Walulu: Tear open a new track in six months

Text by: Guofu

Recently, Chengdu Yuzhilingdong Technology Co., Ltd. (hereinafter referred to as Yuzhilingdong) announced the completion of angel - round and Pre - A - round financing totaling tens of millions of RMB.

Currently, its investor lineup is quite impressive, including KK Group, 4399, iFlytek, and a well - known financial investment institution, covering four major dimensions: financial capital, channel resources, the gaming industry, and AI technology. Achieving this level basically covers the core elements required for an AI consumer goods company.

The fact that Yuzhilingdong can attract these four types of investors simultaneously indicates to a certain extent that their project not only appeals to the intuition of financial investors but also meets the market expectations of product providers for its products.

The core product of Yuzhilingdong is the AI bionic plush pet, Walulu. In the vision of the company's founder and CEO, Gu Xiaomao, they hope to build Walulu into the "Pokémon of the AI era."

This was not Yuzhilingdong's earliest idea. In 2023, when they decided to go all - in on AI, their first choice was to develop a self - owned vertical large - scale model. The team immediately purchased GPUs and formed an algorithm team, and continuously delivered multiple government and B - end projects worth millions.

During the same period, they also began to explore the 2C path. Relying on AI video generation capabilities, they produced several popular content pieces, with a single video achieving a maximum of tens of millions of exposures. Subsequently, they also plan to build a one - stop IP generation platform and have signed contracts with multiple studios and individual creators.

Display of AI video creation cases on the company's official website

However, when the production line was up and running, CEO Gu Xiaomao realized a rather fatal problem: the pure online model was difficult to form a stable business closed - loop. Although the content data might look good, the actual monetization path remained unclear.

It was also at this time that the team gradually discovered that the existing AI capabilities needed a physical carrier, a product that users were willing to pay for. Even self - created popular IPs needed a specific form to be realized. Otherwise, the IP itself could only stay at the traffic level and could not be transformed into real brand assets.

So, they began to look for answers from their past experiences. Before this, the core team members of Yuzhilingdong had been deeply involved in the gaming field for more than a decade. They had launched a single product with a gross profit of over 100 million and were responsible for products with more than one billion non - duplicate registered users. With the experience of a complete product cycle from scratch to a large - scale success, they had developed a set of self - summarized methodologies regarding user payment willingness, product iteration rhythm, and operation logic. Even when going all - in on AI, they should not abandon the advantages that the team had honed over nearly a decade.

Specifically in decision - making, they abandoned the idea of "technology looking for scenarios" and instead emphasized "scenarios looking for technology" - for example, introducing AI technology into the trendy toy market to create AI toys.

On the one hand, the trendy toy market now has a clear demand foundation. Companies such as Pokémon, Sanrio, and Pop Mart have verified the potential of aesthetic consumption and collection consumption with real achievements. On the other hand, AI technology can just solve the shortcoming of traditional plush toys, which lack interactivity and are "only for viewing, not for playing." The combination of the two naturally has great differential competitive advantages.

Introduction to the Walulu product

After finalizing this direction, they immediately abandoned the work on the underlying large - scale model and instead chose to access the mature AI capabilities in the market. They devoted all their energy to product definition and IP operation. Without the burden of underlying R & D, the team was able to design products and select technical solutions entirely around user needs, and thus completed the entire process from project initiation to product launch in just six months.

Choosing this direction also largely helped Yuzhilingdong avoid a large amount of potential future competition. Most AI startup companies follow the path of "emphasizing technology and neglecting products," that is, building barriers first and then looking for scenarios. After Yuzhilingdong decided to focus on products, their relationship with these companies naturally changed from competitors to potential cooperation partners.

The company's ability to quickly finalize multiple rounds of financing follows the same logic. For financial investment institutions, the team demonstrated verifiable business capabilities; for KK Group and 4399, the team chose a path that did not compete with them but was complementary; for iFlytek, it was the synergy in the implementation scenarios of voice technology. The combination of the four institutions essentially represents collective recognition of this path. Making a general product with a score of 70 is more sustainable than making a vertical product with a score of 90. This sentence is often quoted by the team and is the underlying logic for their choice of this path.

Walulu is the result of Yuzhilingdong's thinking based on this set of ideas.

The thinking of most AI toys is AI - first. Manufacturers often try to stack functions such as voice and emotion recognition first and then figure out how to turn it into a specific product. However, Yuzhilingdong's thinking is completely the opposite. They hope that users will first recognize Walulu, find it cute, and want to take it home. Then, they tell users that this little cat can actually talk, act coquettishly, and come over to rub against you when you are unhappy. In short, Walulu must first be a trendy toy that people want to own, and then an AI toy.

Introduction to the Walulu product

In this way, the design path of Walulu naturally becomes clear.

First is the image design. The team uses the AI capabilities accumulated in the past to continuously verify different IP characters on short - video platforms. Through feedback such as views, likes, comments, and more direct demands like "want to buy the same style," they judge which images can really appeal to users. Only those images with excellent data verified by the market will be advanced to the next stage.

Walulu is such a cat that stands out from many candidate images. Before its physical product was even formed, a makeup - related video had already achieved 30 million exposures across the network; another video of it peeping out from an eggshell had also received 600,000 likes on Douyin. In the comment section, many users had already started asking "how to buy" and "where to buy." At that time, the product hadn't even entered the mold - making stage.

High - view - count video on Walulu's official Douyin account

After that, the team found the chief designer in the trendy toy industry. Combining the designer's design concept and AI - assisted creation, they polished more than 100 versions of the image, adjusting details that are not easily noticeable in daily life, such as the size of the eyes, the angle of the ears, the texture of the fur, and even the overall roundness.

Secondly, there are the additional details of the product. For example, the team specifically invited a well - known voice - over teacher to record Walulu's meows in different scenarios and emotions - even though they all ultimately sound like just a "meow." Correspondingly, they also designed a series of rich actions for Walulu, constructing a complete set of behavioral logic for a living being through sounds and actions. When users pat it, Walulu will express grievance; when users gently stroke it, Walulu will start to act coquettishly towards its owner.

Finally, there is the growth memory system. CEO Gu Xiaomao describes this function as "the inheritance of digital life": Walulu will record the interaction data with its owner, continuously learn and update its memory. As the usage time increases, its understanding of the owner's habits and preferences also deepens. With OTA online upgrades, the product experience is theoretically expected to continuously evolve, and user stickiness depends on the actual performance of this system.

In short, all of this is to make Walulu "come alive" - and in fact, they have indeed achieved this. During offline sales, in addition to praising Walulu for being cute, many users subconsciously thought it was a real cat.

The subsequent sales data also prove that Yuzhilingdong's direction is completely correct. During the National Day in 2025, Yuzhilingdong opened its first brand flagship store at Joy Mall in Wangfujing, Beijing. The initial stock was only more than 300 pieces, and the team was not sure about the sales volume at that time. As a result, all the Walulu products were sold out within three days. In the following days, they could only meet the demand through pre - orders, and even the display prototypes in the store were bought by customers. During Christmas in 2025, the X11 co - branded store at Chengdu Global Center opened; during the May Day holiday this year, the first brand flagship store in Southwest China was opened at a well - known scenic spot in Chongqing. Currently, Walulu has been introduced into many trendy toy chain stores, including leading channels, and its offline brand influence is gradually expanding.

Walulu's flagship store in Wangfujing, Beijing

During the peak sales season from January to February online, the daily sales volume has also reached hundreds of pieces. This figure basically verifies the users' willingness to pay within a specific price range. So, Yuzhilingdong no longer holds back and has pushed Walulu into the full - scale sales stage, and simultaneously started the layout of the overseas market.

Sales popularity of Walulu on online channels

According to CEO Gu Xiaomao, they divide the overseas expansion into three stages: In the first stage, they mainly rely on self - operated e - commerce. The products have completed compliance certifications in major markets such as the United States, Europe, and Japan, and will enter the market through crowdfunding, Amazon, and independent websites; in the second stage, they will cooperate with overseas channels. Currently, they have reached a cooperation with an agent in Japan, with a guaranteed annual sales volume of 100,000 units; in the third stage, they will customize IP images that meet the local market needs for overseas customers, such as koalas in Australia, but still maintain the design language and style of Walulu to ensure the brand tone.

Of course, the current Walulu is not the end. Yuzhilingdong is still continuously iterating the product. For example, the pendant version of Walulu is in the development stage, targeting a lighter - weight and lower - priced market; the 2.0 version will be equipped with electronic - screen eyes, and the interaction form will be expanded from pure voice to visual feedback; IP derivatives are also on the agenda.

In addition, the company has launched co - branded cooperation with well - known IPs. Recently, it will launch the Garfield version and the giant panda version of Walulu, with the goal of cooperating with world - class IPs such as Pokémon and Disney to launch AI co - branded products. The direction of co - branded cooperation is not limited to the IP itself but also includes cooperation with technology companies, consumer brands, local governments, and cultural and tourism bureaus. This means that Walulu's business model is expanding from selling products to selling IPs. Using co - branded products to build brand awareness and obtaining large - scale revenue through licensing fees, it has gradually built its own brand influence.

Garfield co - branding

In the future, they will face more and greater challenges - when this category is proven to be profitable, the influx of followers will quickly lower the average profit.

Currently, Walulu has two major moats: one is the first - mover advantage, with a market - verified product and a certain level of user recognition; the other is IP operation ability, which is a soft power that requires time to accumulate and cannot be directly replicated by technology. However, these two moats are not deep enough. Once large companies decide to enter, the supply chain can be replicated in a short time. Whether Walulu can maintain its current premium ability depends on whether the brand can be truly established.

This is a competition against time. In six months, the team answered a question about ability; next, they need to answer a question about barriers at a faster pace. Whether they can solidify brand recognition before the influx of competitors is the key to determining how far this company can go.

A notable detail is that the main investors in the Pre - A round are also the investors in the angel round, with only a small share going to new investors. This means that the angel - round investors have exercised their pre - emptive rights to the greatest extent, which usually only happens when they highly recognize the company's current development. For Yuzhilingdong, this round of financing is not just about getting money but also about getting an entry ticket into the industrial circle. The endorsement of the four institutions is like a business card, which can help the team open some doors that were originally difficult to open, whether in channel cooperation or talent recruitment.

This article was first published on 36Kr Gaming.