Product
On Daniel Ek's departure, what he said, and what he didn't
After nearly two decades as CEO of the world's largest music streaming platform, Daniel Ek has stepped down — into the role of Executive Chairman, where he will continue to set strategy, oversee the two co-CEOs replacing him, and, by his own account, work from the same room as the people nominally now running the company.
'We actually sit in the same room, the three of us,' said incoming co-CEO Alex Norström on the investor call. 'People are very surprised when they see that.'(1)
The surprise presumably belongs to people who expected the announcement of a leadership transition to describe an actual leadership transition. Ek built Spotify into a $90 billion company and is its largest individual shareholder. He is not leaving. He has changed his title.
This would be unremarkable — founders stepping into executive chairman roles while maintaining operational influence is a standard corporate manoeuvre — except that Spotify's relationship with the artists whose work built that $90 billion company has become sufficiently strained that the question of who is making decisions at the platform is no longer merely a governance footnote. It is the story.
In June 2025, Ek's investment firm Prima Materia led a €600 million funding round in Helsing, a German-French defence technology company developing AI-powered software and hardware for military applications, including autonomous combat systems.(2)The investment was made through Ek's personal vehicle rather than through Spotify, a distinction that Spotify emphasised when artists began withdrawing their catalogues in response.
The distinction satisfied nobody in particular. King Gizzard and the Lizard Wizard, Deerhoof, and other artists pulled their music from the platform. Deerhoof's statement was direct: they did not want their music killing people, and they described Spotify as a data-mining scam operating under the guise of a music company. The framing — that the platform had never been primarily about music — was the more pointed part of the criticism, because it located the Helsing investment not as an aberration but as a clarification.
Ek's personal fortune, estimated at $9.9 billion, was built on a platform that pays artists a fraction of a cent per stream. The €600 million available for autonomous weapons investment represents, in one sense, the accumulated returns from a decade and a half of that pricing model. The artists who are no longer receiving fractions of a cent from their Spotify streams contributed, in aggregate, to the capital that was available for redeployment into defence technology. Whether this constitutes a moral problem depends on your view of the obligations that attach to wealth accumulated through other people's creative labour. Artists appear to have formed a view.
Ek's farewell statement, published on the occasion of his transition to Executive Chairman, is worth reading in its entirety. Not because it is unusual — it is a thoroughly standard piece of corporate communication, warm without being specific, reflective without being revealing — but because of a particular absence.
The word 'music' does not appear in it. The word 'product' appears three times.
To me, this is no small thing. Spotify was founded on the proposition that it could save the music industry from piracy by making legal streaming more convenient than illegal downloading. The company's founding narrative is inseparable from music — from the specific claim that what Ek was building was a better deal for artists and listeners than what came before. The farewell statement of the founder who built the company on that narrative does not contain the word that the narrative was built around. It contains 'product.'
Musicians are accustomed to being told that their work is content, their expression is a product, and their artistic output is a commodity to be optimised for streaming metrics. Most of them are accustomed to it precisely because Spotify's decade of market dominance has made this vocabulary the default language of the digital music industry. Ek's statement offers the unusual transparency of saying the quiet part without apparently noticing it is quiet.
The Helsing investment and the royalty structure are the most visible points of friction, but Liz Pelly's 2025 book Mood Machine documents a more structurally significant problem: Spotify's use of its own recommendation infrastructure to promote music it had commissioned from production partners, often at rates substantially lower than standard royalties.(3)
A Swedish newspaper investigation found that as few as 20 songwriters were responsible for more than 500 'artists' on the platform — generic, mood-tagged productions designed to occupy the functional music categories (sleep, focus, background) that generate high stream counts with minimal listener engagement.(4) These productions were surfaced by Spotify's recommendation algorithm at the expense of independent artists working in the same categories, who were competing without knowing that the platform itself was a participant in the competition — and a participant with a structural advantage, given that the algorithm was making the recommendations.
Spotify removed approximately 75 million tracks classified as spam in 2025.(5) The company has not published data on AI-generated music on the platform, which is a notable omission at a moment when AI music generation has become sufficiently accessible that the volume of algorithmically produced content is a meaningful question for any platform's royalty economics. The removal of tracks the company acknowledged as spam is the disclosure. What remains on the platform, and what its composition is, is not a question Spotify has chosen to answer publicly.
The pattern across these issues — below-minimum royalties, fake artist promotion, AI content opacity — is consistent. In each case the platform's commercial interests diverge from the interests of the artists whose work gives the platform its value, and in each case the platform's interests have taken precedence. This is not surprising. It is what a platform optimised for shareholder return rather than artist welfare would do. The useful question is not whether Spotify has behaved in ways that prioritise its own interests, which it has, but whether the structure of the platform makes any other behaviour possible.
Spotify's per-stream royalty rate is approximately $0.003 to $0.005, depending on the listener's geography and subscription tier.(6) At the midpoint, an artist needs roughly 250 streams to generate a dollar in royalties. An artist with a dedicated following of several thousand listeners, streaming their music consistently, may generate annual Spotify income in the low hundreds of dollars — an amount that does not cover the cost of recording the music that generated it, let alone the ongoing costs of maintaining a musical career.
Deerhoof's Satomi Matsuzaki described the rate as insufficient to purchase a can of soda per hundred streams. This is arithmetically accurate. It is also a description of the terms on which a platform with a $90 billion market capitalisation has chosen to compensate the artists whose work it distributes. The market capitalisation and the per-stream rate are not unrelated figures. The platform is worth what it is worth in part because its content costs are structured the way they are.
The standard defence of the royalty structure is that Spotify pays out approximately 70% of its revenue to rights holders, and that the low per-stream rate reflects the low subscription price rather than an unjust distribution of available revenue. This defence has merit as far as it goes. Where it stops going very far is in the distribution of that 70% — the pro-rata model allocates royalties according to share of total streams, which concentrates revenue among the artists with the largest global audiences and distributes fractional amounts to everyone else. An independent artist with ten thousand dedicated listeners generates a different absolute royalty from a hundred million streams than Taylor Swift does, and the difference is not proportional to the difference in audience sizes.
Daniel Ek is still at Spotify. The royalty structure is unchanged. The co-CEOs who replaced him were chosen by him, sit with him, and will answer to him in his capacity as Executive Chairman and largest shareholder. The press release described a transition. The investor call described a continuation.
The artists who pulled their catalogues have made a statement. Whether the statement has economic consequences for Spotify depends on whether enough artists follow, and whether enough listeners follow the artists — both of which are genuinely uncertain. Spotify's catalogue is sufficiently large, and its switching costs for listeners sufficiently high, that individual artist withdrawals are more symbolic than financially material. The symbolism matters, but symbols require scale to become structural pressure.
The structural alternatives are developing. User-centric payment models, which distribute a subscriber's fee to the artists they actually listen to rather than to the global play count distribution, have been implemented by some smaller platforms and proposed as a reform at Spotify without adoption. Artist-owned and cooperative platforms are in development in several markets, including in Australia. These are not currently at a scale that challenges Spotify's market position. They are at the scale of demonstrating that different royalty structures and different ownership models are operationally viable — which is a different and more durable kind of argument than a catalogue withdrawal.
The departure that is not a departure, the product that is not called music, the fake artists and the genuine ones paid equivalently nothing: these are not separate problems with separate solutions. They are descriptions of a platform that was built to serve one set of interests and has served them consistently. The question for artists and listeners who find that arrangement unsatisfactory is not how to reform the platform from within its existing logic, but whether the conditions exist to build something outside it.
They do. The interesting period is now.
* * *
References and inspirations
1. Alex Norström quote from the Spotify investor call, Q2 2025, widely reported in music industry press including Music Business Worldwide and Billboard. The 'same room' detail was confirmed in multiple outlet transcripts of the call.
2. Prima Materia's €600 million investment in Helsing was reported in June 2025 across financial and music industry press. Helsing's product portfolio, including AI-powered military software and hardware systems, is documented on the company's own website and in defence industry reporting. The investment was made through Ek's personal fund, not through Spotify corporate.
3. Liz Pelly, Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist (Atria Books, 2025). Pelly's reporting is based on extensive interviews with former Spotify employees and industry participants and documents the Perfect Fit Content program in detail.
4. The Swedish newspaper investigation — conducted by Dagens Nyheter — identified the concentration of production behind multiple Spotify 'artist' identities and was published in 2023. The findings were subsequently reported internationally and contributed to Spotify's 2025 platform cleanup.
5. Spotify's removal of approximately 75 million tracks was announced by the company in 2025 and reported across music industry press. The company's decision not to publish data on AI-generated music remaining on the platform has been noted by Pelly and by industry analysts including the Music Industry Blog.
6. Spotify per-stream royalty range from the Trichordist Streaming Price Bible 2024 and independent analysis by Music Business Worldwide. The range varies by geography, subscription tier, and the rights holder's licensing agreement; the figures cited represent the commonly reported range for independent artists without major label deals.