[Edaily Reporter Park Jung-Soo ] M eighty-three Co.,Ltd., a company specializing in visual effects (VFX), is set to fully enter the synthetic data business for physical AI training in collaboration with its AI R&D subsidiary, D-Blat.
M eighty-three Co.,Ltd. announced on the 24th that it will collaborate with Diblat to pursue projects involving the creation of synthetic data and virtual environments for AI training, which will be used in humanoid robots, autonomous driving, and industrial robots.
As global big tech companies have recently entered a race to develop humanoid robots and autonomous driving technology, securing AI training data has emerged as a key competitive advantage. Physical AI requires learning from a wide range of situations that can occur in real-world environments, but relying solely on real-world data is considered to have limitations in terms of time and cost.
Synthetic data is created by replicating real-world environments in a digital space and repeatedly generating various variables and scenarios. It is gaining attention as a key infrastructure for next-generation AI development, as it not only reduces data acquisition costs but also enables the simulation of rare or dangerous scenarios that are difficult to capture in real-world settings.
Diblatt is advancing its technology for recreating real-world spaces and objects in digital environments, leveraging M eighty-three Co.,Ltd.’s VFX production capabilities and content production expertise. The two companies plan to build training data for physical AI by utilizing the high-precision 3D asset production technology and live-action-based environment recreation technology they have accumulated through the production of movies, TV dramas, and Over-The-Top (OTT) content.
In particular, Dblat is utilizing physics-based simulation technology to create virtual environments that reflect real-world physical laws, such as gravity, collisions, friction, and deformation. A key feature is its ability to provide training spaces that closely resemble real-world environments, going beyond simple image generation.
Through this, the company aims to expand its business scope into various industrial sectors, including robotics, drones, defense, and smart factories.
A Diblatt representative stated, “In the era of Physical AI, the ability to accurately recreate real-world environments in digital spaces will be a key competitive advantage,” adding, “We possess the technical capabilities to generate diverse scenarios and variables based on real spaces and objects.”
A representative of M eighty-three Co.,Ltd. stated, “We will combine the VFX technology and AI R&D capabilities we have accumulated through the content production process to target the Physical AI market,” adding, “We plan to expand our business scope beyond AI-based content production into the industrial AI sector.”