Original · Unique Research · 2026-03-26 · Shanghai
Editor's note: This is a historical account of remarks made at the March 26 summit. Model-market assessments, collaboration examples, demand descriptions and efficiency figures are attributed to the original report and speakers, not independently validated results or current product announcements. The 90% and 20-fold figures describe claimed work or efficiency effects, not measured workforce reductions or guaranteed business returns. Views about copyright and AI relationships are reported positions, not legal determinations or evidence of AI consciousness; the source's safeguards for real-world relationships are preserved. Several names are provisionally romanized, and the source does not define opc/bcp. All extracted article text is translated here; substantive source-image content remains to be checked before publication.
Unique Awards · Hangzhou AI Week
As AI Begins Taking Over Entertainment's Supply Side,
This Hangzhou Summit Examined What Comes Next
3,000+ Registrations as the AI Entertainment Industry Enters a New Stage
On March 26, during 2026 Unique Awards · Hangzhou AI Week, the Unique Content AI Entertainment Summit took place in Hangzhou's Binjiang district.
Organized by Unique Research and co-organized by 43Talks, the event brought together practitioners working across models, tools, content, platforms, international expansion and investment. More than 3,000 people registered, and AI companies, founders and investment institutions were among the groups represented on site. The discussion had moved beyond “Can AI make content?” to a more practical question: as AI genuinely enters entertainment's supply side, how will the industry be rewritten?
From video generation, music and audio to animated series, short dramas, companionship and social interaction, global content, artistic boundaries and monetization, the summit put nearly all of today's central AI entertainment questions on the table. More importantly, the focus had clearly shifted from “Are models getting stronger?” toward “How do we produce content at scale, build IP, expand overseas and generate revenue?”
Starting with a Broader Assessment: China's AI Entertainment Industry Is Moving from Catching Up to Setting the Direction
Opening the summit, Jason, head of research at Unique Research, presented the “2026 AI Entertainment Industry Trends Report.” His central assessment was clear: China's AI entertainment industry is gradually moving from being a follower to helping define important new paradigms for global content production.
In his account, OpenAI's Sora appeared to be undergoing a strategic retrenchment rather than simply exiting. In parallel, Chinese companies were continuing to move up the field in entertainment models such as video and music: ByteDance's Seedance 2.0, Kling 3.0 and Kunlun Tech's Mureka V8 had all entered the leading global competitive group. Meanwhile, domestic products' user metrics were still growing rapidly. Chinese teams were no longer competing over “whom to replace,” but over who would define the next generation of content-supply standards.
Jason further argued that in 2026, competition would hinge not merely on model image quality and isolated effects, but on industrial-scale content production capabilities. Character consistency, narrative stability, cost efficiency and controllable workflows were becoming the variables that genuinely determined implementation speed. At the same time, AI content was moving beyond generation alone into a new stage of “generation + interaction.” Emotional companionship, digital life forms and interactive content were emerging, rewriting IP life cycles and the logic of platform competition.
Beyond these trends, he also stressed the other side that the industry must confront: copyright, compliance, aesthetics and narrative competition will not simply disappear. The future winner will not be one isolated model, but a system combining models, workflows, IP and an end-to-end commercialization loop.
Aishi Technology's Answer: Large Video Models Must Ultimately Reach Deep into Industries
In the next keynote, Sun Weizhe, head of enterprise services at Aishi Technology, presented “Growing Upward: Expanding the Horizons of Imagination in the Era of Visual Storytelling.”
His presentation clearly represented how some leading companies now think: if large video models remain confined to technology circles, they can quickly become exercises in self-congratulation. What will determine the next competitive landscape is whether they can reach deep into industries and real business settings.
Sun Weizhe reviewed Aishi Technology's large-video-model work: from entering model training early, through successive generations of models, to a multilayer structure comprising V-series foundation models, C-series models specialized for animated series and film and television, and R-series real-time models. Aishi Technology was seeking to build not an isolated product, but a complete technical foundation for industrial-scale production.
He emphasized three directions for Aishi Technology's next steps: further strengthening model stability and capabilities; deepening its presence in key settings such as animated series, short dramas, film and television, marketing and e-commerce; and building a tiered creator ecosystem that feeds creators' experience back into the models.
Notably, the examples he offered were no longer purely conceptual. Collaborations with Mango TV, iQIYI, JD.com and Zhangyue, alongside implementations in film and television generation, finished marketing advertisements and batch production of e-commerce materials, pointed to a deeper phase of application. Sun Weizhe's closing position was clear: no single company will build the future of AI video alone. Model companies, creators, platforms and industry customers must work together to keep pushing the boundaries of visual storytelling.
The Discussion of Text and IP Addressed Not “How to Make Content,” but “What Content Is Still Worth Making”
The first trends roundtable, “Text and IP: AI-Driven Content Creation and the Reconstruction of IP Value,” was moderated by Huang Lin, head of business at StoReel. Panelists were digital artist An Shi; Sandra, founder of Xuyuan Yanjin; David Ju Dawei, founder and CEO of ADGo; and Yubo, founder and CEO of YouMind.
The discussion was particularly interesting because the speakers did not agree with one another.
From an artistic perspective, An Shi urged caution: as AI technology matures, content supply will indeed expand dramatically, but it may also quickly dilute user attention. If content becomes increasingly similar and cheap, large quantities of realistic human-like content will not necessarily have inherent commercial value. She was more optimistic about stylized, artistic and highly distinctive work.
From the perspective of tool development and application engineering, Sandra added another reality: AI is changing supply-side efficiency, but consumer demand has not changed in tandem. Commercial content may therefore have only two paths forward—extreme industrialization or exceptional craftsmanship. Content that resonates with the mindset of mainstream consumers is what can endure.
David Ju Dawei brought the conversation back to commercial marketing. He saw the possibility of AI advertising moving from “different content for different people” to “one or multiple versions for each individual.” Content production, brand placement and integration with user data could all become part of an automated creative process.
Yubo's view was the most provocative. He argued that current AI creative tools remain early-stage and are not easy to use. At the same time, he declared that “IP is dead”—at least in some fields, large, unified brand IP in the traditional sense is being replaced by more personalized and segmented patterns of demand.
The discussion did not produce a single shared answer,
but it clarified the question:
The AI era is not simply about “more content.”
The value of content itself is being reassessed.
Changes in Music and Audio Are More Practical Than Many Imagine
The second trends roundtable, “Music and Audio: AI-Driven Industrialization of Content Production,” was moderated by Lu Xiaoxu, CEO of Xiaoxu Music. Panelists were Jiang Tao, CTO of Ziyou Liangji; Zhao Wenxiao, co-founder of ACE Studio; Song Kaifa, co-founder and COO of VMEG; and Duan Ruilei, head of the Music Engineering Department at Zhejiang Conservatory of Music.
The discussion broke the AI audio industry into fine distinctions: large music models, music-creation tools, AI voice products for overseas markets and music education are almost four separate markets.
Jiang Tao introduced Ziyou Liangji's Chinese-developed large music model “Yinchao 3.0,” emphasizing its closer fit with popular music creation and film, television and gaming use cases. Drawing on ACE Studio's professional creative workstation, Zhao Wenxiao discussed the significance of open-sourcing full music models for the creator ecosystem. Song Kaifa shared VMEG's overseas experience in AI speech translation and dubbing, particularly the continuing scale of demand for lifelike multilingual voice production. Duan Ruilei returned to education, discussing AI's role in music learning and teaching.
More importantly, the panelists did not avoid the most practical issues: copyright, disruption to occupations and adaptation in education.
On copyright, they generally favored recognizing users' rights and interests in AI-created music while acknowledging the importance of copyright governance for training data. On occupational replacement, their conclusion was not extreme: lower-end, standardized dubbing without strong emotional character would be affected earlier, but high-quality vocal work requiring strong performance remained difficult to replace completely. In education, AI was more an aid than a substitute, because learning music ultimately still depends on aesthetics, understanding and emotional communication.
To summarize the discussion in one sentence:
The AI audio industry is not waiting for a single answer.
It is accelerating industrialization within its separate niches.
Companionship and Social Products Are Addressing a More Sensitive, Practical Question: Is AI a Tool or a Participant in a Relationship?
The afternoon's first roundtable, “Social Interaction and Companionship: AI Innovation in Emotional Connections within Human Relationship Networks,” was moderated by Jerry Huang Jingrui, VP at Unique Capital. Panelists were Jiang Yizhuo, founder and CEO of Yuejie Chuangwan; Ma Jianjun, founder and CEO of Tuike AI; Kevin, co-founder of Huihua Intelligence; Li Kechen, founder of Halo; and Cheng Zhenhua, founder and CEO of Lingda Intelligence.
What stood out was not that everyone spoke about “emotional value,” but that several speakers had clearly moved beyond discussing AI as a chat companion toward treating it as a participant in an emotional relationship.
Whether it was the deeply immersive emotional involvement Tuike AI observed in overseas social data, Huihua Intelligence's companionship relationships for older adults, or Lingda Intelligence's use of light content such as astrology and fortunes to guide users back toward real-world social interaction, the examples suggested that AI was no longer merely an intermediary: it was entering real relationship networks.
The panelists were broadly positive about the trend, but all placed strong emphasis on boundaries.
One important consensus was that AI emotional products must have an “emotional firewall.” Its design should not abruptly cut people off, but guide them gradually back toward real-world relationships.
In other words, the industry did not deny that users could form deep emotional connections with AI. Yet almost everyone believed AI should not replace real-world intimate relationships; it should be a bridge that helps repair, extend and maintain them.
Commercially, the speakers also agreed that essential emotional needs are the real driver of payment. Users are willing to pay for emotional value, but motivations vary greatly across groups; one uniform product cannot serve everyone.
Looking further ahead, the discussion touched on a deeper trend: future social networks may not be only networks of “people and people,” but also of “people and AI,” “AI and families,” and “AI and relationship assets.” The idea may still sound new, but people at the summit were already working seriously on it.
The Boom in Animated Series and Short Dramas Will Require More Than Efficiency: Content Differentiation Will Ultimately Decide the Outcome
The next roundtable, “Animated Series and Short Dramas: AI-Enhanced Creative Expression and Market Potential,” was moderated by Yi Yating, head of marketing and branding at 43AI and founder of Yilun AI. Panelists were Huang Chujie, founder of Aimen Weisi; Zhou Zhipeng, co-founder of Shuimu Intelligence; Yang Xiaofeng, founder of Huanzhou AI; and Luo Yuan, chief director at Shanzhu AI.
This was a particularly direct discussion because all the speakers were doing frontline business work.
Huang Chujie discussed their work in regional content and commercially customized AI livestream dramas. Zhou Zhipeng shared Shuimu Intelligence's exploration of AI animated series and creator tools. Yang Xiaofeng focused on industrial-scale AI drama production—how AI could do 90% of the manual work and thereby empower highly leveraged individuals. Drawing on experience with overseas short dramas and hit original productions, Luo Yuan discussed platform rules, emotional pacing and content methods.
They all acknowledged that AI had significantly increased production capacity. But greater capacity does not automatically make the content market better.
The real keyword was therefore “how to escape sameness.”
Huang Chujie's approach was personalized, highly customized content, linking character libraries with IP. Zhou Zhipeng emphasized data-driven work and premium IP, particularly science fiction, fantasy and historical genres. Yang Xiaofeng focused on a “personal memory system” and support for creators throughout the entire process, aiming to leave repetitive labor to AI and real creativity to people. Luo Yuan reminded the audience that short drama is ultimately an emotion-driven industry: it needs memorable moments and emotional pull, not production efficiency alone.
The panelists' outlooks were also remarkably consistent: AI today mainly enhances existing models, but the real inflection point will arrive when interactive, real-time-generated and individually personalized content matures. At that point, the organization and business models of film and television content will be redefined.
AI Tools Are Not Replacing Creators; They Are Rewriting the Content Industry's Efficiency Curve
“Creation Is No Longer a Test of Stamina: How AI Tools Reconstruct Content Production's ‘Efficiency Curve’” was moderated by Duan Hongyu, partner at Unique Research. Panelists were Chen Weiyu, deputy general manager of technology and head of AI at China Literature; Liang Wei, co-founder of MovieFlow; Yan Huapei, founder of Shanjian Intelligence; Zhao Lu, China president of Xufeng Technology; and Zhong Xiangyu, Senior Product Engineer at MeshyAI.
The roundtable resembled a comprehensive examination of the content-production toolchain.
Zhao Lu emphasized that AI's central value is freeing creators from “the manual labor inside mental work.” In settings such as real estate, cultural tourism, gaming and robotics, pure efficiency gains could even reach 20-fold. Yan Huapei and Zhong Xiangyu, drawing on 3D printing, games and film, and 3D generation, described how AI tools were already producing real efficiency gains in areas such as concept design and turning ideas into physical objects.
Chen Weiyu's contribution was important because it brought the conversation back to premium content. AI tools can reduce the cost of scriptwriters' early information organization, style training and image generation, but story ideas, the ability to build settings and production aesthetics remain central to high-quality content.
Liang Wei broadened the perspective further. He argued that AI would penetrate more deeply into traditional film and television production, reshaping everything from preparation and asset construction to shooting and generation. A minimal production unit comprising a director, a screenwriter and an editor could become the norm.
The central consensus was therefore clear:
AI is not here to bring creators to an end.
It is here to reshape the efficiency curve.
It frees creators from large amounts of low-value repetitive work. But the higher one moves, the more important creativity itself becomes.
Overseas Expansion Is No Longer an Additional Option, but AI Content's New Main Arena
“A New Window for Global Content: How AI Enables Animated Series and Short Dramas to Be ‘Born Global’” was moderated by Abner Zhao Liang, partner at Unique Capital. Panelists were Liu Meiling, business director at SeaArt AI; Ding Wei, AIGC content director at Xinshijue; Wang Jian, CTO of Jingying Technology; and Tong Chao, co-founder of MOOTION.
The panelists all pointed to a shared backdrop: AI visual content is currently expanding overseas during a distinctive window of opportunity.
Model capabilities are advancing rapidly, global distribution platforms are maturing, and users' attitudes toward AI content are shifting from looking at the technology to looking at the content itself. Technology, channels and users are beginning to align at the same moment, allowing many forms of content to be genuinely “born global” for the first time.
But that does not mean a single general-purpose product can conquer every market.
Liu Meiling shared SeaArt AI's experience testing multilingual and locally relevant themes in different regions. Tong Chao emphasized keeping tools general-purpose while adapting operations and content guidance locally. Wang Jian argued that as AI lowers technical barriers, creativity and taste will become the real differentiators. Ding Wei urged content companies to use the window to secure talent and IP, create original work and build more systematic AI film and television production processes.
The most compelling part of the discussion
was how it moved from the methods of “going overseas” to the underlying logic:
The core of future global content
is not just translation and distribution.
It is the ability to identify emotions shared across humanity
and frame them within different cultural contexts.
Those who understand human nature better have a greater chance of creating global hits.
When AI Becomes the Paintbrush, Art Needs Human Responsibility Even More
“When AI Becomes the ‘Paintbrush’: Where Are Art's Boundaries Once Technology Becomes a Tool?” was moderated by Wang Chaochao, partner at Unique Capital. Panelists were Liu Rushan, founder and CEO of MulanAI; Zhou Changyin, founder and CEO of Vozo AI; Roy Cao Renyi, co-founder of Cutout.pro; Huang Naiyuan, COO of Jifan Liuxing; and Xitang, founder of Xiyangshi.
The discussion brought a technological question back to an artistic one.
The panelists all acknowledged that AI greatly lowers barriers to creation, allows more people to express themselves, and changes creative processes, role divisions and production methods. But they were equally clear that AI has not changed the essential nature of art.
What determines whether an AI-generated work can be art is not how convincing it looks, but whether it embodies clear human intent, subjective judgment and meaning capable of moving people.
A clear consensus emerged: AI content without active human choice and the attribution of meaning is difficult to regard as a genuine artwork.
This also means that as generated outputs become cheaper, the hesitation, choices, trade-offs and tensions within the process become more precious. Excellent creators will increasingly resemble directors, responsible for coordination, judgment, expression and accountability. Meanwhile, AI will take over more execution-oriented, mechanical graphic-production work.
Ultimately, AI has not erased art's boundaries; it has brought them back into focus. The stronger the technology, the more important human taste, creativity, emotion and responsibility become.
The Final Discussion on Monetization Voiced an Uncomfortable but Unavoidable Truth: Being Able to Make Something Is Not the Same as Being Able to Survive
The final roundtable, “Creation and Monetization: A New Order of Content Value and Business Models in the AI Era,” was moderated by Yi Yating, head of marketing and branding at 43AI and founder of Yilun AI. Panelists were JadeWu, founder of Guangnian Huanjing; Yanchuan, AI director at Yanchuan Culture Studio; and Jichuan, head of Jichuan AI Studio.
It was a pragmatic discussion, and the closest to the real circumstances of many creators today.
All the panelists worked on the front line and described a common reality: AI tools are getting stronger, but creators' real difficulty is not “whether they can make something.” It is “how to be seen, how to become recognizable, and how to monetize consistently.”
They therefore repeatedly highlighted “distinctive IP.” IP is not merely a character concept or a single visual style; it combines style, a world, stories and long-term accumulation. Especially as tools converge, creators' own experiences, taste and methods are becoming new sources of differentiation.
The panelists did not avoid the industry's harsher realities.
They observed that supply already exceeds demand in AI short dramas and animated series, placing growing pressure on medium-sized studios. Demand for television commercials (TVC) still exists, but cost-reduction and efficiency requirements are continually squeezing margins.
Opportunities have not disappeared; they have shifted. One lies in premium, standard-format long-form dramas. Another lies in combining interactive content with new technologies, such as AI interactive short dramas, scripted role-playing murder-mystery games, instantly generated content and new entertainment formats driven by world models. Platforms will gradually establish content tiers, and creators must break the habit of earning through old approaches and start understanding new ways to earn.
In the Breakout Venue, a More Practical End-to-End Industry Process Was Taking Shape in Parallel
The breakout venue also offered a comprehensive program.
Zhang Wei, vice president of Baidu AI Cloud and general manager of its Pan-Technology Business Division, attended the opening ceremony. Chang Jiashuo, director of AIGC industry solutions at Baidu AI Cloud, then presented “Introduction: From AIGC Short Animated Dramas to IP Derivatives—How Can Good Stories Keep Gaining Value?” Yang Yang, head of Duanju Zixishi, shared “From a Production Industry to a Content Industry: AI-Driven Industrial Upgrading.” Xing Huaifei, video-cloud technical architect at Baidu AI Cloud, focused on “Efficiency, Content and Overseas Expansion: Intelligence Across the Full Animated-Series Workflow.” Yuan Mingnan, senior AIGC instructor at Yuguang Tongchen, explained “The Full AI Creative Process, Animation Generation and Techniques for Adding Motion.” Gu Lei, director of AI film and television business at Xuanjia Technology, examined “Practical opc/bcp Applications in the AI Animated-Series Business Chain.” Jiang Tao, CTO of Ziyou Liangji, shared “AI Dubbing, Sound Effects and Immersive Audio Design.” Zhang Zhongnan, head of animated-series operations at Baidu Baijiahao, discussed “Helping Good Content Grow into Good Business: The Evolution of Baidu's Animated-Series Ecosystem and a Path to Shared Success.” Pang Xueting, business director at Mofang Animation, presented “AI Animated Series: Moving from the Traffic Era into the IP Era.” Li Wenzheng, head of LibTV ecosystem partnerships at LiblibAI, focused on “LibTV: Rules for Human–Agent Coexistence on an Infinite Canvas.” The event ended with a graduation ceremony and the induction of new members into the “Baidu AI Cloud AIGC Creator Alliance.”
While the main venue concentrated on industry trends and central questions, the breakout venue resembled a practical class covering the industry chain. It linked the issues most relevant to AI animated series today: how to make content, increase efficiency, produce audio, understand IP, build an ecosystem and expand overseas.
For that reason, the breakout venue did more than supplement the information: it translated “trends” into “methods.”
What the Summit Really Demonstrated: Competition in AI Entertainment Has Entered Its Next Stage
Looking back over the day's discussions, a clear change emerges.
In the past, discussions of AI entertainment easily became fixated on technical metrics such as model capabilities, image quality, speed and parameters. Today, the industry's real concerns have become a different set of questions:
• How to build industrial-scale workflows,
• How to make content distinctive,
• How to develop IP,
• How to achieve localization and global distribution,
• How to build sustainable business models,
• And how to preserve genuinely valuable creativity as efficiency keeps increasing.
This is also one of the most important signals in AI entertainment in 2026:
The technology race is not over.
But what truly separates competitors
is increasingly not technology itself,
but the content systems built around it,
organizational capabilities and end-to-end commercial processes.
The Hangzhou summit examined this trend in depth.
It did not stop at the surface observation that “AI is hot.” It opened up a more practical industrial question: as AI begins taking over entertainment's supply side, those who can turn tools into workflows, workflows into content, content into IP and IP into long-term value are the ones with a real chance of reaching the next round.
What the summit truly clarified
was the next stage of the AI entertainment industry:
Competition between systems, not just models;
end-to-end loops, not isolated breakthroughs.
And that, in 2026, is
the industry signal that truly deserves attention.
This article was produced by Unique Research.