The online music industry is currently witnessing a remarkable phenomenon: the rapid ascent of Soda Music. The platform has experienced two major highlights since its inception: first, its initial launch positioned it as a disruptor in the music sector, and second, its monthly active users have now surged to levels approaching those of
Soda Music was launched quietly in 2022, a period marked by stricter regulation of domestic internet platforms and a policy focus on anti-monopoly measures. In its early stages, the product emphasized vertical music videos, mimicking TikTok's swipe-to-change-song feature, yet it refrained from large-scale promotional campaigns, allowing for organic growth. Operationally, aside from active updates on its Douyin account, Soda Music's presence on social media channels like Weibo and official WeChat accounts remained nearly dormant. For the first few years, it maintained a modest profile, far from becoming a significant concern for giants like
How, then, has Soda Music managed to achieve such rapid growth in just over a year? Whose established order is being disrupted by its rise?
What Drives Soda Music's Swift Growth?
According to data from research firm QuestMobile, as of the third quarter of 2025, Soda Music's monthly active users have surpassed 120 million, closely rivaling
A comparative analysis reveals that while
A pivotal turning point occurred in 2024. As reported by Tech Planet, ByteDance implemented a crucial copyright bundling strategy. When Douyin's copyright agreements with rights holders were up for renewal, the company stipulated that if licenses were extended to cover both Douyin and Soda Music, Soda Music would pay standard copyright fees. This move effectively elevated Soda Music's copyright library to a functional baseline, amassing a catalog of 50 million songs and laying the groundwork for its subsequent explosive growth.
Soda Music further integrated AI-generated music, Douyin viral hits, covers, live versions, and works from independent musicians into its recommendation system. This approach helped fill gaps in classic copyright holdings, creating a content supply chain highly aligned with the instant aesthetic preferences fostered by short videos, effectively capturing that demand. Industry insiders suggest that Soda Music's music recommendation algorithm is currently the closest replica of the defunct Xiami Music's system, ensuring that even free users consistently discover appealing songs.
Building on this foundation, the deep integration between Douyin and Soda Music—spanning account systems, data assets, and product features like the Douyin Music Card—has further strengthened the platform's late-mover advantage. This synergy not only reduces user migration costs but also embeds the Douyin-Soda music ecosystem as a closed loop within multiple key segments of the music industry chain. From content production and distribution to user consumption and feedback, a highly interconnected structure has formed, significantly enhancing the platform's ability to consolidate industry resources and user attention.
Furthermore, the launch of the "Soda AI Music Creation Lab" in November last year, integrating functions like AI lyric writing, composition, recording, arrangement, and mixing, provides a one-stop creative environment for creators. This facilitates a shift for AI music production from technical experimentation towards scalable content supply.
Compared to traditional platforms, Soda Music, utilizing Douyin's traffic distribution mechanism and precise recommendation algorithms, has enabled a cohort of AI singers and AI music creators to rapidly build a user base. For instance, AI singers like "Ding" and "Pencil Sharpener" have attracted over 1.2 million fans solely through AI-generated content, aided by coordinated distribution across Douyin and Soda Music.
This signifies that Soda Music has constructed an efficient conversion pathway from short-form video content consumption to dedicated music listening. The integrated account systems, data interoperability, traffic diversion, promotion, and content collaboration create an ecosystem-level linkage that transforms deep listening on Soda Music into an almost subconscious daily routine for users.
Data disclosed at Douyin's 'See Music Plan' mid-year industry forum last year revealed that on the user side, Douyin's daily music video exposures exceeded hundreds of billions, with over 6,000 songs achieving annual exposures in the billions, establishing short videos as a crucial channel for music dissemination. Soda Music, in turn, caters to deeper consumption needs, with its user base growing 100% annually and average usage time per user surpassing 80 minutes.
The low-price strategy is also a critical factor: a free ad-supported tier combined with affordable membership options.
Similar to the approaches of Tomato Novel and Hongguo Short Drama, Soda Music employs a membership exchange system based on user time investment. Users receive a free membership upon initial login without any payment. Subsequently, they can exchange 80 seconds of ad watching for 24 hours of membership benefits, with the option to accumulate time.
Regarding VIP membership pricing, Soda Music remains at the lower end of the industry spectrum—approximately 8 yuan for a standard monthly card and 88 yuan for an annual card. This contrasts with QQ Music's 18 yuan monthly card and 168 yuan annual card, and
Thus, Soda Music's rapid rise against the trend in just over a year is the result of a multi-dimensional strategy encompassing copyright, product, and operations, collectively fueling its explosive growth from 2025 to 2026.
Online Music at a Crossroads The landscape of China's online music market, shaped by over a decade of copyright battles and capital maneuvering, began to shift away from the pure "copyright is king" competition model only with the emergence of Soda Music.
In 2016, Tencent announced a partnership with China Music Corporation, merging QQ Music, KuGou Music, and KuWo Music into
In the ensuing years, the music industry entered a phase deeply intertwined with copyright monopolies and capital operations. Baidu Music and Xiami Music gradually fell behind in this intense copyright competition. During this period,
Copyright barriers once served as the core moat for streaming platforms until regulatory changes promoting the adjustment of exclusive copyright strategies and encouraging library openness created real entry opportunities for new players.
In 2022, Soda Music entered the scene post-anti-monopoly, leveraging its inherent synergy with the Douyin ecosystem and introducing a fundamentally different product logic into this relatively stagnant market.
In terms of strategy, Soda Music abandoned the old path of "competing for copyrights and exclusives." Instead, it adopted a new approach, aiming to foster music growth within the platform itself. By utilizing Douyin's vast UGC ecosystem and revenue-sharing mechanisms, along with initiatives like the "See Music Plan" and "Hello Musician Support Program" to attract creators, and叠加 experiments with AI-generated music, the platform gradually shifted towards internal content production.
Compared to the traditional model reliant on existing copyright systems, Soda Music is more akin to building a content circulation system where music is not only distributed but also produced, amplified, and recreated within the platform.
Clearly, the Chinese online music market is now exhibiting three coexisting and distinct development models, arriving at a crossroads of era differentiation.
The first is the copyright and ecosystem model represented by TME. Relying on traditional copyright accumulation, a massive library advantage, and a mature commercial system (membership fees, concerts, merchandise, etc.), it maintains stable growth with content resources as its core competitiveness.
The second is the community and aesthetic model represented by
The third is the integrated content generation and distribution path represented by Soda Music. Leveraging the traffic synergy of the Douyin ecosystem, the platform no longer merely hosts existing music but enables continuous internal music production through UGC conversion, creator incentives, and AI generation, followed by highly efficient algorithm-based distribution.
The current "new tripartite balance" in online music competition differs from the past linear contest based solely on library size and copyright scale. It resembles a clash between different value orientations and product logics. On one end is the traditional system centered on copyright and resources; on the other is a new path driven by AI algorithms and distribution efficiency. Determining the ultimate victor remains an intriguing question.
Looking deeper, the true battle is no longer over music itself, but over capturing the user's moment of engagement with music. In the ever-changing digital consumption environment, whoever grasps the user's emotions and decision-making chain earlier is more likely to retain them.
The AI Era: How Are Platforms Evolving? Every transformation in the music industry fundamentally revolves around "scarcity." What changes with the times is the role and raison d'être of platforms.
In the唱片 era, music production costs were high, and distribution channels were limited; content itself was scarce. In the streaming era, the internet enabled near-infinite replication and dissemination of almost all music; scarcity shifted to distribution and selection, with platform value lying in helping users discover their next favorite song.
For over a decade following the rise of streaming, this core logic remained largely unchanged. Users pay a fixed monthly fee for unlimited music access; platforms reduce choice costs through playlists, algorithms, and recommendation systems. This model, since Spotify's official US launch in 2011, has been the engine of music industry growth.
However, signs of this growth slowing globally are now apparent. According to IFPI's "Global Music Report 2025," streaming revenue reached $20.4 billion, a 7.3% year-on-year increase. This growth has gradually slowed compared to 2021 (23%), 2022 (11%), and 2023 (9.8%). Concurrently, paid subscription streaming revenue, the mainstay of income, grew by 9.5%, down nearly 2 percentage points from 11.2% in 2023. Consequently, the growth rate of global recorded music revenue plummeted from 10.2% in 2023 to 4.8% in 2024, halving and marking the lowest annual growth rate since 2015, indicating that streaming is entering a distinct plateau.
The challenge today is that while we recognize streaming's maturity, the next viable model remains unclear. However, with the explosion of AI music, both the platform and supply sides are outlining an answer.
On one hand, changes are occurring at the surface level of platform capabilities. Domestically, for example, the AI singer "Ding," created by KuGou Music's Apollo Sound Lab, boasts over 27 million monthly listeners on a single platform. Tencent's Yuanbao Pai has also integrated with QQ Music's member content library, and QQ Music has added AI accompaniment features. Within the TME ecosystem, the Morning Star Studio has formally incorporated AI music into its distribution pipeline.
In 2025,
Internationally, Spotify recently disclosed its internal AI tool, Honk, built on Anthropic's Claude model, capable of quickly fixing platform bugs and adding features. Notably, following the release of Claude Opus 4.5, Spotify's top engineers have completely ceased writing code manually, focusing solely on supervising AI-generated code, marking a revolutionary change in the platform's R&D model.
In terms of content strategy, Spotify categorizes AI music into "AI Original New Songs" and "Music AI Derivative Works," focusing efforts on the latter. The platform believes derivative works like AI remixes and covers represent new opportunities for artists to monetize existing IP, with the relevant technology now fully mature, limited mainly by copyright licensing frameworks.
Overall, whether through AI extensions in functionality and scenarios,布局 in recommendation and creation ecosystems, or even at the R&D level, a common trend is evident: AI is progressively moving into a central position within music platform ecosystems, becoming a new core connecting products, content, users, and social relationships.
On the other hand, changes are happening at the foundational structure of content supply. Historically, music supply was constrained by creative ability and production costs. Now, anyone can generate music. According to internal investor materials, Suno users generated approximately 7 million musical works daily last year, equivalent to producing a volume of music every two weeks that matches the entire scale of Spotify's catalog.
Furthermore, data from Deezer indicates the platform receives over 50,000 fully AI-generated songs daily, constituting about one-third of daily uploads. More notably, a significant proportion of plays associated with this content are flagged as anomalous or fraudulent, indicating that AI content not only increases supply but also directly impacts the platform's revenue distribution ecology.
Simultaneously, a vast amount of music is effectively never heard. Luminate's year-end report data shows that although net new ISRCs (International Standard Recording Codes) in 2025 increased by 37.9 million compared to 2024, 86% of these new tracks garnered annual plays concentrated at ≤100 plays. Approximately 88% of tracks had annual plays of 1,000 or below (falling short of Spotify's minimum threshold for royalty payments).
When music can be generated infinitely, library size ceases to be a competitive advantage, and the marginal value of recommendation algorithms begins to diminish. What becomes truly scarce for users is time, attention, and trust in the source of content.
In this context, platforms are thrust into an unavoidable role, requiring them to assume new responsibilities: deciding which content deserves visibility, what requires labeling, which acts constitute creation versus mere system exploitation, and ultimately, where revenue should flow. These are currently unresolved dilemmas.
History has proven that when platforms relinquish governance, leaving everything to growth and traffic, ecosystems often spiral into失控 in short order.
The former social platform MySpace is a prime example. In 2006, it was the most imaginative breeding ground for independent music, adding 230,000 accounts daily. Music became MySpace's defining cultural hallmark, even altering the career trajectories of artists like Lily Allen and Arctic Monkeys. However, it was ultimately consumed by spam, fake engagement, and fraudulent information, leading to its decline.
With this precedent, and given AI's aggressive arrival, platforms no longer have the luxury of remaining "neutral."
Conclusion As Marshall McLuhan stated, "We shape our tools, and thereafter our tools shape us."
Reflecting on the industry's changes over recent decades reveals that each technological leap alters how people access music.
Now, AI is pushing this into a new phase.
As AI becomes integrated into the user experience itself, discussions surrounding Soda Music, AI music, and copyright boundaries fundamentally transcend the得失 of any single platform. They resemble a transitional period for the entire music industry, where old operational models are loosening their grip, and new rules and boundaries have yet to be fully established.
In the future, the fate of music streaming may not hinge on library size or traffic volume, but rather on their ability to establish more credible curation, clearer rules, and more stable distribution systems in the AI era.