Zhou Yahui, the low-profile founder of Kunlun Tiangong, has stepped into the public eye to outline a bold strategic pivot: the company will concentrate on building an AI-native music platform overseas rather than fighting for share in China’s bruising tech duopoly. Zhou says his team’s latest music model, Mureka V8, marks a step-change in capability and gives the company a six-month window to seed creator and consumer ecosystems before rivals catch up.
Mureka V8, Zhou claims, outperformed Suno V6 in large blind tests and addresses long-standing shortcomings in AI-composed music such as structural coherence and expressive nuance. The improvements stem from larger training datasets, scaled-up model architectures and an upgraded reinforcement-learning pipeline that refines the reward model behind composition. Kunlun Tiangong plans to stitch those technical gains into product and distribution: a creator-focused Studio, a C-end ad-supported app patterned after social discovery platforms, and APIs for commercial use.
Rather than trying to go head-to-head inside China with ByteDance and Tencent, Zhou argues it makes strategic sense to cooperate domestically and compete overseas. “In China it’s too cutthroat — fighting ByteDance and Tencent is like fighting Google and Meta abroad,” he told local technology reporters. The company plans to partner with Chinese streaming incumbents such as NetEase Cloud Music and QQ Music at home, while taking its consumer-facing AI music service to markets where Spotify is the dominant incumbent.
Zhou and his engineers set out a four-step plan to create an overseas “AI Spotify”: advance the core models, validate content forms and quality with users, build recommendation and interaction systems that let users co-create, and establish partner-ready monetisation and revenue-sharing mechanics. They argue AI music is not merely a faster way to produce tracks but the seed of a new media category that can broaden listening audiences and raise overall royalty pools if handled correctly.
The timing matters. Zhou says 2026 represents an inflection point: large generative models are now able to produce short-form audiovisual content at scale, making the rapid production of 30–60 second music-driven short dramas feasible. That predicted wave could create fresh demand for easily produced, highly shareable soundtracks and seed an ecosystem of micro-creators, much as streaming once did for recorded music.
Zhou frames the competitive landscape bluntly. He does not believe China’s AI-first start-ups will become Super Apps able to match incumbent platform depth; only the major tech groups have the “thickness” — scale, distribution and embedded services — to claim that status. He names Zhipu, KIMI, MiniMax and Baichuan as AI-native firms he respects and urges them to stay the course. On the global stage, he offers a mixed read on OpenAI: uneven performance in 2025, but recent model updates like 5.2 have shown resilience against rivals such as Anthropic and Gemini.
Kunlun Tiangong’s public roadmap also touches on thorny legal and cultural issues. Zhou acknowledges that copyright, attribution and the broader legal definition of AI-generated works remain unsettled internationally, but treats regulation as an evolving backdrop rather than a showstopper. He points to early adoption by professional musicians who use the tools to augment workflows — uploading demos and using AI for arrangements — as evidence that the technology is already being integrated into the industry’s creative process.
If Kunlun Tiangong can deliver on quality and build compelling discovery mechanics, it could change the supply-demand calculus of global music. Large-scale AI production promises to flood the market with tracks and reduce costs for advertising-driven distribution, potentially expanding total listenership beyond current streaming limits. Yet that same abundance raises questions about discoverability, compensation, and the role of human artistry in a world of algorithmic abundance.
The company’s choice to pursue growth overseas rather than battle domestic giants highlights a broader strategic logic in China’s AI sector: when local competition is dominated by a few massive platforms, scale-seeking start-ups may find faster growth and fewer political and commercial frictions by focusing on international markets. How Western incumbents and regulators respond to the arrival of high-quality, AI-generated music at scale will be the decisive factor in whether an “AI Spotify” becomes a disruptive force or simply another content pipeline.