Spotify's current AI bet is about being the music service that agents choose | TechCrunch
Short summary
Spotify's co-CEO argues that AI agents will help rather than hurt the company, because people want agents to handle chores but not entertainment, so Spotify wants to be the music service those agents call on. The company is betting on integrations with assistants like Meta's Muse, ChatGPT, and Claude, while keeping listener data in-house.
- Spotify reported 788 million monthly active users and nearly €5 billion in quarterly revenue, with the earnings call focused on how agents could drive listening.
- Spotify was the first music service integrated with Meta's Muse assistant, and it also works with ChatGPT and Claude, so users can get music and podcasts from inside those tools.
- Spotify says it doesn't send listener data to partners; the agent asks Spotify for something and Spotify's own systems choose what to play.
- Amazon has blocked Meta's Muse from shopping on its platform, a sign that major platforms may restrict outside agents over access, security, and commercial control.
Full summary
Spotify's third quarter was strong on the numbers: 788 million monthly active users and nearly €5 billion (about $5.6 billion) in revenue. But co-CEO Gustav Söderström spent much of the earnings call on a different question, which is how people will listen to music once AI agents become the front door. His argument is that Spotify sits in a category that benefits from agents rather than losing to them. Agents are good at taking over chores, and nobody wants to be doing the chore of finding music. Entertainment, in his framing, is the opposite of a chore, so people want to enjoy it themselves and won't hand it off to an assistant.
Spotify's strategy follows from that. Rather than fighting agents, it wants to be the service they call on. It was the first music service to integrate with Meta's Muse assistant, and it also works with ChatGPT and Claude. Söderström said the company is seeing "really good adoption" inside these agents and contrasted Spotify's approach with competitors who are "leaning out" while Spotify is "leaning in." The picture he paints is of someone asking an assistant for a running mix, a podcast tied to a conversation they've been having, or an audiobook recap that resumes where they stopped, with Spotify doing the work underneath.
The article also flags a complication. Amazon has blocked Meta's Muse from shopping on its platform, which shows that major platforms may restrict outside agents over access, security, privacy, or control of commerce. Spotify's answer on the data side is that it does not send listener data to partners. Instead, an agent makes a request and Spotify's own systems decide what to play. The personalization pitch rests on years of listening history, including repeat artists, skips, and discoveries, and Söderström offered a concrete example: asking Spotify to sort a user's favorite tracks back to 2008 by key and BPM, stretch their tempo, and assemble a playlist matched to a running cadence. Whether agent platforms will keep letting Spotify in is the open question the earnings call didn't settle.