Yuewen’s AI Pivot Is Only Halfway There

💡Yuewen’s AI content revenue is surging, but falling users and margins reveal the harder monetization problem.
⚡ 30-Second TL;DR
What Changed
Yuewen’s first-half revenue rose 10.7% to RMB 3.53 billion, while online business revenue fell 7.3% and copyright-related revenue grew 40.3%.
Why It Matters
For AI content companies, Yuewen demonstrates that lowering production barriers can expand IP supply without improving unit economics. Platforms must measure incremental revenue, cannibalization of legacy products, content cost per successful title, and user-retention effects rather than relying on output volume.
What To Do Next
Run a cohort experiment comparing AI-translated or AI-adapted titles with human-produced controls, tracking incremental revenue, retention, production cost, and cannibalization of paid reading.
Key Points
- •Yuewen’s first-half revenue rose 10.7% to RMB 3.53 billion, while online business revenue fell 7.3% and copyright-related revenue grew 40.3%.
- •Short dramas and AI comics generated more than RMB 430 million, over three times the prior-year figure, but the company does not separately disclose AI-comics revenue.
- •WebNovel has published more than 30,000 AI-translated works, which account for 40% of platform novel revenue; small-language revenue grew 160% year over year.
- •Online monthly active users fell 5.1% and average monthly paying users fell 10.9%, showing that AI is both a production accelerator and a competitive threat.
- •Yuewen’s adjusted EBITDA margin declined from 12.1% to 10.9%, indicating that revenue growth has not yet translated into stronger profitability.
🧠 Deep Insight
AI-generated analysis for this event.
🔑 Enhanced Key Takeaways
- •Yuewen has integrated its 'Yuewen Miaobi' large language model directly into the writer's backend, providing features like character profile management, world-building assistance, and automated scene generation to lower the barrier for new authors.
- •The company's strategic shift toward 'IP-centric' operations involves a 'Global IP Operation' strategy, which leverages AI to accelerate the adaptation of Chinese web novels into international formats, specifically targeting Southeast Asian and North American markets.
- •Yuewen has established a dedicated AI-driven 'Short Drama' production pipeline that utilizes generative AI for storyboard visualization and rapid script iteration, significantly reducing pre-production costs compared to traditional methods.
- •Despite the push into AI, Yuewen faces significant regulatory headwinds in China regarding the copyright status of AI-generated content, which has created legal uncertainty for their 'AI-comics' and 'AI-translation' business models.
- •The decline in monthly active users (MAU) is partially attributed to a strategic decision to phase out low-quality, ad-supported content in favor of high-value IP development, a move intended to improve long-term monetization efficiency.
📊 Competitor Analysis▸ Show
| Feature | Yuewen Group | Chinese Literature Competitors (e.g., Fanqie) | International Platforms (e.g., Wattpad) |
|---|---|---|---|
| Core Model | IP-centric/Premium | Ad-supported/Free-to-read | Community/User-generated |
| AI Integration | Deep (Writing/Translation/Comics) | Moderate (Recommendation/Editing) | Low (Basic writing tools) |
| Monetization | Subscription/IP Licensing | Advertising/In-app purchases | Subscription/Brand partnerships |
| Content Strategy | High-end IP/Adaptations | High-volume/Viral content | Diverse/Niche genres |
🛠️ Technical Deep Dive
- Yuewen Miaobi Model: A proprietary large language model fine-tuned on Yuewen's massive corpus of web literature, specifically optimized for long-form narrative coherence and character consistency.
- AI Translation Pipeline: Utilizes a hybrid approach combining neural machine translation (NMT) with human-in-the-loop (HITL) post-editing to maintain cultural nuances in web novel localization.
- Generative Comics Workflow: Employs a multi-stage pipeline where LLMs generate script-to-image prompts, followed by Stable Diffusion-based models for character consistency and automated panel layout generation.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: 虎嗅 ↗

