#874: Guy Oseary — The Legendary Hollywood Power Broker on 5-Minute Decisions, 36 Years of Managing Madonna, 26 IPOs, and Spotting Magic First

What It Covers

Guy Oseary, manager of Madonna for 36 years and co-founder of Sound Ventures with nearly $2 billion under management, traces his path from a teenager using a borrowed Beverly Hills address to attend a better school, through signing Alanis Morissette and Candlebox, producing Twilight, and investing early in Airbnb, Uber, Spotify, and Anthropic at a $5 billion valuation.

Key Questions Answered

• •Five-Minute Decision Framework: Oseary makes roughly 90% of investment and talent decisions within the first five minutes of a meeting. The muscle developed from running a boutique label competing against major corporations — waiting meant losing deals to better-resourced rivals. When he saw Candlebox perform to 30 people, he visualized thousands singing along and called his partner from a payphone immediately. Speed signals conviction to founders and artists, which itself becomes a competitive advantage in deal-making.
• •Visualization as Due Diligence: Before committing to any investment or signing, Oseary mentally maps what he personally can contribute — marketing strategy, media partnerships, narrative storytelling, and relationship introductions. If he cannot visualize a concrete value-add, he passes. This self-assessment filter, combined with gut reaction, prevented him from spreading thin across deals where his specific Hollywood-to-tech network would add nothing beyond capital, keeping his portfolio focused on situations where his presence on the cap table is genuinely differentiated.
• •Diversification as Non-Negotiable: Oseary lost his entire life savings by concentrating everything into Idealab just before the dot-com crash. At the same moment, he passed on finalizing deals with Research In Motion (Blackberry) and Vitamin Water — both of which became massive. The lesson he applied to Sound Ventures: deploy across multiple foundational positions rather than concentrating. The AI fund deployed 80% of capital within a three-month window but spread across OpenAI, Anthropic, and several other foundational model companies simultaneously.
• •Thematic Window Investing: When ChatGPT launched in November 2022, Oseary and Ashton Kutcher identified a narrow three-month window before the AI investment landscape would become crowded and valuations would spike. They raised a dedicated foundational AI model fund, communicated the deployment timeline upfront to LPs, and committed to placing 80% of capital before April 2023. This resulted in early positions in both OpenAI and Anthropic, the latter entered at a $5 billion valuation now valued near $60 billion.
• •Artist Development as Company Building: Oseary applies the same framework to startups that he used building Candlebox from a 30-person showcase to 4 million albums sold — weekly incremental progress, consistent audience-building, and resisting pressure to rush. He views every founder as a recording artist with an album to share, and his role as identifying them before the market does, then helping craft the narrative, grow the audience, and sequence the "singles" (product launches, press moments, partnership announcements) to build sustained momentum rather than one spike.
• •Access Through Demonstrated Value, Not Credentials: Oseary built his early network at 15 by cold-approaching roughly five classmates at Beverly Hills High School, asking to meet their industry-connected parents. He arrived at those meetings with a physical portfolio — artist photos, demos, a logo — despite having no money, no experience, and no industry contacts. When Bernie Brillstein offered $25,000, Oseary declined and asked for three phone calls instead, correctly calculating that relationships compounded faster than early capital at that stage of his career.
• •AI and Music Rights — Structural Problem: AI music generation companies, including Suno (valued at $5 billion), have trained models on recorded music without compensating any artist. Oseary argues the fix is straightforward: build an opt-out mechanism for artists who refuse training use, and a revenue-sharing system for those who consent. He draws a parallel to Napster's collapse and Spotify's subsequent emergence — the entity that solves the rights infrastructure problem, not just the technology problem, captures the long-term value in AI-generated music.