Spotify's quarter was strong (788 million monthly active users, nearly €5 billion, about $5.6 billion, in revenue), but the more interesting part of its October 22, 2026 earnings call was a strategic bet about AI agents. Co-CEO Gustav Söderström told investors that Spotify is "in one of the few categories that actually stands to thrive in the agentic age," because agents are good at taking over chores and entertainment is the opposite of a chore. The quotes below come from TechCrunch's report on the call; Spotify's own results are on its newsroom.
This is a company strategy statement, not a measured result. This guide sets out what he claimed, what can be checked, and why it matters for other platforms deciding whether to let AI agents in.
TL;DR
| Claim | Source | Status |
|---|---|---|
| 788 million MAU, nearly €5 billion revenue in Q3 | TechCrunch, citing Spotify results | Reported by the outlet from the earnings release |
| Agents are a tailwind for Spotify, since "nobody asks an agent to summarize a song" | Söderström on the call | Strategy claim |
| Spotify integrates with Meta Muse (first music service), ChatGPT and Claude | TechCrunch | Reported; no usage data |
| "Really good adoption of Spotify inside these agents" | Söderström | Unquantified |
| Spotify does not send listener data to partners | Söderström | Company statement; not independently verified |
What Söderström actually argued
His reasoning has four parts, per TechCrunch's account.
- Agents replace chores; entertainment is not a chore. "People love getting rid of chores. But entertainment is the opposite. Nobody asks an agent to 'summarize a song' for them."
- Agents will do the legwork, not replace listening. Instead of opening Spotify, searching for an artist and building a playlist, a user might ask an assistant for music for a run, a podcast about a topic they have been discussing with an AI, or an audiobook recap that resumes exactly where they left off.
- Spotify wants to be the service agents choose. "Our goal is to be the media service that you use from any agent, really turning the growth of agents into a tailwind rather than a headwind." He contrasted Spotify with "other companies" that are "maybe trying to figure out what to do and many are leaning out."
- Spotify itself is an agent. He described the company as "an intelligent media service encapsulated as an agent," one you can click in the app, or talk to, ask questions of, and instruct.
His example of what that enables: ask Spotify to go through your listening history song by song back to 2008, pick your true favorites, speed them up or down with beat-stretching, sort them by key and BPM, and mix them into a running playlist at your exact cadence.
What is checkable and what is not
- Integrations. TechCrunch states that Spotify was the first music service to integrate with Meta's Muse assistant and also integrates with ChatGPT and Claude. Those integrations are product facts you can test yourself in each assistant, but the call quotes give no usage figures, so "really good adoption" is not measurable from the coverage.
- The data claim. Söderström emphasized that Spotify is not sending its listener data to partners; the agent asks for something, and Spotify's own systems reason about the request and choose what to play. That is a design description. Whether an agent's request text, the user's identity, or interaction data are shared in the process is not explained in the coverage, and we cannot verify it.
- The example. The history-sorting workflow is described as something "no other media services can do." It is a demo claim on an earnings call, with no release date or availability stated in the report.
- The competitive framing. Saying others "lean out" is rhetoric. TechCrunch's counterexample is concrete: Amazon has blocked Meta's Muse from shopping on its platform, as we covered in Amazon blocks Meta Muse shopping agent.
Why Spotify's position is different from retail's
Retailers and banks have reasons to hesitate about agents: security, fraud, privacy and loss of the customer relationship. Banks are restricting third-party agents, as in our report on banks blocking AI agents from customer accounts, and Amazon's block was discussed by Paul Graham as a startup opportunity. A music service has lower risk: an agent that picks songs cannot drain an account, and every stream still happens inside Spotify's catalog, with its subscription and ad model intact. That is the structural reason the "tailwind" argument is plausible for streaming and harder for shopping. The agent becomes a new front door to the same room.
The argument has weak points. If listeners ask an assistant to "play something for my run" and never see Spotify's interface, the service loses the surface where it sells podcasts, audiobooks, merchandise and ads, and the assistant owner gains the relationship and the data about requests. Söderström's answer, that Spotify's catalog and listening history make recommendations better than a generic assistant's, is the key assumption, and it is untested in public data. Standards for how agents and services talk to each other, like the open protocol covered in Personal Agent Protocol with Sierra, Meta, Walmart and Stripe, could shape who captures that value.
The AI music backdrop
Spotify's agent push sits alongside its AI-content policies. Our coverage of Spotify removing 50 million "slop" tracks and changing recommendations describes how the platform is handling AI-generated music, an issue distinct from agent access but tied to trust in what an assistant might play for you. On the Muse side, see our Meta Muse safety verdict and the report on Muse linking bank accounts through Plaid, which show what users are asked to trust when a personal agent acts across services.
Three scenarios for how this plays out
Scenario one: the agent as front door. Assistants become the default way many people start listening, Spotify remains the catalog behind them, and the company's growth comes from more listening hours as friction falls. In this world, Söderström is right: the agent adds convenience and Spotify keeps the subscription. The risk is bargaining power: if three or four assistants control the front door, they can negotiate terms, prefer their own music services, or demand data.
Scenario two: the agent as substitute. Assistants build or license their own music offerings and treat Spotify as one source among many, optimizing for whichever is cheapest. Spotify's catalog and listening history would then be the differentiators it must defend, which is why the claim about personal recommendations matters so much. Nothing in the call coverage shows this happening yet.
Scenario three: the walled garden. Spotify's own assistant, the "agent you can talk to" inside the app, becomes the main experience, and external agents are a secondary channel. The company's remark that it can "do things today that no other media services can do" hints at this path: make the in-app agent so capable that users do not need to leave.
These are not predictions; they are the three outcomes that the call's statements leave open. Watching which one shows up in app-store rankings, partner announcements and usage data over the next few quarters is the real test of the "tailwind" claim.
How this fits Spotify's recent messaging
Spotify has been public about both embracing and policing AI. On one side it positions itself as agent-friendly, as described above. On the other it has taken action against low-quality AI music and has talked about putting human artistry first, a theme in the newsroom's recent posts. The two positions are consistent: an assistant is only a good front door if what it plays is trustworthy, and a catalog flooded with synthetic filler would undermine the "agents choose Spotify" pitch. That is why the agent strategy and the content-quality policy should be read together.
What to watch
- Numbers. Any figure for plays, sessions or new subscribers that come from agent integrations.
- Terms. Whether Spotify's developer or partner terms restrict how agents can use catalog data, caching or recommendations.
- Privacy details. What an assistant sees (play history, playlists) and what Spotify logs when it is called by an agent.
- Feature availability. When the history-sorting playlist workflow ships, and to whom.
- Copycats. Whether other media services (video, podcasts, news) adopt the "be the service agents choose" stance, or follow Amazon's block.
Practical notes for users and builders
If you use Spotify through an assistant, check which account permissions the assistant requests, and review them as you would for any connector. If you build agents, expect a split web: services that welcome agents with proper authentication and terms, and services that block them. If you run a service, decide deliberately which camp you are in, and write it into your robots and API policy before agents decide for you. For the wider agent landscape, see our comparison of proactive agents from Claude, OpenAI, Meta and Grok and the opinion piece on agent teams and a web closing its doors.
Related reading
- Amazon blocks Meta Muse shopping agent
- Banks block AI agents from customer accounts
- Paul Graham on Amazon banning agents
- Personal Agent Protocol
- Spotify removes 50 million slop tracks
- Official: Spotify newsroom and TechCrunch report
Details are accurate as of October 22, 2026 and come from TechCrunch's report on Spotify's earnings call. Statements about adoption and data handling are Spotify's own.