3 E Network Technology Group Limited (NASDAQ: MASK) released strategic insights from CEO Dr. Tingjun Yang regarding the company's robotics technology roadmap. The Hong Kong-based B2B IT solutions provider outlined plans to develop edge computing infrastructure for robotics applications.

Dr. Yang stated that AI models are moving to edge computing, requiring robots to develop "spatial intelligence" capabilities beyond basic semantic understanding. The CEO noted that robotics technology is shifting from rule-based programming toward systems that integrate reinforcement learning and simulation-to-reality transfer methods.

The company identified computational bottlenecks in current hardware when processing multi-modal sensor data from high-framerate vision, spatial radar, and tactile arrays. Traditional merchant silicon faces limitations related to the "Von Neumann bottleneck" and "Memory Wall" issues, contributing to higher power consumption and latency problems.

3 E Network plans to develop heterogeneous computing architectures and Edge AI system-on-chips optimized for multi-modal workloads. The company's approach involves optimizing data paths at the silicon level and integrating tensor acceleration engines with low-latency on-chip SRAM to reduce energy consumption from data movement.

The strategy includes hardware-software co-design, with underlying real-time control algorithms and customized compilers designed for integration with proprietary chips. This approach aims to translate semantic instructions from edge models into motion control commands for joint motors with reduced latency.

3 E Network recently formed a strategic partnership with California-based robotics company Aladdin Alaris AI in the smart healthcare sector. The company targets the eldercare robot market as an entry point, with plans to expand to industrial collaborative robots, logistics, and general-purpose humanoid robots using a standardized computing platform.

The information is based on a company press release statement.