NetDragon (NDWTY) AI Game Output Rises at Least 10%

NetDragon Websoft Holdings Limited (NDWTY) said on its August 28, 2026 earnings call that AI had entered core game development and increased content output for core products by at least 10% versus the second half of 2025.[1] The company did not disclose a corresponding revenue, expense, or profit contribution.

NetDragon develops and operates online games, and this AI application sits directly inside game-content production. Game producers, planners, artists, and developers give instructions in natural language. AI employees receive and analyze the assignments, then generate art assets, code, and playable content. The system supports the core research-and-development workflow from production brief to delivered game content; it is not a separate player-facing feature.

From playable demos to core game development

The AI game-content application expanded from a limited production example disclosed on March 27, 2026 to daily work on core products disclosed on August 28, 2026. At the starting point, more than 10 AI agents helped produce a playable demo in two to three weeks.[2] By the later disclosure, management was no longer presenting only a single demo: it said AI had gradually penetrated the development process for important core game IP. The users, production objects, and outputs remained consistent across the two disclosures. The change was in scope and measurement level, with no evidence that the application was renamed or merged into another workflow.

Adoption scale and operating result

The latest figures describe both adoption and output. NetDragon said AI handled 35% to 40% of overall workload in the first half of 2026, while content output for core products increased by at least 10% versus the second half of 2025.[1] The workload share shows that AI participates in a substantial portion of development tasks, while the output figure places the reported result on content throughput. However, the company did not define its workload calculation, the set of core products, or the unit of content. The two figures therefore cannot be used to calculate unit productivity, and the content increase cannot be converted directly into game revenue.

Financial link remains unquantified

Higher content throughput could improve R&D spending efficiency and the game business's profit margin, but the evidence is directional rather than financially reconciled. If the same team can consistently deliver more usable content in a comparable period, development cost per unit of content could decline; faster updates and greater content supply could also support player activity and revenue. Management also said development headcount may remain stable and that capacity released by efficiency gains would be reinvested in AI games, new businesses, and system expansion. The gains may therefore appear as reinvestment rather than lower total R&D expense or personnel cost. NetDragon did not disclose application-specific development cost, R&D expense, margin change, or revenue contribution.

NetDragon has moved AI content production from a playable-demo example into core game development and has reported higher content output. It has not established that the increase produced cost savings, margin improvement, or revenue growth. A comparable bridge from AI-covered products to unit content-development cost or R&D expense, content releases, and commercial revenue would be needed to test the financial effect.

Application assessment

• AI Game Content Production | Business position: core operations | Deployment stage: limited production | Scope: one business unit | Value type: cost reduction

Sources

[1] Drillr · NetDragon Websoft Holdings Limited (NDWTY) · 2026-08-28 · Earnings call

Original: At least in terms of core model products, we have increased content output by at least 10%.

[2] Drillr · NetDragon Websoft Holdings Limited (NDWTY) · 2026-03-27 · Earnings call

Original: In fact, in the second half of last year and the second half of this year, we have achieved one to two main creators. With the help of more than ten AI agents, we can produce a playable demo in two to three weeks. This is unimaginable. This is at least 10 times more efficient.