TMTPOST — Digital mapping platform Amap has released a universal world model workshop named ABot-World Studio on Tuesday to unify interactive video generation and three-dimensional spatial rendering within a single interface. Developed by the company's computer vision research laboratory, the open-source software suite enables users to convert text prompts or static images into real-time interactive digital simulations saveable as standard videos or 3D Gaussian Splatting files. The software platform architecture utilizes optimized computing kernels to achieve hour-long continuous first-person inference sequences, running at 16 frames per second with a localized 1.2-second processing latency when deployed on a single consumer-grade desktop graphics processor. By deploying these generative spatial assets, corporate software developers can instantaneously construct multi-angle cinematic storyboards and synthetically map localized physical boundaries to train specialized industrial autonomous hardware without building expansive physical mockups. Generative artificial intelligence infrastructure across the Chinese mainland is shifting capital decisively toward spatial intelligence as major internet ecosystems seek direct commercial monetization avenues beyond traditional chat interfaces. As the operational overhead of manually reconstructing real-world geographic metadata remains high, major domestic technology conglomerates are prioritizing generative world models to accelerate autonomous driving and embodied robotics training loops. Consequently, platform architectures are competing heavily on computing-efficiency boundaries, aiming to compress specialized simulation costs to secure a foundational role within the emerging industrial spatial-computing ecosystem.