Spotify’s latest earnings call makes a business case for welcoming AI agents rather than blocking them. TechCrunch covered the October 22 discussion.
The strategic question is who benefits when a listener gives instructions to another company’s assistant. Spotify’s answer is that a different interface need not mean a different listening service. Its financial disclosures show the scale of the business making that bet, but do not isolate the revenue attributable to agent-driven listening.
A distribution argument, not just a new interface
In the full original earnings-call transcript, co-CEO Gustav Söderström describes agents as another distribution channel. His argument is that Spotify can continue earning through subscriptions and audio advertising while operating behind another interface, much as it does through navigation apps and connected devices. That is management’s commercial thesis, not an independently established forecast.
Söderström also says the outside agent requests something from Spotify, which uses its own understanding of the listener to decide what to play. He describes this as keeping the reasoning over Spotify’s listener information within Spotify, rather than handing that information to partners. This is the company’s description of its architecture, not a technical privacy audit of every integration.
The important distinction is between the service that receives a request and the service that delivers the entertainment. If an assistant helps someone find a song, the economic result depends on where listening occurs and how that listening is monetized. Counting conversational interactions alone would not answer either question.
Quarterly growth does not identify an AI contribution
Spotify’s official third-quarter shareholder deck reports 788 million monthly active users and 306 million Premium subscribers. Revenue reached €4.969 billion, up 16% as reported or 14% at constant currency. Operating income was €672 million, gross margin was 33%, and the company’s non-IFRS free cash flow was €735 million.
Those are results for the wider business. They should not be read as a measured return on agent integrations. The deck also identifies temporarily elevated marketing and cloud/AI spending within operating-expense growth, without providing a separate profit calculation for listening initiated through outside agents.
For investors, the missing connection matters more than a simple claim that AI and revenue are both growing. To establish an incremental contribution, a useful disclosure would distinguish listening that an agent added from listening that merely moved from Spotify’s own interface. It would also connect the additional activity to revenue and the cost of serving it. The quarterly totals do not supply that comparison. Comparisons would also need consistent definitions of users and periods, so that an initial trial of an integration is not mistaken for a durable change in listening or subscription behavior.
An existing integration gives the strategy a concrete base
This is not an announcement that Spotify has just discovered external agents. In a clearly dated September 23 company announcement, Spotify described connecting its service to Meta’s Muse, including playback, saving items and creating playlists through a conversation. That earlier integration provides context for October’s earnings argument; it is not a new launch being republished today.
The opportunity and the dependency come together. Another interface could bring useful requests to Spotify, but it could also stand between Spotify and the initial decision about which service to use. A sensible commercial assessment would therefore examine both additional listening and how reliably Spotify remains the chosen destination.
The next convincing evidence would be consistent measures of agent-originated use, retention and monetization, with definitions that allow comparison over time. Until then, the new development is management’s explicit argument for participation in the agent ecosystem, supported by a functioning integration and a substantial existing business, rather than proof that agents have already produced the quarter’s growth.
