K3AI Launches 3D Metaverse Data Library Platform for AI Training

K3AI has recently embarked on developing a new growth engine by launching a business based on its own platform modeled after Nvidia's "Omniverse." The company is now entering the artificial intelligence (AI) market with its high-quality 3D data assets. On May 29, K3AI announced the launch of its 3D asset library platform, "Neuro Asset."

"Neuro Asset" is a large-scale, high-quality 3D asset library platform based on K3AI's proprietary XR (extended reality) technology, designed for use across various industries. Leveraging more than 20 years of accumulated expertise in 3D simulation and modeling for sectors such as manufacturing and education, K3AI has refined and platformized 3D data that has been validated in real-world industrial settings.

Through "Neuro Asset," K3AI plans to utilize its 3D data not only for simple content production but also for implementing its own AI services and supplying 3D data via its platform. Recently, data collected from the metaverse has been used for training generative AI. Notably, the global big tech company Meta is currently training its proprietary generative AI model, "Llama," with data collected from the metaverse environment.

K3AI possesses a total of approximately 30,000 high-definition 3D data assets. These data assets are not limited to specific industries, offering broad applicability, and are refined to a level suitable for immediate use in AI training. Unlike text and images, high-quality, refined 3D data is in limited supply. However, the company explained that it has maximized scalability based on its in-house production and commercialization references.

A K3AI representative stated, "For the launch of this platform, we have completed the integration and refinement of high-quality 3D data collected while creating metaverse components, making it suitable for AI training. As global big tech companies are increasingly utilizing metaverse data for AI training, we will leverage our own platform to convert high-quality 3D data built in metaverse environments into external training datasets, driving our AI data-centric business forward."

He added, "The core technology lies in separating, processing, and transforming metaverse content into a structure that AI can directly learn from. We will demonstrate this through XR middleware and the library platform. By supplying 3D data to AI companies and continuously advancing our technology, we also plan to accelerate our business transformation through the development of proprietary AI technologies."