Valuation of 3 Billion USD in Three Years! How is Yanyu Technology Leveraging a Head Start to Pursue an IPO on the Hong Kong Stock Exchange?

Amid the global surge in large AI model enthusiasm, can the "intermediary" between models and end users sustain a massive capital story without directly developing underlying foundational large models? Recently, according to media reports, Evoken has officially started preparing for its Hong Kong IPO, with a new round of financing valuation reaching $3 billion. This company, which rapidly rose within three years through multiple AI creation and application products, once again sparked in-depth discussions in the industry about the commercial value and competitive barriers of AI application layers.

Starting from a Model Community: The Evolution of Liblib and the Business Logic of the "Intermediary"

Evoken's development journey began with the release of the image model community LiblibAI in 2023. At that time, the open-source image model ecosystem was booming, and the platform gradually evolved into a vertical community similar to a code hosting platform by aggregating models uploaded and fine-tuned by creators. With the rapid advancement of multimodal large model capabilities, Liblib completed a major upgrade to version 2.0, transforming from a single model fine-tuning community into a multimodal creation station. It integrated several mainstream multimodal large models at the core and introduced a more user-friendly interface for one-click generation.

Although the product positioning continued to expand, its underlying business operation logic remained centered on service packaging by integrating multimodal model capabilities. As an "intermediary" of large models, its profit model mainly relies on obtaining API cost advantages from official or channel sources and selling them through bundled packages. For example, the platform uses mechanisms like member points expiration to offset actual usage costs. This model is not traditional tool sales but rather a refined operation based on platform scale and user actual consumption.

Diverse Product Matrix: Rapid Breakthroughs from Design Agents to AI Video Creation

To overcome the limitations of merely acting as a middleman, Evoken quickly laid out multiple vertical application products, building its own growth engine.

Firstly, it launched a specialized agent assistant product targeting the design field. After the explosion of the general agent concept, the team quickly formed a relevant product team, focusing on full-process automation and efficiency enhancement in the design field. Its core idea is to use agent logic to break down design tasks, optimize specific functions to address designers' pain points in text and image processing, text modification, and effect alignment, thus achieving deep optimization of traditional design processes.

Secondly, it introduced an AI video platform for the video creation field. With the boom in the AI short drama and video generation sectors, the related products rapidly entered the market by modularizing script generation, storyboard design, visual generation, and post-production editing, contributing significantly to the company's ARR (Annual Recurring Revenue). By keenly capturing industry trends and rapidly iterating product features, the company has gained a first-mover advantage in multiple niche markets.

Moat and Time Gap: Multiple Real Challenges in the AI Application Layer

Although Evoken has shown impressive revenue growth and product iteration speed, its business model still faces many practical challenges and objective scrutiny.

On one hand, the core moat of AI application layer products largely depends on the "time gap." When large model capabilities continue to decline and internet giants enter vertical markets, the differentiated barriers brought by first-mover advantages may be quickly compressed. Major manufacturers, leveraging their vast ecosystems and resources, are accelerating the layout of similar creative intelligent agents and creation platforms.

On the other hand, the migration cost for users on AI tools is relatively low, and market competition often evolves into price wars and speed races of function iteration. How to maintain rapid growth while improving long-term user retention and stickiness, and build more solid technical and ecological barriers, will be a key issue that such AI application companies need to continuously solve in their long-term development. As the IPO process advances, the market will also give the final answer to the profitability quality and sustainability of this "intermediary" model.