DeepMaze AI and MakinaRocks advance Physical AI and industrial automation

DeepMaze AI has been selected for the 2026 General TIPS program by South Korea's Ministry of SMEs and Startups. The company aims to develop ultra-high-speed and high-precision sensor fusion solutions to enhance the spatio-temporal perception of Physical AI, such as robots and autonomous driving systems. This technology seeks to reduce the trade-off between processing speed and accuracy by combining data from heterogeneous sensors, allowing machines to operate reliably in unstructured environments with noise, dust, or varying light.

In a separate industrial application, MakinaRocks is set to implement AI-based control optimization solutions at Samsung Display's mass production lines in Vietnam by the end of this year. Following successful proof-of-concept testing, the company will use Large Language Models (LLMs) to analyze and clean up accumulated PLC (Programmable Logic Controller) code. This process aims to reclaim memory capacity and prevent the need for costly factory shutdowns by automatically identifying redundant or unused commands within the manufacturing equipment.