back-of-the-envelope" math suggested generating high-quality music was orders of magnitude away in terms of compute and model size due to audio's continuous nature.
Core Breakthrough: The team realized audio need not be modeled using discrete musical constraints (e.g., the 12-tone scale or specific instrument sets) but rather as continuous sound waves sampled at 48,000 Hz (float 32 numbers).
First Principles Approach: By discarding traditional music theory constraints, the model avoids limitations on instrument sets, allowing for the emergence of novel sounds and microtonal music rather than just replicating existing genres.
Early Validation: Initial 12.5-second clips released on Discord were technically poor but generated immense user engagement, confirming the product-market fit despite the poor quality.

Creative Entertainment Model: Suno redefines music from a passive consumption medium to an active "creative entertainment" platform where the primary value is the act of creation itself, similar to cooking or coding.
User Demographics: Approximately 90% of daily users are creators, driven by the psychological fulfillment of production rather than the intent to distribute content elsewhere.
Quality vs. Formula: The model excels in formulaic genres (country, pop) but has also raised the "ceiling" of creativity by enabling bizarre genre blends (e.g., trap with sitar) that were previously impossible.
Voice and Lyrics: The model uses LLMs to generate lyrics based on user prompts, with a specific focus on full songs with vocals (stories) rather than short background loops.
Technical Trade-offs: The company prioritized autoregression over diffusion to ensure the model generates coherent narratives (full songs) even if the immediate audio fidelity is lower than competitors focusing on crisp, short clips.

Revenue and Scale: The company is targeting a $300 million revenue run rate, though Shulman notes the product is still in early stages with significant growth potential.
Warner Music Partnership: Suno secured a landmark settlement/partnership with Warner Music, signaling a shift from adversarial relations to collaboration on products that deepen artist-fan interaction.
Industry Integration: Shulman asserts that most professional music will eventually incorporate AI (similar to the ubiquity of auto-tune), aiming to avoid a bifurcated "AI vs. Non-AI" market.
The "Slop" Rebuttal: Shulman rejects the term "slop" for high-volume user-generated content, arguing that increased production volume historically leads to more great music and new stars, citing the laptop-beat revolution of the past.
Chart Success: Suno has already facilitated chart-topping hits, record deals, and viral singles, demonstrating that AI-generated art can achieve mainstream commercial success.

Product Moat: Shulman believes a moat cannot rely solely on model scale (where giants like Google have advantages) but must be built through superior UX, UI, and product differentiation.
Consumer Focus: Unlike many AI startups targeting enterprise automation, Suno focuses on consumer creativity, a sector requiring high creativity to envision rather than just efficiency.
Upcoming Features (Next 12 Months):
- Social Co-Creation: Enabling users to share song templates for others to riff on, and facilitating asynchronous collaboration.
- Voice Cloning: Allowing users to use their own voices in songs to increase emotional attachment and personal resonance.
- Music Video Beta: Launching a feature to generate music videos that elevate the narrative of the song, distinct from background video generation.
Concert Innovation: The long-term goal includes interactive concerts where audiences can participate in making music with artists, potentially debuting at festivals like Coachella within a year.
Research Approach: The company relies heavily on human preference data to align models with creative tastes, creating a virtuous cycle where product usage drives model research and improvement.