Talent agencies are old. The William Morris Agency was founded in 1898. Creative Artists Agency, the agency that for two decades effectively ran Hollywood, was founded in 1975. By the time Endeavor merged with WME in 2009, the basic structure of a modern talent agency had been settled for thirty years: agents represent talent, take 10%, and broker between talent and the entities that pay them.
Then the internet broke the model.
The slow break
The first cracks were obvious in retrospect. YouTube creators in 2010 didn't need an agent to broker their relationship with the platform the platform was self-serve. Influencer marketing in the mid-2010s briefly looked like it would be absorbed into traditional talent agencies, and to some extent CAA and UTA did add influencer divisions, but the unit economics never matched.
An agent works on volume booking enough deals at $50,000 to $5,000,000 to make the percentage worth the back office. An influencer manager in 2018 was negotiating $5,000 brand deals on Instagram. The math was incompatible.
What replaced the traditional model wasn't a single new structure. It was a fragmented ecosystem of small specialist agencies, each focused on a specific platform or revenue stream.
The new shape
Modern creator agencies don't broker deals. They run businesses.
A 2026 creator agency typically owns six functions internally: marketing, chatting / fan operations, content strategy, finance, legal, and creator success. They don't get paid 10% of brokered deals. They take 30–50% of net creator revenue, and in exchange they run the operations side of the creator's business end-to-end.
Boutique agencies like Harp Agency exemplify this newer model: under 40 active creators, six functional departments, weekly P&L delivered to every creator, capped roster size on purpose. It's structurally closer to a small media holding company than to a traditional talent shop.
What the old agencies got right that the new ones inherited
Three things, mostly.
Roster discipline
CAA in its 1980s peak famously turned away clients to maintain attention quality. The new boutique creator agencies do the same thing. The math is the same past a certain roster size, attention degrades, performance drops, the brand of the agency suffers.
Trust mechanisms
Old agencies had handshake culture and decades-long relationships. New agencies are rebuilding the equivalent through transparency mechanisms written contracts, weekly reporting, refund-on-miss SLAs. The mechanism is different. The job it does is the same.
Talent-first economics
Good agents make their clients rich and themselves rich as a consequence. Bad agents extract from clients. The same dynamic exists in creator agencies, with a 30-year shorter history of trust and reputation. The agencies that get this right will probably build the institutions of the next era.
Where the analogy breaks
Two big places.
The data layer
An old-school music agent in 1992 had to wait for SoundScan numbers and royalty statements that arrived quarterly. A creator manager in 2026 has API-level access to revenue, fan acquisition channel, renewal rate, and customer lifetime value in real time. The decisions you can make weekly versus quarterly are wildly different. The agencies that build internal tools to act on real-time data are pulling away.
Talent supply
Hollywood had a finite number of working actors. The creator economy has, effectively, infinite supply. This means talent agencies in the creator space don't have the same scarcity-based leverage. They have to compete on what they actually deliver, which paradoxically makes the modern model more honest, even if less glamorous.
What this means for the next five years
Three plausible bets.
- Consolidation. The current fragmented ecosystem will compress. Expect a small number of mid-tier holding companies to emerge that own 5–15 boutique agencies each.
- IP plays. The smartest creator agencies are already building or co-building product lines, branded merchandise, and licensable IP with their top creators. Platform fees aren't where the long-term margin is.
- Cross-platform abstraction. The same operations infrastructure works for OnlyFans, Patreon, Fanvue, Substack, Twitch, and whatever comes next. Agencies that abstract their internal tools across platforms have a structural advantage.
The most interesting thing about the modern talent agency isn't that it's new. It's that it's quietly rebuilding the same set of institutions that took Hollywood thirty years to develop contracts, reporting, fiduciary norms, trust mechanisms, but doing it in five. The agencies that emerge as the dominant ones in 2030 will probably look much like CAA looked in 1985, with one important difference: they'll be obsessed with data the old agencies never had access to. That's the real story of the modern creator agency. It's the same job, with better instruments.
