“Evolution of on-device AI in automobiles requires high-performance and high-capacity memory”
Published2024.08.14 17:41Kang Wook-sung, Vice President of SK Hynix, delivered a presentation on "The Direction of Intelligent Vehicle Memory-Storage Development in the AI Era" at the "7th Artificial Intelligence Semiconductor Forum Breakfast Seminar" held on the 14th at JW Marriott Seoul. He stated that automobiles are rapidly commercializing AI functions, including performing AI inference in ADAS, utilizing LLM-based on-device AI in the cockpit, and leading the earliest adoption of robotics in autonomous driving, and that the necessity for high-performance, high-capacity vehicle memory storage is consequently increasing sharply in the future.
Related news
“AI Semiconductor: Small-quantity, multi-variety, NPU technology-equipped foundries are key”
Lee Ju-seok, Vice President of Intel, asserted that the future AI semiconductor market will focus on edge devices where differentiated data is paramount, and that foundries with advanced technical capabilities to produce such semiconductors will become critical in the AI semiconductor era.
2024.06.26by 배종인 기자
"A new era where vehicles learn and judge for themselves"…Edge AI becomes the brain of automobiles
Edge AI is a technology that directly analyzes data within vehicles to immediately process critical functions such as braking, steering, and motor control with ultra-low latency and ultra-low power consumption. In electric vehicles, it increases efficiency and safety through battery charge and health state prediction, predictive maintenance, and virtual sensor implementation. In motor control, it complements existing algorithms to improve responsiveness and efficiency, and through anomaly detection, it identifies abnormal patterns in various components such as batteries, motors, and tires at an early stage. STMicroelectronics (ST) is actively working to expand the development ecosystem.
2026.09.04by 배종인 기자
"Taking AI Out of the Cloud"…Edge AI Opens a New Industrial Revolution
Edge AI is a technology that analyzes and makes decisions directly at the point where data is generated, reducing latency while enhancing security and energy efficiency. It is emerging as essential in fields where real-time performance is critical, such as industrial automation, automotive, and healthcare. STMicroelectronics (ST) identifies reduced latency, cost savings, energy efficiency, reliability, and privacy protection as core advantages. For successful implementation, hardware selection, model optimization tools, and collaboration between data scientists and embedded engineers are important, and future AI is expected to evolve into a hybrid structure combining cloud and edge.
2026.09.04by 배종인 기자
In the Edge AI Era, 'Self-Thinking Sensors' Transform Intelligent Factories
With the proliferation of edge AI, manufacturing is undergoing transformation into intelligent factories that analyze data on-site in real time. STMicroelectronics' latest MEMS sensors measure diverse information including vibration, temperature, and acoustic signals, and execute machine learning within the sensor itself through in-sensor AI to assess anomalies in real time. This enhances predictive maintenance efficiency and reduces network load. Developers can directly embed trained models in MLC and ISPU-based sensors, and development difficulty has been lowered through tools such as MEMS Studio. Sensors are growing in importance as the neural network of AI, and smart sensors are evaluated as a core enabling technology for intelligent factories.
2026.09.04by 배종인 기자
"Measurement, Design, and Safety Verification Also by AI"…Physical AI at the Center of Palletizing Revolution
Physical AI is emerging as a core technology in manufacturing automation, and collaborative robot palletizing processes are undergoing significant transformation. Universal Robots, Robotiq, and Safetics have presented next-generation automation solutions combining AI-based design, verification, and operation technologies. Robotiq introduced a platform where AI designs the workcell and analyzes ROI simply by inputting box information, while Safetics showcased technology that evaluates collision risks through simulation-based safety verification and proposes optimal speeds. In actual field operations, improvements in productivity and shifts in worker roles are becoming evident, and future automation competitiveness is expected to be determined by Physical AI capabilities rather than hardware alone.
2026.09.04by 배종인 기자
