Artificial intelligence is rewriting the underlying logic of the IP industry from the ground up: IP is no longer merely a static asset such as film, animation or comics, but has evolved into an "all-scenario emotional infrastructure" that spans multiple terminals. Analysys (易观分析) reported in its October 2026 research report, The Value Reconfiguration and Ecosystem Map of the IP Industry in the AI Era, that the full penetration of AIGC into content production has pushed the supply side into an era of "infinite abundance" — in 2026 the market size of AI short dramas and motion comics is projected to reach RMB 40 billion, up 138% year on year, while the cost of a single AI motion comic has been compressed to roughly 1% of that of a traditional live-action drama. At the same time, demand for content that is "credible, emotional and interactive" is rising rapidly; the hit rate of AI short dramas is only 0.47%, and oversupply combined with emotional scarcity has created a structural mismatch.
IP scarcity has therefore shifted from "production capacity" on the supply side to "governance rights": whoever can guarantee personality consistency (that is, no OOC) and verifiable copyright ownership holds the core strategic resource. The report proposes an IP value formula — IP Value = Active Creators (AC) × Average Daily Generated Assets (AG) × Monetization Efficiency (ME) × Adjustment Variables — and divides the path into two parallel tracks: "AI transformation of classic IP" and "AI-native IP". Over the next three to five years the industry will confront six major challenges while evolving toward a circular asset network of "rights anchoring at the base, compute underpinning, tools empowering, platforms amplifying, terminals penetrating, and revenue flowing back". In an era of infinite supply, a stable persona, a credible narrative and clear rights attribution are the scarcest assets an IP can have.
Traditional IP has generally been understood as a "finished content product" that can be licensed repeatedly, flowing in a linear fashion through channels such as cinemas, satellite TV, games and derivatives, with its value ending at each transaction. Since generative AI became widely available, this paradigm is breaking down: content production costs have collapsed, and the relationship between IP and users has evolved from "one-way viewing" to "real-time interaction" and "long-term companionship". Analysys judges that the IP industry has already moved beyond the early stage of competing on generation quality and has entered a race to secure positions in infrastructure.
I. Macro trends: AI + IP becomes a high-certainty track for the content industry
The combination of AI and IP has become the most certain direction of evolution for the content industry. Its core contradiction lies in a "structural mismatch between the two ends of supply and demand": content expands without limit on the supply side, while emotional connection is scarce on the demand side. Together these two forces push IP into the position of a "certainty barrier for emotional connection".
1.1 Supply side: the marginal cost of content generation approaches zero
In 2026, AIGC technology has fully permeated the content production chain, pushing the supply side into an era of "infinite abundance": the market size of AI short dramas and motion comics is projected at RMB 40 billion, up 138% year on year; on Douyin alone, 221,900 new AI short dramas were added in the first half of the year, with daily output exceeding 1,200 titles and monthly view growth for AI motion comics surpassing 70 billion plays. The decline on the cost side is even steeper: the production cost of a single AI motion comic has been compressed to RMB 3,000–5,000, only about 1% of a traditional live-action drama, and the cost of a one-minute motion comic has fallen to RMB 200–300. Content generation has shifted from a "scarce good" to a "daily commodity", with marginal cost accelerating toward zero.
1.2 Demand side: rising demand for content that is credible, emotional and interactive
Infinite supply inevitably leads to content redundancy and homogenization. When "AI-flavoured" content floods the market, purely randomly generated content with no emotional anchor will quickly lose user patience, and the core demand shifts to content that is "credible, emotional and companionable". This demand has already been validated by the market: in the AI short drama The Girl Who Was Laid Off, the AI protagonist "Fang Taozi" gained 455,000 followers in short order thanks to her vivid personality and emotional resonance, and a single commercial endorsement was priced as high as RMB 258,000 — proof that users are willing to pay for AI characters capable of generating emotional bonds, and that emotional value converts directly into commercial revenue.
1.3 Supply-demand mismatch and the migration of scarcity
The popularization of AIGC has greatly lowered the barriers to production, leaving content supply in a state of "thermodynamic entropy increase", while human attention and emotional bandwidth remain fixed and finite. The two form a structural mismatch: the overall hit rate of AI short dramas is only 0.47%, and 99% of all content is drowned out the moment it is born. Content glut makes "relationships" the scarce commodity, and IP is the most effective carrier of relationships — holding consensus through "limited anchor points" and countering infinite noise with certainty. Its irreplaceability stems from a "finiteness" that is hard for AI to replicate: the finiteness of rules (a self-consistent worldview constituting an established "grammar"), the finiteness of personality (stable personality boundaries that prevent OOC), and the finiteness of memory (shared memories accumulated across years that become social currency).
Correspondingly, the logic of IP scarcity has shifted fundamentally. In the industrial era, scarcity arose from barriers on the supply side (high production cost, channel monopoly). In the generative AI era, as content production approaches infinity, scarcity instead derives from "control" over IP assets, expressed along four dimensions: the core resource shifts from "content production rights" to "IP governance rights"; the form of copyright shifts from "static attribution" to "dynamic attribution"; channel logic shifts from "traffic distribution rights" to "ecosystem embedding rights"; and the profit model shifts from "one-off sales / licensing fees" to "continuous usage-based revenue sharing".
II. Value reconstruction: valuation models and the dual-track path
Against the backdrop of infinite supply and scarcity migrating to governance rights, the valuation, development path and business model of IP assets all need to be redefined.
2.1 The IP value formula: AC, AG, ME and adjustment variables
Traditional valuation systems are based on industrial-era scarcity, focusing on "the depth of the copyright library" and "play counts / sales" — essentially a financial inventory of static, existing assets. The new measurement system of the AI era introduces an "asset turnover logic", focusing on the dynamic activity and cross-scenario penetration of IP: IP evolves from a "finished content product" sold once into a "productive factor" that can be called on demand, reused at high frequency and combined in real time. The core metric accordingly shifts from "the scale of assets held" to "the frequency of calls per unit of time".
On this basis the report proposes an IP valuation formula: IP Value = Active Creators (AC) × Average Daily Generated Assets (AG) × Monetization Efficiency (ME) × Adjustment Variables. AC represents the co-creation capacity and community stickiness of an IP: the larger the AC, the more vibrant the derivative-creation ecosystem and the stronger the long-tail influence. AG directly reflects the diminishing marginal cost brought by AI and can grow exponentially. ME represents the ability of the rights holder to recoup revenue through automated value capture and revenue sharing across the ecosystem. The adjustment variables cover implicit factors such as personality consistency (OOC prevention), compliance risk and brand reputation, and are often the "invisible levers" that determine long-term value.
2.2 Two parallel tracks: AI transformation of classic IP and AI-native IP
The report summarizes IP value growth along two parallel paths. Path A is the AI transformation of classic IP, characterized by "a long slope with deep snow and high LTV certainty": a large base, strong stickiness, a deep foundation for derivative creation and fast cold start, with high monetization certainty and easy access to brand licensing and offline revenue sharing. Its core risks centre on OOC (persona collapse), emotional backlash from longtime fans, and copyright disputes. Path B is AI-native IP, characterized by "pulsating bursts with LTV yet to be proven": a small value base and a cold start that requires algorithmic boost, but AI tools grant it enormous freedom and easily attract pioneering creators to co-nurture it. AG is extremely high because the entire chain is AI-generated, which allows it to spread rapidly across multiple terminals. Monetization relies on immediate means such as advertising and tipping, so it tends to be "a flash in the pan", and it also faces compliance constraints including content safety review and the obligation to label AI-generated content. Overall, classic IP uses AI as an "activator" and a "life extender", forming the ecological "ballast"; for AI-native IP to achieve high LTV, it must elevate the user relationship from "watching" to "companionship", substituting "emotional subscription" for "payment for content".
2.3 The cost curve and the restructuring of the cost structure
Generative AI restructures the IP cost curve from a "high-threshold J-curve" into a "flat-bottomed pot" — heavy accumulation up front, near-zero marginal cost afterwards — providing the physical precondition for the unlimited amplification of AG. This triggers three structural changes: the IP supply curve shifts rightward and long-tail IP emerges at scale, with the market evolving from "head monopoly" toward "diverse coexistence"; lifecycle costs are amortized and the gross margin structure improves, but IP must shift from "selling goods once" to "continuous services / API calls" in order to break through the LTV ceiling; and the cost structure for going overseas shifts from "asset-heavy production" to "heavy prompt operations and cultural risk control".
The cost structure of IP is likewise shifting from "labour-intensive" to "compute- and compliance-driven". Traditional IP relies heavily on offline production teams, with labour costs often accounting for over 50% of the total budget. In the AI era, core labour shifts to prompt engineers and AI trainers, while compute (token consumption) becomes the central cost of development and operations. The production model becomes "AI batch generation plus precise human calibration", and compliance costs (AI labelling, copyright attribution, cultural risk control) rise markedly, becoming a new cost black hole.
2.4 Reconstruction of the IP–user relationship and the new pattern of value distribution
The relationship between IP and users is shifting from "one-way reach" to "symbiotic co-evolution", which shows up along four dimensions: role displacement (users move from passive audience to co-creation nodes), the emotional contract (from idol worship to long-term companionship, with the narrative agent NAE acting as a 7×24-hour digital companion that moves the value anchor from "exposure" to LTV), power migration (the right of interpretation is decentralised from an official monopoly to the "personality range"), and interaction frequency (from low-frequency pulses to high-frequency online presence). This inversion of the relationship also breaks the traditional binary split between "rights holder + platform", spawning a "four-party dynamic distribution": the unit of distribution sinks from "title level" down to "call level", "token level" and "derivative-creation level"; settlement evolves from periodic manual reconciliation to real-time smart-contract clearing; and the granularity of distribution is upgraded from coarse to fine-grained and traceable.
III. Competitive landscape and ecosystem map: dominant players and breakthrough challengers
In the AI era, the participants, competitive positions and value-distribution relationships of the IP industry have been comprehensively restructured. The report depicts a complete ecosystem map covering traditional rights holders, emerging AI-native IP studios, the AIGC production toolchain, content communities and streaming media, smart terminals, and offline scenarios, and summarizes five major changes in the industrial ecosystem.
3.1 Rebuilding the foundation: rights attribution and compute elevated to the infrastructure layer
Before AI entered the picture, rights attribution relied on static contracts handled by legal teams and computing power was confined to the rendering clusters of IT departments — both were back-office functions that responded only after the fact. In the new AI + IP ecosystem, both are elevated to core infrastructure, and the traditional linear industry chain is restructured into a "double-deck sandwich": rights attribution on the top layer sets the ceiling for trusted transactions, while compute on the bottom layer makes intelligent production possible. The attribution layer evolves from static "one-off registration" into a full-lifecycle governance middle platform: watermarking and evidence preservation enable traceability across the entire lifecycle, support the real-time splitting of "original / training / orchestration" tripartite ownership, and embed compliance checks into the generation side. Compute steps out of its supporting role in post-production rendering and is elevated to the physical carrier and pricing benchmark of AIGC content generation: from a post-production tool to the "first mile" of generation, metered precisely by tokens, frame rate and polygon count. Elevating both at once solves the trust problem (clear ownership of digital assets lets all parties dare to license, dare to use and dare to create derivatives) and the pricing problem (making it possible to share revenue in real time per call, per volume and per usage).
3.2 The rise of the technology / platform layer: the largest "new addition" in the ecosystem
In the traditional model, technology served only as an implicit tool or a channel pipeline, floating outside core value distribution. In the AI era, the technology / platform layer rises to become the largest new addition in the ecosystem and its central hub, transcending the "gold digger" positioning to become an asset operator that commands "the alchemy": it supports backward the industrialized production of film, games and animation and consolidates premium IP barriers; forward, it lowers creative barriers and activates long-tail UGC and derivative-creation capacity; downward, it connects in-car and XR terminal experiences; and upward, it restructures the logic of AI editing and distribution. Its value capture shifts from "selling shovels" to "subscription plus data": every AIGC generation and interaction platform charges token fees, and user interaction data feeds back into model iteration, forming a "data flywheel". Ownership of behavioural assets is becoming the new focal point in bargaining between platforms and rights holders.
3.3 The fission of IP supply and the restructuring of operations and distribution
Under AI empowerment the "pyramidal" structure of traditional IP supply (monopolistic head, hollow middle) has fissioned into "four echelons": the first and second echelons (traditional rights holders and traditional content producers) use AI to revitalize existing assets and improve industrial efficiency, consolidating a head "plateau"; the third and fourth echelons (emerging AI-native IP studios and UGC / PUGC creators) release a flood of innovative "peaks". Under the orchestration of the technology / platform layer these echelons form a "revolving door" mechanism — head IP opens licensing to the mid-to-long tail, while long-tail hits feed back into the head. Traditional rights holders are transforming into "IP asset management companies" to activate dormant assets; traditional producers into "AI process factories"; AI-native studios achieve high-quality mass production with extremely small teams and extremely low cost; and the vast ocean of UGC / PUGC is elevated from "consumer" to "co-creative supplier".
The logic of operations and distribution is likewise restructured: it traditionally relied on channel monopoly and one-off transactions. In the AI era, distribution rights migrate toward "algorithmic recommendation + creator ecosystem + closed-loop data", and operations split into "professional operations building height, community co-nurturing building heat". On the professional operations side, promotion and distribution transform toward full-lifecycle IP value-added service providers, leveraging generative AI to run "IP pre-visualization" and precise reach across audience tiers. On the community co-nurturing side, content communities rise to become IP "co-nurturing bases", with official channels guiding co-creation through AI toolkits and turning users from "bystanders" into "a community of shared interest" through voting, crowdfunding and tipping.
3.4 Expansion of terminal scenarios: an exponential surge in IP "consumption outlets"
Traditional IP consumption relied on cinemas, television and print media, constrained by the linear experience of "sitting down and watching". In the AI era, ubiquitous computing pushes terminals from "players" to "entry points and carriers" of IP, and consumption scenarios fission into a three-dimensional network of "all-scenario, multimodal, instantly interactive" experiences. "IP everywhere" becomes the new normal, and all-scenario LTV formally replaces single-title play count as the core metric. In specific scenarios: AI designer toys embody IP characters, shifting from "selling toys" to "selling companionship"; XR and wearables are populated through 3D models and AR filters; offline real-world settings extend into multidimensional immersive experiences such as murder mystery games and theme parks; and in-car space captures commuting "golden hours" with audio dramas and NAE companionship.
IV. Outlook: the new IP distribution ecosystem of the next three to five years
Looking ahead three to five years, the report notes that the industry remains constrained by six major challenges, while continuing to evolve along the direction of an "asset network".
4.1 Six core challenges
First, a technical bottleneck: multimodal generation struggles to maintain a high degree of consistency in character imagery and worldview, and high-precision generation comes with enormous computational cost — if the "flat bottom" is not flat enough, the ROI logic of the dual-track path breaks through. Second, a red ocean of content: convergence in underlying model architectures pushes generated content into the "average face" trap, while the sheer volume of PUGC dilutes attention instead, hindering the conversion of AI-native IP from "pulsating bursts" to "long-tail LTV". Third, a rights-attribution dilemma: the copyright ownership of AI training data is ambiguous and the boundaries of derivative-creation rights remain in a vacuum, making "fine-grained revenue sharing" hard to implement and directly suppressing monetization efficiency. Fourth, an emotional shallow: existing AI companionship largely stays at the level of stitching together scripted lines, while misuse of emotional data and the risk of youth dependency constitute ethical red lines that limit the premium of "emotional subscription". Fifth, a commercial fog: weak willingness to pay on the consumer side and continued entrapment in "free logic" create a risk of inverted input-output ratios. Sixth, heavy compliance pressure: content safety review drives costs up, and mandatory labelling of AI-generated content may break the immersive experience. These six challenges directly constrain any improvement in AC, AG and ME, and add uncertainty to the adjustment variables.
4.2 Development trend: from "selling content" to "operating an asset network"
The traditional IP industry follows a linear pipeline of "creation → distribution → consumption", with value ending at each transaction. The AI + IP ecosystem, by contrast, is restructured into a circular system of "rights anchoring at the base, compute underpinning, tools empowering, platforms amplifying, terminals penetrating and revenue flowing back", and the industry core shifts from "producing blockbusters" to operating a callable asset network — IP elements (characters / worldviews) are packaged as APIs and flow continuously across all scenarios, metered like water and electricity.
Looking ahead, the ultimate form of IP is that personality-based IP becomes "emotional infrastructure", accompanying users around the clock across phones, XR devices, in-car systems and embodied robots. Copyright attribution, AI content compliance and the consistency of character personas are the industry's biggest bottlenecks; infringement by AI-generated content has shifted from verbatim copying to disguised infringement, making rights enforcement harder. The logic of going overseas is turning toward prompt tuning and localized cultural risk control. Supplementary research from institutions such as Extreme Innovation likewise points to trends in which AI upgrades from a tool to an autonomous creative "agent", content production moves toward "AI-led, human-supervised" collaboration, the new concept of "AI copyright" is established, and AI becomes a lever for breakthroughs in the globalization of entertainment.
Conclusion: who will win the right to define the new ecosystem
AI is pushing the IP industry from a "content business" toward an "infrastructure business". When content supply trends toward infinity and generation cost approaches zero, what is truly scarce is no longer production capacity, but the "control" to deliver stable characterization, credible narrative and clear copyright governance. The decisive factor over the next three to five years will be whether three capabilities can be built at the same time: establishing, on the compliance dimension, a system of rights attribution, labelling and cultural risk control for AI-generated content; mastering, on the AI capability dimension, systematized technology ranging from personality-consistency maintenance to cross-terminal interaction engines; and developing, on the globalization dimension, cross-cultural prompt tuning and localized operational capability. Whoever closes the gaps in all three dimensions at once will hold the discourse power of the ecosystem throughout the new golden decade of the IP industry.