Carving out a position in the emerging new market of AI casual games, we are currently seeking angel round financing.
When the torrent of AI technology surges into the gaming industry, how can small teams seize the industrial inflection point to explore a differentiated development path?
Casual games are becoming the segment with the most certain growth in the global mobile game market. According to data from Sensor Tower, the global revenue of the casual game segment reached about 22 billion US dollars in 2025, among which the revenue growth rate of the hyper-casual track exceeded 100%; against the backdrop of the general decline in the overall download volume of mobile games, the casual game segment still maintains growth relying on its extremely low access threshold. This trend is mainly carried by the mini-game ecosystem in the Chinese market. According to the public disclosure of the WeChat Mini Games platform, its monthly active users have exceeded 500 million, the number of developers has exceeded 500,000, among which the monthly active users of IAA products dominated by advertising monetization have exceeded 400 million; the "2025 China Game Industry Report" released by the Game Publishing Committee of China Audio-Video and Digital Publishing Association shows that the advertising monetization revenue of domestic mini-program games in 2025 reached 17.071 billion yuan. The installation-free, click-and-play form eliminates the download and registration processes, enabling casual games to reach a large number of people who were not previously regarded as game users.
The scale potential is supported by the inherent characteristics of the casual game segment: targeting general users without pre-learning cost; each round takes a short time, which fits fragmented scenarios; the gameplay is light with high revisit frequency, and user stickiness is built on daily habits; the creative space is open, a single point of innovation can form differentiation and spread rapidly relying on social relationship chains. This determines that the ceiling of casual games is not delimited by the capacity of a certain sub-category of gameplay, but continues to rise with the expansion of the general user scale.
After the threshold is lowered: supply expansion and homogenization dilemma
The situation on the R&D side is more complex. AI has greatly lowered the creation threshold — links that originally required division of labor such as art, code, and planning can now be completed by a single person with the help of tools to produce basic outputs, and the number of entrants has increased rapidly accordingly. However, the improvement of efficiency has not been equally transformed into cost advantages. Most traditional game companies introduce AI by embedding it into existing processes to replace partial links. The tools save working hours at a single point, but the organizational structure and collaboration links have not changed; the saved working hours cannot be converted into manpower reduction, while tool procurement, computing power consumption and team adaptation constitute new additional expenditures, so the total R&D cost often rises instead of falling.
The more fundamental restriction lies in the underlying logic. The traditional R&D system is based on manpower division of labor and process connection, and the process design and decision-making links are built on the basis of positions; the AI-native workflow requires organizing production by capabilities, allowing single-point decision-making to run through the entire link. The migration between the two is not just tool replacement, but the reconstruction of organizational form, which is extremely difficult to promote in teams that have formed a stable structure. The real AI Native workflow is more likely to be implemented first in small teams without historical burdens.
The congestion on the supply side is finally transmitted to the user side. Under the mature user acquisition system, what products players come into contact with does not come from active selection, but is determined by the delivery rules — the material click rate, the first-day payback efficiency and the ROI model jointly screen the pushed content. This set of rules rewards certainty rather than novelty. Once a certain gameplay is verified to be viable, a large number of products with similar structures will appear in a short period of time. As a result, a contradictory situation is formed: the number of products increases year by year, but the perceptible choices of players are narrowing instead.
At the same time, the classic gameplay of the casual game segment has been fully explored, and the cost and failure rate of building differentiation by relying on original gameplay have risen rapidly. Most teams therefore prefer to make minor adjustments on the mature framework, which further strengthens the homogenization cycle. The reaction of players is clear: aesthetic fatigue is intensifying, patience for new products is declining, and advertising acceptance and willingness to pay are decreasing simultaneously. This is not an operational problem of a certain type of product, but an inevitable result of the entire supply system under the existing rules.
Build AI-native workflow from scratch
In summary, the casual game market is large enough and still expanding, while threefold dilemmas exist on the supply side at the same time: traditional game companies find it difficult to complete the migration to AI-native, homogenized competition leads to an increase in product quantity but a narrowing of choices, and players' demand for new content has not yet been matched by a mature form. Beijing Boudong Yixia Network Technology Co., Ltd. has taken these three dilemmas as its strategic starting point since its inception. Traditional companies need to complete transformation on the basis of existing processes, while this project is built from scratch, directly organizing production based on AI capabilities without historical burdens; the OPC model brings lower fixed costs, shorter verification cycles and more flexible decision-making links, enabling it to complete far more trial and error than conventional teams with the same amount of resources.
In terms of product direction, the project uses light casual games to carry AI-generated content — the development cost is controllable, and the experience difference brought by generative content is more significant in this form, with low investment and clear feedback signals. This direction has been implemented since the first product, and is promoted as a whole according to the structure of "three ends and three stages": vertically, the product end, R&D end and enterprise end are simultaneously AI-native, and horizontally, the AI-native content goes through three stages from light embedding, to continuous deepening as the pipeline matures, and finally to the formation of a complete new product form. The core purpose of this arrangement is to seize the capability and position before the rules are reshaped.
Pipelines, products and verification, current achievements of the enterprise
The project was launched in December 2024, and the company was officially established in September 2025. Over the past year, the project completed the development of 7 complete projects within 11 months, the development cycle of a single product was compressed to 2 to 6 weeks, and a set of mature local R&D pipelines was precipitated. The formation of the pipeline follows a clear principle: do not take the demonstration effect of the model as the basis for model selection, but directly put new capabilities into real commercial projects for verification, and only solidify the parts that have been polished by multiple projects and are stable and reusable. The resulting delivery system does not rely on any single model, and includes complete local assets such as code reuse specifications, AI collaboration processes and platform adaptation standards; the localized AI Agent workstation built by the founder has accumulated more than 50,000 lines of prompts, and trained exclusive digital avatars that undertake repetitive work in planning, art, code and operation links. The feasibility of this path is also related to the founder's industry experience — Mi Yunmao has 14 years of experience in visual and commercial design in the game industry, and has participated in the R&D of projects such as Final Fantasy: Awakening and The Walking Dead: Survivor.
Five of the products have been launched on the WeChat Mini Games ecosystem, with a total of more than 20,000 accumulated users. Good Luck Fortune Sticks has accumulated more than 9,300 users relying on natural growth, with an active retention rate of about 30%, and the retention rate remains at 9% 90 days after launch; Pull a Big Pile has a next-day retention rate of about 37% and a three-day retention rate of about 28%. This set of data confirms two things: the pipeline can stably produce complete commercial products that can be launched and operated, rather than staying in the prototype stage; the market demand targeted by the products truly exists.
A more directional verification appears in the AI fortune-telling function of Good Luck Fortune Sticks. Relying on generative content, this function provides personalized interpretation feedback for players. Although the form is light, it completes a key pre-verification — AI-generated content can be naturally accepted by users in the light casual usage scenarios, and becomes one of the reasons for players to continue using the product. Based on this, the project has established a unified principle for subsequent products: every new product must be embedded with generative content capabilities, so that the content generation capability can continuously obtain real user data along with the product, including usage rate, actual impact on the retention curve and the structure of generation cost, instead of putting it into use all at once after the conditions are fully mature.
The advancement on the enterprise end is carried out simultaneously. The project has settled in the Zhongguancun North Latitude AI Community OPC Incubation Center, obtained 3 software copyrights, advanced to the final of the Haidian District OPC Special Competition of the Hicool Global Entrepreneurship Contest, and obtained the OPC certification of Haidian District, Beijing. So far, the first-stage survival verification has come to an end, and the closed loop has been completed at the product, R&D and enterprise three ends.
In terms of business model, the project currently focuses on IAA advertising monetization of light casual games, obtains revenue relying on mini-game platforms such as WeChat and Douyin, and adopts self-operation or co-distribution for products with excellent performance to amplify momentum; in the medium term, the verified R&D pipelines will be gradually modularized, and authorization and module services will be explored for small and medium-sized teams to form a second revenue curve.
The essence of the light casual track is creativity-driven. There is high uncertainty about whether a single product can succeed, and the number of trial and error directly determines the hit probability. The OPC model reduces the cost of a single trial and error to a fraction of that of the traditional link. The core purpose of this round of financing is not to make a single product heavier, but to obtain far more verification opportunities than conventional teams with the same amount of capital, converting uncertainty into affordable costs. The project has not yet obtained external financing, and the founder has promoted the implementation with own funds. This round plans to raise 2 million to 3 million yuan, and the funds are mainly used to support the large-scale implementation of generative content in products, including computing power consumption, local equipment supplement and development and operation investment. It is planned to recruit no more than 5 employees in the follow-up, who will be responsible for pipeline optimization, product output, business and financial work respectively.
Mi Yunmao said that AI technology is at the inflection point from R&D integration to content explosion. Whoever completes the combination of generative content and real commercial products first in this window period will occupy a favorable position when the rules are reshaped. The most urgent problem of the project at present is the shortage of funds and manpower. The core goal is to realize positive cash flow as soon as possible and complete the first-stage survival verification. In the future, the project will continue to adhere to the positioning of "small but strong, continuous evolution", not blindly expand the number of personnel, continue to accumulate R&D capabilities, re-organize the production mode of light casual games with AI, and bring more innovative AI game experiences to global general users.