Recently, Google DeepMind's release of Project Genie, an experimental prototype allowing creators to generate, edit, and explore 3D virtual worlds, has caused significant volatility in capital markets, leading to sharp declines in the stock prices of several US-listed gaming companies. However, a Bernstein research report points out that this panic is not only excessive but also reveals a superficial market understanding of the essence of game development.
According to the report released on February 2nd, the core point for investors is that generative AI is currently a probabilistic tool, while games are deterministic systems; their fundamental natures are different. While AI can accelerate asset production, it cannot replace the moats built through game rule design, numerical balance, and accumulated intellectual property. The report clearly states that compared to the increasingly strong resistance from Western developers, Asian gaming companies are in a more advantageous position in leveraging AI to enhance R&D efficiency. This irrational sell-off, instead, highlights the investment value of leading Asian companies that possess top-tier IP, robust long-term operational capabilities, and are actively embracing AI technology.
The technical reality is that the "hallucinations" generated by AI struggle to replace the deterministic rules of a game. Market fear surrounding Project Genie stems from an impressive video demonstration, but the devil is in the details. The report notes that this technology remains in an extremely early stage: generated virtual world sessions are limited to 60 seconds and cannot be saved for reuse. A deeper logical contradiction lies in the difference in technical principles. Generative AI models are inherently "probabilistic," guessing the next pixel or scene based on statistical patterns; whereas video games must be "deterministic," requiring rigorous rules to provide a consistent experience. When playing an action game, you need precise hit feedback, not randomly generated "surprises" from AI. Although the technology behind 3D world models will iterate rapidly, it currently addresses only one aspect of the many disciplines in game development—rendering—and remains far from building a fully functional game system.
The barrier of "fun" involves more than just the accumulation of 3D modeling. The moat in game development is far deeper than generating a beautiful 3D scene. The report emphasizes that "fun" is an extremely complex concept. Relying solely on AI to generate visual assets cannot solve core issues such as whether the narrative is engaging, the combat feedback is precise, or the character progression curve is reasonable. In multiplayer games, this complexity increases exponentially. Multiplayer games are arenas where players express deep desires like survival and conquest, requiring highly complex systems to support them, including in-game balance (a dynamic game between developers and players seeking exploits), anti-cheat systems, and matchmaking mechanisms. These are all "implicit variables" that current generative AI cannot comprehend or replace. Therefore, the view that AI can quickly supplant traditional game engines overlooks the complexity of games as a comprehensive interactive medium.
Content supply and the IP moat: Quantity does not equal quality. AI does lower the barrier to game production, leading to a surge in market supply. Data shows that over 20,000 new games were released on Steam in 2025, partly attributable to AI-assisted creation. However, this "explosion in quantity" has not altered the "monopoly on quality." Report data indicates that among the massive number of games released in 2025, only about 1,300 received more than 500 reviews; raising the threshold to 10,000 reviews causes this number to plummet to just 137. This demonstrates that while AI has increased the supply of relatively cheap, disposable experiences, it has not increased the supply of unique, differentiated content. The human brain experiences diminishing dopamine returns from repetitive experiences, meaning that merely using AI to generate vast amounts of homogeneous content cannot achieve commercial success. Conversely, the strength of IP, brand recognition, and player nostalgia remain critical factors determining commercial success. Whether it's Nintendo's classic IP or the *GTA* series, the emotional connections built over the long term cannot be replicated by AI overnight.
The challenge of long-term operations: A dynamic art difficult for AI to master. For mainstream live-service games, AI's ability to replace is even more questionable. The report argues that there is no "one-size-fits-all" solution for long-term operations; even AI trained on massive amounts of human behavioral data would struggle to outperform top-tier human teams. Historical cases show that even games with superficially similar appearances, like *Battlefield 6* and *Arc Raiders*, exhibited vastly different player retention rates post-launch. Sustaining a game's vitality requires constantly adjusting content pacing and balance based on player feedback—a dynamic, context-specific decision-making capability for which there is currently no evidence that AI is competent. If even industry veterans cannot guarantee operational success every time, expecting AI to automatically manage a complex online community is clearly unrealistic.
Future outlook: AI as the co-pilot, not the pilot, of the engine. Regarding the relationship between AI and game engines, the report proposes a rational path for integration: generative AI will be subordinate to the game engine. As Epic Games' Tim Sweeney stated, engines have advantages in stable world representation and physics simulation, while AI has advantages in handling diverse content. The future development model will likely be: the game engine is responsible for building the "skeleton," such as physical rules, lighting, and asset layout, while generative AI is used to fill in the "flesh and blood," such as crowd behavior, random events, or assisting in generating UGC content. Tencent's *Peacekeeper Elite* is a prime example; although its UGC content currently accounts for only a single-digit percentage of revenue, its Daily Active Users (DAU) increased by 25% year-over-year. This indicates that AI acts more as an enabling tool to help expand the content ecosystem, rather than completely颠覆ing the existing development process.
Geographical differences: The structural advantages of Asian manufacturers. Finally, the report presents a crucial investment perspective: the global gaming industry's attitude towards AI shows a clear divergence. The 2026 GDC State of the Industry report indicates that Western developers are becoming increasingly hostile towards AI, even displaying a "Luddite" tendency, which could lead to self-destruction during technological iteration. In contrast, gaming companies in Asia, particularly China, have demonstrated greater adaptability. Tencent is considered one of the industry's most AI-ready companies and is one of the few, besides Google and OpenAI, to have released a 3D world model (HY-World v1.5). Asian developers not only benefit from lower R&D costs but are also more proactive in embracing new technologies and leveraging AI to enhance productivity. In this wave of AI, Asian companies that can pragmatically integrate AI into their workflows and translate it into commercial value will have a greater chance of success compared to their Western counterparts constrained by ideological debates.
Comments