Translated from Founder Park

Creating an IP universe was once the privilege of companies like Disney and Nintendo. Nieta is turning it into an ability anyone can have.

Recently, the AI-native community platform Neta announced that it has raised over $10 million in a Pre-A+ funding round, led by Jiukun Ventures, with participation from BV Baidu Ventures. Existing investors Source Code Capital and MiraclePlus (formerly YC China) also increased their stakes in this round.

The new funding will primarily be used for three directions:

• Recruiting top global talent to support international expansion
• Increasing investment in technology research and development
• Expanding new product lines

01

Nieta: From “Creating Characters” to “Building Worlds”

Launched in March 2024, Neta originally focused on AI character creation. Users can create virtual characters using natural language and build images, comics, short videos, and interactive gameplay experiences around them.

The founder, Hu Xiuhan, graduated from Peking University and Columbia University, and previously worked at Meta, where he was responsible for the technical architecture of short-video publishing platforms.

The team includes members from leading internet and gaming companies such as ByteDance, Bilibili, and TapTap, bringing strong expertise in multimodal AI technologies, product growth, and global community operations.

Currently, the platform has:

• 12 million+ users
• Average daily interaction time exceeding 110 minutes
• 5.84 million virtual characters created
• 400+ virtual worlds and community spaces, each with their own scenes and rules

On the commercial side, Neta’s revenue in China has already covered user acquisition and compute costs, achieving positive unit economics.

As AI technology evolves and user creativity deepens, themes on the platform have expanded beyond entertainment to include traditional culture such as Dunhuang and intangible cultural heritage.

Thousands of users now collaborate under shared worldviews and unified artistic styles. Neta is evolving from “character creation” to “world construction.”

02

Defining the “World Expression Protocol Layer”

The creation of a “world” can be divided into three stages.

Stage 1: The PGC Era

Represented by companies such as Disney and Marvel.

Building a universe like the Marvel Cinematic Universe, including characters, scenes, rules, stories, and social interactions, typically requires:

• Tens to hundreds of millions of dollars
• Thousands of creators
• Over a decade of production

Stage 2: The UGC Era

Represented by platforms such as Minecraft and Roblox.

The right to create worlds was decentralized to users. However, the tools were still relatively primitive, lacking:

• deep characters
• narrative structures
• high-quality visual production

This era proved that “everyone wants to build worlds.” But the supply-side creative tools were not powerful enough.

Stage 3: The AI Era

Platforms like Neta represent the next stage.

AI dramatically lowers the barrier to world creation and reconstructs the cost structure. A single creator can now express a complete world through:

• Presentation layer — characters, objects, visuals
• Rule layer — interactions, gameplay mechanics

Compared with the Disney/Marvel era, the cost of building a world is now roughly one ten-thousandth of the previous cost.

As models continue to improve and creators refine their skills, AI-driven world creation will redefine content consumption experiences, unlocking entirely new business models and technological advantages.

Neta’s capabilities have expanded from text-image generation to interactive gameplay and video creation.

The platform supports:

• natural language generation of virtual characters
• multi-character interactive comics
• AI-generated video storytelling

Its open-source Neta model series topped the HuggingFace text-to-image leaderboard in 2025, gaining attention from the global developer community.

Unlike many AI applications that rely on a single model capability, Neta has built a “World Expression Protocol Layer” above the models, defining standardized assets and interaction rules for:

• characters
• objects
• scenes
• behaviors inside virtual worlds

This means that even if the underlying models evolve — from image generation to video models, and eventually to world models — the accumulated world assets and creator ecosystems on the platform will continue to grow rather than become obsolete.

By continuously integrating the latest model capabilities, Neta is building the first application scenario for sustainable real-time interactive world models, enabling multi-user collaborative world creation and participation.

03

Targeting the World’s Largest Creative Community

Neta is launching an upgraded global creator toolset, targeting audiences from global fandom and creative communities such as Comic-Con and AO3.

The company is currently recruiting product designers and cultural strategy experts with global perspectives.

The demand for fictional world creation is universal:

• Japanese anime and otaku culture
• Western fantasy traditions
• global fan-fiction communities

All contain large, passionate communities of young creators.

Founder Hu Xiuhan stated:

“In the past, building a Disney-level virtual world required decades and billions of dollars.

What we are validating is this: when AI lowers the barrier to creation far enough, anyone with imagination can build their own theme park.

AI characters will live, perform, and construct inside these worlds.

What we are building together with creators is the standard for world expression.

This form of expression naturally crosses cultural boundaries, and we are bringing it to the global stage while searching for people who truly understand world creation and want to define the rules with us.”

An investor from BV Baidu Ventures commented:

“Neta starts from the cultural and creative needs of young users. Hu Xiuhan has a clear vision for the evolution of AI technology and content communities.

Through its self-developed multimodal models, the platform continues to reduce the barriers to personal creation, allowing AI-native Gen-Z users to participate in content production and expression.

BV has long valued opportunities in AI-native content platforms, and we look forward to seeing the team further explore creator ecosystems and community forms.”

An investor from Source Code Capital added:

“Neta’s strength lies not only in its accurate understanding of model and technology trends, but also in its deep understanding of ‘world expression.’

The team has captured the needs of global young creators and communities.

As multimodal AI interaction becomes inevitable, their community structure is naturally suited to continuously generate AI-native creators.

Transforming creation from a privilege of a few elites into an ability accessible to everyone is something we are excited to see.”

首发 | 捏 Ta 完成超千万美金 PreA+ 轮融资，定义 AI 时代世界创作的基础设施

创造 IP 宇宙曾经是迪士尼和任天堂的特权，捏 Ta 正在把它变成一种个人能力。

近日，AI 原生社区捏 Ta 宣布完成超千万美金的 Pre A+ 轮融资，由九坤创投领投，BV 百度风投跟投，源码资本、奇绩创坛等老股东超额跟进。

这一轮融资将主要用于三个方向：招募面向全球拓展的顶尖人才，技术研发投入，以及新产品线拓展。

⬆️关注 Founder Park，最及时最干货的创业分享

超 22000 人的「AI 产品市集」社群！不错过每一款有价值的 AI 应用。

邀请从业者、开发人员和创业者，飞书扫码加群：

Image

进群后，你有机会得到：

• 最新、最值得关注的 AI 新品资讯
• 不定期赠送热门新品的邀请码、会员码
• 最精准的 AI 产品曝光渠道

01

捏 Ta ：从「创造角色」到「构建世界」

上线于 2024 年 3 月，捏 Ta 最初是一个专注于 AI 角色创作的平台。用户可以用自然语言创建虚拟角色，围绕角色创作图像、漫画、短片和可交互的玩法等。

创始人胡修涵，毕业于北京大学和哥伦比亚大学，曾在 Meta 负责短视频内容发布平台的技术架构。团队成员来自字节跳动、Bilibili、心动 TapTap 等互联网和游戏知名企业，在 AI 多模态技术、产品增长和全球社区运营等领域拥有深厚积累。

目前，平台已积累超 1200 万用户，活跃用户日均互动时长超过 110 分钟。在平台上，创作者们构造了 584 万个虚拟角色，上线了 400+ 包括场景、规则的虚拟世界和社团空间。商业化层面，捏 Ta 在国内的收入已覆盖获客与算力成本，实现单位经济模型打正。

随着 AI 技术迭代、用户积累和玩法的深入，主题也逐渐扩充到包括像敦煌、非遗等传统文化，上万人在同一个世界观中统一风格参与创作，捏 Ta 平台正在从「创作角色」升级为「构建世界」。

Image

捏 Ta 团队合影

02

定义「世界表达协议层」的底层能力

关于「世界」的创作，可以分为三个阶段。

第一阶段：PGC 时代

以迪士尼、漫威为代表，构建一个包括角色、场景、规则、故事、社交等多个维度的「漫威宇宙」，需要数千万至数亿成本，数千人团队，耗时十年以上。

第二阶段：UGC 时代

以 Minecraft、Roblox 为代表，造世界的权利下放给用户，但工具相对原始，缺乏角色深度、叙事规则与视觉品质。这证明了「人人都想造世界」这个需求是真实的，但供给侧的能力跟不上。

第三阶段：AI 时代

以捏 Ta 为代表，AI 将世界创作的门槛进一步降低，完整成本结构大幅重构。AI 让一个有创作欲望的普通人，可以通过表现层（角色、物件、视觉）和规则层（互动、玩法）进行完整的世界表达。

相比迪士尼、漫威时代，「造世界」的成本是原有的万分之一。随着模型能力成熟与用户创作技巧提升，AI 时代的世界创作将重新定义内容消费体验，构建出全新的商业模式和技术优势。

捏 Ta 的技术能力已从图文生成扩展至交互玩法与视频创作。平台支持自然语言生成虚拟角色、多角色互动漫画与视频内容创作等功能。其开源推出的 Neta 系列模型于 2025 年登顶 Huggingface 文生图模型榜首，获得全球开发者社区关注。

不同于多数依赖单一模型能力的 AI 应用，捏 Ta 在模型之上构建了一套「世界表达协议层」，定义虚拟世界中的角色、物件、场景的资产标准和互动规则。

这意味着，无论底层模型如何迭代——从图文生成到视频模型，乃至未来的世界模型——平台所积累的世界资产与创作者生态都能够持续增强，而不会因技术换代而失效。

平台持续引入最新的模型能力，正在构建可持续实时动态交互世界模型的第一个应用场景，打造让多人可以参与互动共创的核心玩法和虚拟世界基础设施。

03

锚定全球最大的泛创作社群

捏 Ta 正在面向全球推出升级版创作工具，目标人群覆盖 Comic-Con、AO3 等全球泛创作社区。公司目前重点招募具有全球视野的产品设计师和文化策略人才。虚构世界的需求是全球性的，从日本二次元文化到欧美奇幻文化，虚拟世界创作在世界各地有着大量年轻、原生、热情的创作者。

捏 Ta 创始人胡修涵表示：

「过去造一个迪士尼级别的虚拟世界需要几十年和几十亿美金。我们正在验证的是——当 AI 把创作门槛降到足够低，每个有想象力的人都能造出自己的主题乐园，而 AI 角色会主动在里面生活、表演和建设。我们和创作者一起做的，就是定义这件事的标准。这种表达天然跨越文化边界，我们正在将其带向全球，并寻找真正理解世界创作、愿意一起定义规则的人加入。」

BV 百度风投投资人表示：

「捏 Ta 从年轻用户文化和创作需求为切入，修涵对 AI 技术演进与内容社区建设有比较清晰的判断，通过自研的多模态模型持续降低个人创作门槛，让 AI Native 的 Z 时代用户能够参与到内容生产和表达中。BV 一直重视 AI 原生内容平台的机会，期待团队持续探索创作者生态和社区形态，推动新一代 AI 内容社区的发展。」

源码资本投资人表示：

「捏 Ta 的厉害之处，不只是对模型和技术方向看得准、看得远，更在于他们真的懂『世界表达』这件事，也精准抓住了全球年轻用户和社区的需求。在 AI 多模态交互已成大势的现在，他们做的社区，本身就适合源源不断长出 AI 原生的创作者。把创作从少数精英的特权，变成每个人都能用的能力，我们很期待给更多用户带来好的体验。」