Deezer has reached a milestone that many considered distant for the French platform: it is now making money, at a time when music streaming is overwhelmed by a wave of AI-generated content. In the first half of 2026, the group reported a net profit of â¬6.7 million, compared to a loss of â¬7.6 million a year earlier. At the same time, it receives nearly 90,000 AI-created tracks every day, accounting for more than half of new deliveries during peak periods.
This contrast tells a story that goes beyond mere financial recovery. It reveals a global battle for the value of music: who deserves to be recommended, compensated, and made visible when machines can produce thousands of tracks in a matter of hours? Deezer, a player much smaller than Spotify, Apple, or Amazon, is trying to turn this threat into a territory of differentiation.
A Profitability Finally Visible in the Accounts
The results published at the end of July show a consolidated revenue of â¬268.2 million in the first half, a modest increase of 0.4% year-on-year. Overall growth appears modest, but its composition is more revealing. Direct subscription revenue grew by 6.7% to reach â¬185.2 million, driven by 3.9 million subscribers in France, an increase of 8.4%, and a 9.2% rise in subscribers worldwide.
Adjusted EBITDA reached â¬8.5 million, compared to â¬2.1 million in the first half of 2025. Operating income improved from a loss of â¬7 million to a profit of â¬6.5 million. It is this combinationâgrowth in direct subscriptions, cost discipline, and better marginsâthat allows Deezer to speak of sustainable profitability rather than a mere accounting accident.
However, caution remains necessary. An analysis released by Boursorama highlights that a significant portion of the cash flow improvement comes from an exceptional effect on working capital. In other words, the trajectory has improved significantly, but the â¬37.4 million in available cash flow cannot be mechanically extrapolated to every semester.
90,000 AI Tracks Per Day: The Figure That Changes the Industry
The real earthquake lies in the catalog. Deezer reports receiving nearly 90,000 AI-generated tracks daily. During peak periods, they represent over 50% of new titles delivered each day. In April, the reported proportion was still 44%, or about 75,000 tracks. The growth is therefore extremely rapid.
This abundance does not mean that half of the music listened to is artificial. Tracks entirely generated by AI account for only a small share of listens. However, their volume can saturate distribution systems, manipulate recommendations, and dilute the compensation available for human artists. A report from the National Assembly reminds us that in 2025, up to 85% of listens associated with 100% AI-detected tracks on Deezer were fraudulent, often produced by bots.
The issue is therefore not only aesthetic. It is economic. When a fraudster automates both the creation of songs and their listening, they can divert a portion of the royalty pool without building a real audience. On a large scale, this mechanism threatens independent artists, producers, songwriters, and labels that already rely on very low unit revenues.
Deezer Transforms Its Defense into a Global Product
Deezer’s response relies on an internally developed tool that searches for acoustic markers of artificial generation. Identified albums are labeled and excluded from algorithmic recommendations as well as editorial playlists. The platform has also made a detector available to the public in 27 languages, capable of analyzing playlists from about twenty major services.
This technology is now becoming a standalone commercial activity. Rights management organizations in Hungary and the Netherlands have acquired a license. BumaStemra can thus use the tool to better protect the rights of its 38,000 members. Deezer is also integrating detection into its business offerings, alongside white-label music services and advertising solutions.
This is a smart strategic choice: instead of competing solely on catalog size or subscription price, the French company seeks to sell trust. If synthetic music continues to grow, the ability to distinguish between human, hybrid, or entirely artificial content could become as important as the recommendation itself.
A Useful Technology, But Not an Absolute Truth
However, the detector is not a machine for certifying artistic authenticity. Manuel Moussallam, head of research at Deezer, explained to TechRadar that the system detects the presence and intensity of markers associated with AI generation. It does not always know which part of a song was created by a human, whether the lyrics are original, or if only certain instruments are synthetic.
This nuance is crucial. Many artists already use AI as a production, mixing, or experimentation tool without delegating the entirety of creation to it. A too rigid filter could penalize these hybrid works. Conversely, a too permissive system would allow automated content farms to slip through. The next technical battle will therefore take place at the level of the different tracks of a pieceâvocals, drums, instruments, texturesâand not just on the final file.
The France Point: A Cultural and Industrial Weapon
For France, Deezer’s success goes beyond the symbol of a technology company that has become profitable again. The country has a powerful musical ecosystem, from Sacem to Ircam, including labels, festivals, and distribution platforms. The National Assembly has already addressed the risks of revenue dilution, transparency, and protection of works in the face of generative AI.
Deezer is also supporting twelve Francophone festivals in 2026, eleven of which are in France. This on-the-ground presence reminds us that streaming cannot be reduced to a battle of algorithms. Its value still depends on artists capable of creating a community, taking the stage, and circulating culture.
A Real-World Test for All Streaming
The new phase of Deezer will be observed well beyond Paris. The company must prove that it can continue to gain subscribers, develop its partnerships, and monetize its tools without turning the fight against AI into a mere marketing argument. It will also need to publish sufficiently transparent results for artists to understand which content is filtered and on what criteria.
But the signal is already powerful. A French platform long considered an outsider has regained profitability at a time when the industry is desperately seeking rules for the era of generative music. If Deezer manages to make authenticity, transparency, and rights protection a profitable model, it will not only survive against the giants. It could help define the global standards of the next age of streaming.
