Recently, tech behemoths across the globe have been unanimously ramping up their exploration of AI application scenarios. OpenAI is reportedly conducting internal trials of an advertising model for ChatGPT, which will prioritize displaying 'sponsored content' when users seek product recommendations, aiming to offset the high operational costs. Meanwhile, whispers suggest Google is secretly developing advertising prototypes, exploring viable avenues to merge AI responses with commercial promotions. Meta, not to be outdone, has acquired 'Butterfly Effect', the developer behind the AI application Manus, for billions of dollars to bolster its AI capabilities. Microsoft and Amazon, too, are delving into the integration of AI agent functionalities into industry-specific solutions.

This flurry of activity collectively signals a clear shift: the AI race has transitioned from the initial phase of 'model competition' to the subsequent stage of 'application competition', centering on identifying practical use cases for the technology.

As Yao Shunyu, a former OpenAI researcher who recently joined Tencent as Chief AI Scientist, aptly put it: 'The AI competition isn't about training; it's about defining and evaluating genuinely useful tasks. We must consider for whom we're solving problems and how effectively we're doing so. This will determine who can unlock trillion-dollar economic value.'

I observe that gaming currently stands as one of the few mature domains capable of harnessing AI's complex capabilities, enabling scalable validation and continuous iteration. However, its potential has consistently been underestimated.

01

AI: A Pathway for Global Game Developers to Enhance Player Experience

Survey data reveals that on the Steam platform, the number of games actively disclosing the use of generative AI has surpassed 10,000, constituting nearly 10% of the total. The estimated total revenue from these games stands at $660 million, with 12 of the highest-grossing titles including last year's domestic sensation 'Emotional Anti-Fraud Simulator' (more commonly known as 'The Gold Digger Game'). In the gaming sector, the application of generative AI faces controversies and challenges. Yet, regardless of player sentiment, AI has undeniably become a force to reckon with. Our task is to leverage it to craft superior gaming experiences.

Recently, Otto Soderlund, AI lead at 'Clash of Clans' developer Supercell, publicly declared AI as 'triggering the greatest transformation since the Industrial Revolution,' with the gaming industry poised to be among the first severely impacted, given the rapid evolution of intelligent agent AI.

He underscored China's astonishingly rapid development pace in this regard.

According to the 'Game Enterprise AI Technology Application Survey Report' released by the Game Industry Committee, the AI penetration rate in intelligent NPC development has reached 68.4%.

For an extended period, NPCs have functioned more as rigid nodes, existing to trigger tasks, provide information, or populate the game world without truly participating in its dynamics.

Today, AI is infusing NPCs with 'souls.' By endowing them with world understanding, emotions, and behavioral logic, game characters are no longer mere puppets reciting lines but digital entities capable of real-time dialogue and forming bonds with players, making each player's actions trigger unique game outcomes.

Whether it's the AI town in university experimental projects where characters spontaneously form social relationships or the role systems in commercial games that can alter attitudes, strategies, and even fates based on player behavior, they all point to the same trend: the gaming world is evolving from a 'tourable space' to a 'symbiotic system.'

For instance, Tencent Games' 'Peacekeeper Elite' introduced Bruce, the industry's first AI combat dog with AI-powered voice recognition. No longer just an electronic pet providing weather and horoscope updates, Bruce is a formidable ally capable of turning the tide in a match:

Shout 'Go lick the box,' and he'll sprint to an open area to grab airdrops, minimizing player exposure risk;

Shout 'Go bite the nearest enemy,' and he'll act like radar to locate enemies lurking silently in corners;

Shout 'Go save the teammate,' and he'll run behind another cover to rescue.

Some netizens even shared that after being knocked down by an enemy, Bruce aggressively dragged and bit the enemy for a second or two, creating precious rescue time.

This character sparked discussions not for its flashy AI abilities but for being embedded in a real tactical decision-making chain. Players aren't just operating a tool; they're collaborating with a 'partner' capable of understanding instructions, assessing risks, and executing teamwork.

The same logic applies to AI teammates, AI commanders, AI companions, and other forms. Ubisoft emphasizes 'collaboration over replacement' in its generative AI teammate project. Tencent, NetEase, and other developers are exploring AI's potential as tactical units, narrative nodes, or long-term companions across various genres. These attempts collectively indicate that AI is becoming an integral part of the gaming experience.

If the changes on the experience front are more player-centric, then on the production side, AI's impact directly influences industrial efficiency and organizational structure.

Game development has long grappled with challenges: content production is becoming increasingly complex, while production efficiency lags. Open worlds, higher-precision assets, and more frequent updates are continuously lengthening development cycles.

AI's integration into game development pipelines offers more and better production tools, transforming game production models. For example, Tencent Games introduced VISVISE, the industry's first AI-powered full-process 3D animation pipeline. By integrating four modules—skeleton generation, intelligent skinning, 3D animation generation, and intelligent frame interpolation—it creates a one-stop workflow, allowing AI to handle highly repetitive yet essential production tasks. This enables creators to focus more on high-value decisions like gameplay design, narrative construction, and artistic style.

Tencent Games' unique approach of 'technological pragmatism' stems from the challenges it faces in serving a massive DAU user base and complex device environments. This drives them to apply and deeply optimize cutting-edge technologies for 'universal' use, ensuring that high-end 3A experiences aren't exclusive to premium hardware but can be stably delivered on the widest range of terminal devices. Making the most advanced achievements accessible to the broadest player base is precisely the most profound implementation value of AI in the gaming sector.

02

Value Reassessment: How Can Game Companies Position Themselves in the AI Era?

As the relationship between gaming and AI deepens, collaborations between game companies and AI/tech firms have never been more robust. In the future, game companies may become even more intertwined with AI firms, as evidenced by partnerships like NetEase with Meshy and DeepSeek, MiHoYo's early investment in MiniMax, and KRAFTON's collaboration with NVIDIA on the ACE project. Tencent, as a global leading game developer, possesses unique advantages and value:

It may currently be one of the few companies capable of establishing a comprehensive AI+gaming ecosystem, building a layered, mutually reinforcing collaborative system.

In terms of foundational models, Tencent's self-developed 'Hunyuan' foundational large model forms the core of its AI capabilities. The platform also deeply integrates advanced 3D generation technologies, enabling efficient creation of high-precision 3D models while leveraging world models for simulation and deduction of complex virtual environments.

In terms of mid-level technologies and toolchains, Tencent decomposes cutting-edge AI technologies into 'usable, practical, and scalable' tools based on foundational model capabilities, embedding them into real-world development processes.

Leveraging its foundational large models and mid-level technologies and toolchains, Tencent holds a natural advantage in commercialized products. It operates multiple top-tier games with tens of millions or even hundreds of millions of DAU, covering core gameplay types like MOBA, shooting, and tactical combat. These large-user-base application scenarios enable cutting-edge AI technologies to directly enter scalable application stages.

Thus, Tencent has forged a clear path of 'technological pragmatism': starting from cutting-edge research but always using real, complex, and large-scale game products as technological testbeds and validators. Under the continuous pressure of massive user bases and highly complex products, AI technologies are continually pushed to the boundaries of 'usability, practicality, and scalability.' In this process, Tencent achieves a delicate balance between engineering capabilities, system integration, and frontier exploration.

As the AI competition enters the 'second half,' what's being contested is precisely this comprehensive strength to translate technological potential into real user value and industrial progress. Therefore, the reassessment of game AI's value is essentially a reassessment of companies like Tencent, which possess full-chain technological integration, implementation, and commercialization capabilities. Gaming is the stage, and Chinese developers like Tencent are proving to be not just the shining stars on stage but also the behind-the-scenes architects of this technological revolution.

References:

AI Base: OpenAI's Strategic Shift: Abandoning the AGI Dream to Sustain ChatGPT

Yahoo Finance: $90 Billion is Just the Beginning—Google CEO: Fully Charged Towards 'Full-Stack AI'

FORTUNE: An MIT Report Revealing 95% of AI Pilots Fail Alarms Investors. But It's the Reason for Failure That Should Worry Executives

Game Research Society: The Rise of University 'AI Town' Projects—Teaching AI to Work Overtime

Tencent Cloud: Tencent's Large Model 'Stirs Up Trouble'—Game Intelligent NPCs 'Come Alive!'

Game Grape: Meet Tencent Games' Most Obedient 'Dog'

Wall Street Journal: Tesla's 'World Simulator' Debuts: Learning 500 Years of Human Driving Experience in One Day, Optimus Can Share the Same 'Brain'

Interface News: 27-Year-Old Former OpenAI Researcher Yao Shunyu Joins Tencent

BCG: How Platforms Are Colliding and Why This Will Spark the Next Era of Growth

IT Home: Around 8% of Games on Steam Use Generative AI, Generating $660 Million in Revenue

Zipdo: AI in the Game Industry Statistics

Dr. Fei-Fei Li: From Words to Worlds: Spatial Intelligence is AI's Next Frontier

Silicon Star: The 'Hidden Battlefield' in the AI Parameter War: Why Global Top Players Are Turning to 'Gaming'

Sina Esports: Tencent Games VISVISE Unveils the Industry's First AI-Powered Full-Process 3D Animation Pipeline at SIGGRAPH Asia

Cool Play Recommendations: Ubisoft's First Playable Generative AI Game 'Teammates'—NPCs Understand Human Speech, Collaborate on Puzzles, and Even Talk Back! Google Gemini3 is Super Powerful!

GameLook: Tencent, NetEase, and MiHoYo 'Speed Up' in AI Game Development Competition!

Supercell Unveils 'AI Technology Strategy'! Lead: 'To Empower Employees with Superpowers,' Chinese Peers' Progress is Astonishing