China’s Big Bet on AI for Games

What the report says

I recently tried an AI-driven detective puzzle game that an independent Chinese developer released on itch.io. The concept is simple enough: playing the role of a detective, you interrogate suspects powered by a large language model that allows them to improvise in character with spoken audio to keep you off balance, requiring you to sift through their answers to solve the crime. The visual presentation still feels a bit rough, and the core mechanic comes across as incremental rather than groundbreaking. Yet, it signals a broader shift: in China, integrating AI directly into core gameplay has evolved beyond isolated experiments; it is now a systemic trend.Ever since generative AI emerged, Western game developers have primarily focused on copyright concerns, ethical questions, and fears over job displacement.

In contrast, studios in China pivoted to a more pragmatic question: how can AI drive practical results? Today, it’s clear that AI is already helping Chinese developers reduce operational costs and streamline production. Yet rather than stopping at efficiency, they continue to push boundaries, investing heavily in unproven, forward-looking applications.A Tale of Two SpeedsWhenever AI is brought up, a sense of lingering anxiety seems to shadow the Western gaming scene (more so than other industries). The 2026 GDC State of the Game Industry survey found 52% of developers think generative AI is hurting the industry — up from 30% a year earlier and 18% the year before that.

As that sentiment has taken hold, AI’s rollout in the West has faced noticeably higher resistance. Most notably, Steam makes developers disclose AI use, and more than 7,300 games had a label as of March. Larian promised to drop generative AI from the concept-art phase of its next game after fan backlash. Call of Duty: Black Ops 7 caught complaints over its AI-made calling cards too. AI still slips into backend work — EA’s Battlefield team runs LLMs to catch duplicate Jira tickets, for instance — but the moment the output lands in front of players, the backlash gets noticeably louder.

Key details

Many companies are bending the knee as a result.In China – not unlike mature Western markets – rising development and user acquisition expenses are squeezing margins, and hit rates are falling as companies increasingly compete for existing audiences. However, in response to these pressures, China’s game industry is moving at a completely different pace.Looking at distribution, Tanwan Games reports an AI adoption rate surpassing 80% across art creation and ad targeting. Similarly, 37 Interactive’s 2024 annual report indicates that AI produces over 80% of its 2D art and assists in more than 70% of its ad video creation.In core gameplay, AI integration is already directly embedded into major titles.

NetEase’s Justice Online markets its dynamic, AI-driven NPCs as creating “a Jianghu that breathes,” while Giant Network’s Supernatural Squad recorded over 25M matches against AI opponents during its launch week.When it comes to art pipelines, Tencent’s VISVISE platform stands out. Character rigging traditionally took one to three and a half days, while generating a ten-second skeletal animation required three to seven days. VISVISE accelerates character rigging by more than 8x and has been deployed across nearly 100 projects, including flagship titles like Honor of Kings, League of Legends: Wild Rift, PUBG Mobile, and Teamfight Tactics.

Beyond a mere feature set, its true strength lies in execution: developed hand-in-hand with the PUBG Mobile production team, the core GoSkinning module (v4.2) was tailored to overcome practical hurdles such as skirt physics and four-legged rigging. Such close integration with an active pipeline provides a competitive advantage that off-the-shelf software would struggle to match.Is Moving Faster with AI Paying Off?Gamma Data estimates AI could add ¥53.3–84.6B (roughly $7.9–12.6B) in new market value for China’s game industry. Net profit across the industry is projected to grow from roughly ¥111.37B ($16.5B) in 2026 to ¥162.66B ($24.1B) by 2030, with AI as one of the major drivers.Among China’s top 50 game companies by revenue, 80% have disclosed AI plans publicly, and 56% have built dedicated labs or teams.

More from the announcement

A 2025 Niko Partners survey puts AI usage at around 60% of Chinese studios. In the first half of 2026, AI-driven titles were just 15% of China’s top 100-grossing mobile games, but they generated 42.6% of the revenue. This also reflects strong products and mature teams that are better positioned to leverage AI, rather than just newer entrants.However, we’re also starting to see examples of new games that benefit from more efficient AI workflows. For example, 37 Interactive’s AI-native mini-game, Heart of the Moon Palace, would normally have required a development cycle of six months or more.