News · industry — By WATMM Desk —
Deezer turned a profit. Here is how its artist-centric model moves the money, and who it moves it from
Deezer reported €8.5 million of net income for 2025, which it calls its first positive net income, and credits its artist-centric payment model among its strengths. The model weights streams by who the artist is and how the fan found the track. We explain the rules, who gains and who loses, and why the profit says little about whether artists are better off.
Key facts
- What happened: on 18 Mar 2026 Deezer reported FY25 net income of €8.5 million against a €26.0 million loss in 2024, adjusted EBITDA of €9.7 million and revenue of €534.0 million, and said 85% of its partners are now on its Artist-Centric Payment System (ACPS).
- How it works: streams of artists with at least 1,000 streams a month from 500 unique listeners count double, actively chosen streams get a further boost (up to four times a basic stream in total), each user counts for at most 1,000 streams a month, and noise is taken out of the royalty pool.
- Why it matters: Deezer has not published any figures on how payouts to artists changed. Its own results put the gross-profit and EBITDA gains down to label terms, white-label deals and cost control. So the profit shows the model can sit inside a viable business, not that it helped artists.
Deezer has made money. Its FY25 results, released on 18 March 2026, show net income of €8.5 million after a €26.0 million loss in 2024. Chief executive Alexis Lanternier said it was the first positive net income in the company's history. Adjusted EBITDA was €9.7 million and free cash flow €10.1 million, on revenue of €534.0 million, down 1.4%. Deezer lists "leadership in fair artists remuneration" among its highlights and says 85% of its partners now use its Artist-Centric Payment System (ACPS). Its 2026 plan is to extend it to all partners.
That makes this a good moment to look at what the model actually does with the money, and whether anyone can show it helped.
How the model moves money
Deezer and Universal Music Group announced the model in September 2023 for launch in France in the fourth quarter of that year. It still pools subscription money and splits it by streams, as before. What changes is how much each stream weighs. Deezer's artist remuneration page and Sacem's member guide set out four rules.
- The "professional artist" boost. If an artist gets at least 1,000 streams a month from at least 500 unique listeners, each of their streams counts as two.
- The active-listening boost. A stream counts double again when the fan searched for the track or album, or played it from a playlist rather than an algorithmic feed. Deezer says the two boosts combined can make a stream worth four times a basic one.
- The per-user cap. Each user counts for at most 1,000 streams a month. Sacem's example: a listener who plays 2,000 streams in a month has each stream count as half. Deezer says the average user streams about 600 times a month, so the cap mostly hits heavy and automated accounts.
- Noise out. White noise and similar functional audio is replaced with Deezer's own content, which earns no royalties. Fraudulent streams are removed.
Deezer also cleaned its catalogue. A year in, it told Music Business Worldwide it had deleted more than 26 million tracks, about 13% of its catalogue: noise, mono-track albums, fake artists and tracks with no plays in a year.
Coverage has grown. By its FY24 results, ACPS applied to about 85% of Deezer's recording royalty pool, and the model had been extended to Merlin's independent members. In January 2025 Deezer and Sacem announced that the model would also apply to publishing rights in France. The first Sacem distribution under the model was due on 6 October 2025, covering streams from the first quarter of 2025, and it applied to all paid streaming on Deezer in France.
How it differs from Spotify
Spotify's answer to the same problem is a cut-off. Since 2024, a track earns nothing until it has had 1,000 streams in the previous 12 months, and the money it would have earned goes to everyone above the line. We go through Spotify's rule in our piece on the 1,000-stream threshold.
Deezer's version differs in three ways. It is measured per artist, not per track, so one popular song lifts a whole catalogue. It is a weighting, not a cut-off: an artist below the bar still gets paid, at the basic rate. And it publishes both numbers, 1,000 streams and 500 listeners, where Spotify says it has a unique-listener minimum but does not share the figure publicly.
Who gains, who loses
Deezer's own 2023 analysis is the best guide to the shape of it. It found that 97% of uploaders generated only 2% of total streams, while about 2% of uploaders had more than 1,000 monthly unique listeners. Noise was about 2% of streams, and roughly 7% of streams in 2022 were flagged as fraudulent.
So the clear losers are noise channels, stream farms and very heavy (often automated) accounts. AI track farms should fall in the same group. Thousands of tracks that each pick up a few plays are unlikely to reach 500 real listeners a month for any one artist, so they would stay at the basic weight while artists with real fans count double or more.
The more contested group is very small human artists. In 2023 Believe, whose TuneCore distributes DIY artists, called the boosts a "reverse Robin Hood" that takes from early-stage artists to pay established ones. Believe also said Deezer and UMG had not consulted artists or independents.
What we think
The profit is real, but it isn't evidence that the model works for artists. Deezer's release puts the gross-profit gain down to white-label deals and "the optimization of labels terms", and the EBITDA gain to that plus tight cost control. Not one figure in it shows how artist payouts changed. Two and a half years in, Deezer still hasn't published the number that matters: how much money moved, and to whom.
We can estimate the scale from Deezer's own 2023 figures. If the 97% of uploaders who drew 2% of streams were all paid at the basic weight while everyone else counted at least double, their share of the pool would fall from about 2% to about 1%, and lower once search boosts are counted. That is our arithmetic, not a Deezer figure. It matters for the critique. Believe's "reverse Robin Hood" point is right in direction, but the sums involved are small, because tiny artists never had much of the pool. Noise and fraud together were about 9% of streams, so cutting those out likely moved more money than the professional-artist boost did.
Our guess at the bigger redistribution is the active-listening boost, which works among artists who already qualify. Two artists with the same stream count can be paid very differently if one is found through search and the other through autoplay and mood playlists. That favours fan-driven genres and established names. It counts against session musicians, library composers and anyone whose career depends on being placed in algorithmic feeds, none of whom choose how their listeners arrive. Deezer and Sacem don't even describe the rule the same way: Sacem counts "personalised" playlists as active, while Deezer contrasts active streams with algorithmic suggestions. A rule this valuable should have one clear published definition.
For artists, the Deezer approach is still better than a hard cut-off. A first single on Deezer earns from its first stream, and the bar is per artist and public. Three things would make it easier to defend. Deezer should publish yearly before-and-after payout shares by artist size. It should show artists in their dashboards how many of their streams were boosted and why. And rights-holders signing up to ACPS, now 85% of partners, should tell their artists what changed. The model sets out to reward real fans. Deezer should be just as open about its own effects.
Sources
- Deezer achieves profitability in FY25 as strategy delivers tangible results, Deezer Newsroom, 18 Mar 2026
- Artist remuneration, Deezer
- Universal Music Group and Deezer to launch the first comprehensive artist-centric music streaming model, Deezer and Universal Music Group, 6 Sep 2023
- Deezer: first royalty distribution with artist-centric model, Sacem, 12 Sep 2025
- Deezer and Sacem partner for fairer remuneration of publishing rights with the artist centric streaming model, Deezer Newsroom, 15 Jan 2025
- Deezer reports strong FY24 results, achieving positive free cash flow for the first time, Deezer Newsroom
- Modernizing our royalty system, Spotify for Artists
- Track monetization eligibility, Spotify Support
- Deezer has deleted 26m 'useless' tracks since it launched artist-centric model with Universal Music Group, Music Business Worldwide, 14 Mar 2024
- Believe, TuneCore decry UMG-Deezer "reverse Robin Hood" artist payouts, Hypebot, Sep 2023
Sources
- Deezer achieves profitability in FY25 as strategy delivers tangible results, Deezer Newsroom (primary source). Accessed 23 Oct 2026.
- Artist remuneration, Deezer (primary source). Accessed 23 Oct 2026.
- Universal Music Group and Deezer to launch the first comprehensive artist-centric music streaming model, Deezer and Universal Music Group (regulatory filing via DILA) (primary source). Accessed 23 Oct 2026.
- Deezer - First royalty distribution with artist-centric model, Sacem (primary source). Accessed 23 Oct 2026.
- Deezer and Sacem partner for fairer remuneration of publishing rights with the artist centric streaming model, Deezer Newsroom (primary source). Accessed 23 Oct 2026.
- Deezer reports strong FY24 results, achieving positive free cash flow for the first time, Deezer Newsroom (primary source). Accessed 23 Oct 2026.
- Modernizing Our Royalty System to Drive an Additional $1 Billion toward Emerging and Professional Artists, Spotify for Artists (primary source). Accessed 23 Oct 2026.
- Track monetization eligibility, Spotify Support (primary source). Accessed 23 Oct 2026.
- Deezer has deleted 26m 'useless' tracks since it launched artist-centric model with Universal Music Group, Music Business Worldwide. Accessed 23 Oct 2026.
- Believe, TuneCore decry UMG-Deezer "reverse Robin Hood" artist payouts, Hypebot. Accessed 23 Oct 2026.
Researched with AI assistance, written and checked by WATMM Desk.