Editor’s Note: There is no shortage of technology news, yet many promising Chinese startups still remain largely unknown to global audiences. At the same time, the constant flow of funding news, product launches, and new technologies makes it harder to see which companies truly deserve attention. DragonLab is TechNode’s effort to close that gap, looking past the noise to discover China’s next wave of innovation—emerging companies, ambitious founders, and breakthrough technologies that have yet to receive global attention.
Making a video game has traditionally been a complex process involving multiple specialized roles. Designers create concepts and gameplay systems, programmers build mechanics, artists develop characters and environments, and producers coordinate development and iteration.
Over the past two decades, several companies have tried to make game creation more accessible. Professional engines such as Unity and Unreal Engine have given smaller teams access to powerful development tools, while no-code platforms have helped people without programming backgrounds experiment with making games.
Now, a new generation of startups is exploring a more ambitious idea: what if creating a game could begin with an idea described in words, rather than years of learning development tools?
That is the vision behind JIYI’s SOON, an AI-powered game creation platform designed to help users turn creative ideas into playable game experiences.
The company is exploring whether AI can take on more of the work traditionally handled by game designers, artists, and programmers. It hopes to help creators move from an idea to a playable prototype much faster.
The ambition behind SOON can be summed up in a simple phrase: a game studio in minutes.
From No-Code Engines to Ai-Powered Creation
To understand what SOON is trying to change, it helps to look at an earlier effort to make game development easier: GDevelop.
GDevelop began as an open-source project in 2008. The company behind it, GDevelop France, was incorporated in 2021 and is based in Paris. It has since grown into a no-code game engine aimed at people who want to make games without first learning a traditional programming language. Instead of requiring users to write most game logic in code, it provides a visual, event-based system for building interactions and gameplay. The platform lets creators publish games across web, desktop and mobile platforms.
That makes GDevelop a useful reference point for SOON. Long before generative AI became part of game development, GDevelop was already tackling the same broad problem: how to let more people make games.
Its answer was to simplify the development environment. Users still build scenes, define events, set behaviors and make design decisions themselves, but they can do so without mastering conventional programming. GDevelop has since added AI features as another layer of assistance, while remaining fundamentally a game engine that creators operate.
SOON starts from the same desire to lower the barrier, but pushes the idea further. Rather than simply making an engine easier to use, it aims to reduce the amount of manual work a creator performs inside that engine-like workflow. The platform presents natural-language interaction as a way to move from an initial idea toward a playable game, with AI taking on more of the execution.
For now, SOON is focused largely on 2D and mini-game projects, a deliberate trade-off that lets it use a more constrained framework to automate more of the production process while keeping output relatively stable. That makes it narrower than GDevelop, which supports both 2D and 3D development, but also reflects SOON’s strategy of going deeper on AI-driven game generation rather than trying to match the breadth of a general-purpose engine.
GDevelop already lowers the technical barrier to game development by removing much of the need for traditional coding. But creators still need to learn how its engine works, structure game logic, manage assets, and assemble the game themselves. SOON is trying to lower that barrier further, allowing users to describe what they want while AI takes on more of the production work.
Large commercial games still depend on experienced teams, precise control, and sophisticated development environments. But for prototypes, small projects, and creator-driven games, SOON is already showing what its approach can do. During its second round of testing, users created more than 3,000 game projects in 45 days, with the platform helping generate code, game logic, characters, scenes, and UI.
A Startup Emerging From the Gaming Ecosystem
Many AI gaming startups in the U.S. have emerged from software or AI research backgrounds, focusing on areas such as AI characters, interactive agents, and generative content. Inworld AI, for example, provides tools for adding intelligent characters, voice, and real-time interactions to games built with engines such as Unreal and Unity.
JIYI’s background is different. The company operates within China’s gaming and technology ecosystem, where developers and investors are exploring AI as a way to reduce the cost and complexity of game production.
In this context, JIYI has positioned SOON as an AI game-creation platform. Its approach reflects a question increasingly facing the gaming industry: Will AI remain another tool in the developer’s toolkit, or will it reshape the way games are produced? SOON is placing its bet on the latter.
JIYI was incubated by Chinese gaming company Kingnet Network and has since attracted investment from Bingo Group, Zhongxu Future and Hengdian Capital. In 2026, the company had raised nearly RMB 100 million ($14 million) across two funding rounds.
China is one of the world’s largest video-game markets, with a substantial domestic development and publishing industry. That gives startups like JIYI a large pool of developers, publishers, investors, and production experience to draw from as they experiment with new ways of making games.
SOON has also begun to attract industry recognition. At the 2025 China International Artificial Intelligence Conference, it was named one of the year’s 10 most innovative AI products, alongside products from companies including Huawei, JD Logistics, and Inspur Cloud. In July 2026, SOON was named Best AI Game Creation Platform at the seventh Feifan Awards, which focuses on the commercial application of AI.
JIYI’s pitch is ultimately about production efficiency. SOON aims to compress more of the game-making process into a single platform, helping smaller teams and individual creators move from an idea to a playable project with fewer steps.
Beyond game creation, in a strategic partnership with CMGE, the company launched GamePartner.AI, an AI-powered agent designed to help developers publish games to platforms including Apple’s App Store and Google Play, while connecting them with advertising monetization services such as AppLovin.
The Rise of the One-Person Game Studio
One of the most interesting ideas behind AI game creation is the possibility of smaller teams and individual creators producing more ambitious projects.
Traditional game development is often limited by resources. A creator may have a strong idea but lack the programming, art, or design skills needed to turn that idea into a finished product.
AI tools could change that balance. A creator who previously needed several specialists may be able to use AI to generate early versions, test concepts quickly, and refine ideas at a lower cost.
The trend resembles changes seen in other creative industries. Digital cameras, editing software and online publishing platforms lowered barriers to creating and distributing photography and video, giving independent creators direct access to tools and audiences that were once harder to reach.
Game development may be heading in a similar direction. If tools like SOON can take over more of the technical work, people with ideas but little development experience could have a much easier path to making something playable.
The Challenge: Creating Games People Want to Play
Generating a game is much harder than generating an individual asset. A successful game depends on how its mechanics, pacing, challenge and feedback come together to keep players engaged.
Some platforms like Meshy in the U.S. and Hyper3D’s Rodin in China, largely focus on helping developers create characters, props, environments, textures, and other production-ready assets. These tools can save artists and studios significant time while fitting into existing game-development workflows.
SOON, on the other side, takes a broader approach, bringing code, game logic, characters, scenes, and UI into the same creation process with the goal of producing a playable project.
Some companies are also expanding beyond asset generation into interactive environments and world-building. Tripo AI, for example, is developing persistent multiplayer worlds through Project Eden, while Runway has been exploring real-time, AI-generated game worlds that respond to player input.
a successful game requires much more than characters, environments, or code. It needs compelling gameplay, balance, progression, storytelling, and a reason for players to return. Research on game design and player experience has repeatedly emphasized the importance of factors such as challenge, feedback, immersion, and engagement.
For tools like SOON, one of the current questions is how to turn those generated pieces into an experience that feels coherent and keeps players interested.
A New Chapter for Game Creation
For years, platforms such as GDevelop have widened access to game development by making the tools easier to use. SOON is pushing that idea further, giving AI a larger role in the work that happens between an initial concept and a playable game.
The platform is still being developed, and large commercial titles will continue to rely on experienced teams and sophisticated production tools. But SOON points to a different kind of creator workflow, one where people can describe what they want, see a playable result quickly, and then shape it through further instructions.
That could make game creation accessible to people who have ideas but little experience with coding, engines, or production pipelines. For them, making a game may start less with learning a tool and more with simply trying an idea.