Supergate has been selected for a government-funded project to develop a TinyML microcontroller equipped with a deep learning unit of 200 GOPS level. As the integration of new technologies like IoT, big data, AI, and 5G accelerates globally, the proportion of AI semiconductors in the system semiconductor market is also continuously increasing. By 2030, AI semiconductors are expected to account for over 30% of the total system semiconductor market. According to the Korea Smart Home Industry Association, the domestic smart home market is projected to be around 21.2 billion USD in 2025.
In the smart home and appliance sector,
the need for premium and new technologies is becoming more pronounced due to stagnation in the traditional appliance market and increased competitiveness in the Chinese appliance sector. Examples of new technology applications include AI upscaling, AI image enhancement, food ingredient recognition and status determination, and custom washing with object and laundry recognition in robotic vacuums.
The demand for edge-type AI SoCs is increasing, enabling AI inference without network connection, thereby enhancing applicability in various fields. Particularly, their use in areas with inadequate communication infrastructure and important sectors like wildfire monitoring is anticipated.
Supergate’s developed product has several competitive advantages. Firstly, as an edge-level AI semiconductor, it is suitable for the smart home and IoT market with its small size, low power, and cost competitiveness. Secondly, it boasts appropriate core performance for running applications. Thirdly, it supports up to 512MB without data bottlenecks or model size limitations thanks to the integrated DRAM controller and secures a bandwidth of 12.8Gbps.
This selection for the government-funded project is expected to be a significant step towards the growth of domestic semiconductor technology and strengthening competitiveness in the AI semiconductor market.
