Qualcomm Technologies Inc. on November 5th participated in the “AI Supergap Challenge Results Presentation” held at the Conrad Hotel in Yeouido, Seoul, where it unveiled collaboration outcomes with South Korean artificial intelligence (AI) startups.
Hosted by the Ministry of SMEs and Startups and the Korea Institute of Startup & Entrepreneurship Development, the event brought together Qualcomm, LG Electronics, Qualcomm’s local partners — DTECT, Integrit, Argosdyne and Bluebird — and the selected startups. Attendees included Vice Minister Noh Yong-seok of the Ministry of SMEs and Startups, Kim Sang-pyo, Qualcomm vice president and head of Korea operations, and Park Hyung-se, president of LG Electronics’ MS Business Division. Participants reviewed months-long joint projects and shared plans for commercialization.
Qualcomm led the Vertical AI track of the government’s AI Supergap Challenge, supporting collaboration in four key areas: autonomous driving robotics, inspection drones, AI edge boxes and industrial handheld devices. Throughout the program, Qualcomm provided the Qualcomm® AI Hub and a suite of software development kits (SDKs) to ensure stable integration and deployment of AI solutions. The company also coordinated efforts to help partners embed startup technologies into real products, creating close working relationships that allowed startups to leverage Qualcomm’s global network, strengthen technical competitiveness and accelerate diffusion of edge AI across industries.
At the venue, local partners Integrit, Argosdyne, DTECT and Bluebird operated demo booths; several presented detailed collaboration results and commercialization roadmaps.
Notable collaborations include:
- DTECT / MindForge: Joint development of an AI box/system based on the Qualcomm Dragonwing QCS6490 processor, with plans to expand into industrial, public safety and road/traffic safety applications and pursue joint exhibition and commercialization opportunities.
- Integrit / Designed AI: Testing AI models optimized via quantization techniques to reduce RAM usage while improving accuracy.
- Bluebird / KeySense: Deployment of Qualcomm-based industrial RFID readers and terminals with QNN SDK to run NPU-based vision AI models, achieving multi-barcode reading and accurate recognition of damaged barcodes to meet production-line efficiency requirements.
- Argosdyne / Deep-ET: Development of an initial drone-based vision AI model for concrete crack detection; Argosdyne will supply collected training datasets to improve model accuracy.
Kim Sang-pyo, Qualcomm vice president and head of Korea operations, said Qualcomm “holds core technology leadership that drives intelligence and connectivity of edge AI across smartphones, PCs, XR and automotive devices,” adding that the program demonstrated how linking local partners and startups can help standardize edge AI across IoT products. He pledged to expand such collaborations to support growth of the domestic AI ecosystem and wider adoption of edge AI.
Vice Minister Noh Yong-seok said startups “are developing breakthrough technologies from novel ideas and growing into key players in the AI ecosystem through collaboration with large firms,” and stated that the ministry will broaden the AI Supergap Challenge into more sectors and gradually increase participation by partner corporations to support startup growth and global expansion.
