[e4ds Tech Day 2026 ⑦] PopcornSAR CEO Chae Seung-yeop, "SDV Orders Difficult Without AI-Native Systems"
Published2026.10.02 15:52PopcornSAR CEO Chae Seung-yeop presented on 'AI-Native SDV Development Architecture: Practical Methods for Converting Legacy C to MBD·AUTOSAR Development Systems' at e4ds Tech Day 2026, emphasizing that "AI is no longer optional but has become an evaluation criterion determining orders and survival." In the SDV era, vehicle software development is being reorganized around companies with AI-native systems, and it has become an essential requirement with 20~35% of technical evaluation points allocated to AI and automation capabilities in domestic bidding processes. Development periods are being compressed from the existing one year to approximately four months, and Japan and China have already implemented AI-based full-process automation to drive new vehicle development within 18 months. Since generic AI alone cannot satisfy AUTOSAR and A-SPICE requirements, PopcornSAR has established ontology-based knowledge systems and the AI SDV platform PARVIS to automate requirements, design, code, and testing. The hybrid development approach combining AI and rule-based engines is effective, and developers are shifting from the role of writing code to reviewing and judging AI results.
Comments
Related news
"Beyond the Era of Knowing Technology to the Era of Proving Technology"…e4ds Tech Day 2026 Concludes Successfully
e4ds' 'e4ds Tech Day 2026' held on September 8th at the Korea Science and Technology Hall concluded successfully. This event, centered on SDV and edge AI, focused on presenting practical application standards and utilization directions in actual industrial settings rather than on technology itself. CEO Myung Se-hwan emphasized the 'era of proving technology' and presented the importance of practical decision-making criteria, while Director Gwak Su-jin of the Korea Automobile Research Institute outlined the future mobility ecosystem based on AIDV and Physical AI. Subsequently, Shin Dong-joo, CEO of Mobilint, explained the role of edge AI semiconductors as core infrastructure in the Physical AI era. At the end of the event, the TechDay Contest 2026 awards ceremony and winning project presentations were held, sharing actual technology implementation cases and presenting a vision for building a collective intelligence ecosystem in the industrial field. The awards in this ceremony were presented respectively by Mouser Electronics and STMicroelectronics, both platinum sponsors of e4ds Tech Day 2026. Daphne Tien, Vice President of Mouser, and Matteo Marabita, Center Director of STMicroelectronics APAC AI Competence Center, presented awards directly to the winners.
2026.09.09by 배종인 기자
[e4ds Tech Day 2026 ①] MPS Director Lee Jun-ho, "Power Semiconductors Will Determine Automotive Performance in the Age of AI Autonomous Driving"
Director Jun-ho Lee of Monolithic Power Systems (MPS) presented on the theme "Power Architecture and the Future of Power Conversion for the SDV Era" at the 'e4ds Tech Day 2026' held on the 8th, emphasizing the importance of power semiconductor technology, stating, "The essence of the SDV era is not software, but the power infrastructure that supports it." Director Lee analyzed that in the SDV era, autonomous driving, artificial intelligence (AI), and high-performance computing within vehicles are becoming essential elements, and high-density power design technology to reliably support these is emerging as a key factor determining automotive competitiveness.
2026.09.15by 배종인 기자
[e4ds Tech Day 2026 ②] Eugene Yoo, Nordic Director, "Innovating Edge AI with Axon NPU and Neuton"…Nordic presents realization of 'On-Device AI'
Nordic presented an ultra-low-power on-device AI strategy centered on Axon NPU, Neuton, and Edge AI Lab. Through an MCP server-based development environment, the company significantly enhanced development productivity by connecting SDKs, documentation, and cloud data with AI. Axon NPU delivers inference speeds up to 15 times faster than CPU and high power efficiency. Neuton performs AI inference with less than 5KB of memory on existing MCUs, achieving speeds up to 10 times faster and ultra-low-power characteristics. Edge AI Lab simplifies the development process by automating model generation and training with just data uploads. Nordic is building an integrated ecosystem spanning wireless SoCs, PMICs, cloud, and security to aggressively pursue the on-device AI market.
2026.09.15by 배종인 기자
[e4ds Tech Day 2026 ③] Matteo Maravita, ST Center Director: "Generative AI is Also Moving to the Edge"
STMicroelectronics highlighted the surging data center power issues due to generative AI proliferation at e4ds Tech Day 2026, emphasizing that the key to AI competition is not 'TOPS' but 'TOPS/W' and sustainability. ST presented an on-device AI strategy based on MCU and MPU to address the limitations of cloud-centric AI. The STM32N6 delivers a highly efficient NPU with 600GOPS and 3TOPS/W, achieving 600x performance improvement compared to the existing STM32H7. Additionally, ST adopted in-memory computing (IMC) to minimize memory movement and maximize energy efficiency. ST aims for low-power AI solutions in the 5-10TOPS range and is expanding the embedded AI ecosystem with next-generation AI technologies such as SLM, Agentic AI, and Physical AI.
2026.09.15by 배종인 기자
[e4ds Tech Day 2026 ④] Sung-il Lim, Executive Director at TI, "Edge AI is no longer a high-cost SoC domain"
Sung-il Lim, Executive Director at Texas Instruments (TI), emphasized at the 'e4ds Tech Day 2026' held on the 8th that edge AI is no longer confined to the domain of high-cost SoCs, and an era has opened where it can be implemented even in one-dollar MCUs. The key is the company's own NPU 'TinyEngine,' which reduces AI inference latency by 90 times compared to MCUs and saves 120 times more power. By separating control and inference through a parallel structure of MCU and NPU, efficiency has been improved, and it can be applied across various industries including predictive maintenance, home appliances, wearables, and robots. The AM13E23019 MCU integrates Cortex-M33, TinyEngine, and analog functions to reduce BOM costs by 30%. TI is also strengthening the development environment with Edge AI Studio and AI-based CCStudio, emphasizing that "the essence of edge AI is not high performance but product commercialization."
2026.09.15by 배종인 기자
