Siemens 與 HD Hyundai 測試 ICX 平台,瞄準韓國六大產業
Siemens Tests ICX Platform With HD Hyundai, Targets 6 South Korean Industries
Siemens 已與包括 HD Hyundai 在內的韓國企業展開企業 AI 智能體平台 ICX 的小規模試點,並瞄準半導體、電池、顯示屏、汽車、造船及鋼鐵六大產業。ICX 以知識圖譜連接 PLM、ERP 和 MES 數據;Siemens 稱,在半導體 HBM 良率下降的情境下,平台可將問題處理時間由數天縮短至 6 分鐘。

Siemens is targeting major South Korean manufacturers with its enterprise agentic systems platform, Intelligence Center X (ICX), having initiated small-scale pilot tests with leading local companies including HD Hyundai in the electrical and electronics as well as shipbuilding sectors.
Siemens Digital Industries Software, a software subsidiary of Siemens, introduced ICX at a press briefing for Agentic Enterprise Connect (AEC) 2026 at Josun Palace in Seoul’s Gangnam district on Oct. 13.
ICX is an industrial artificial intelligence (AI) platform designed to consolidate data scattered across an enterprise and enable AI agents to operate in real-world business environments. It connects data from product lifecycle management (PLM), enterprise resource planning (ERP), and manufacturing execution systems (MES) through a unified knowledge graph.
Siemens is initially targeting six major South Korean industries with ICX: semiconductors, batteries, displays, automobiles, shipbuilding, and steel.
“We are currently working with several major South Korean companies on small-scale scenarios to validate the technology through proof-of-concept (PoC) projects and assess the return on investment (ROI),” said Lee Sua, Vice President at Siemens Digital Industries Software.
Asked to identify customers beyond HD Hyundai, Lee said the company is conducting pilot projects with leading South Korean businesses in the electrical and electronics, shipbuilding, and semiconductor sectors. She added that validation takes an average of three months and that no projects have been completed yet.
ICX can also connect easily to data from Siemens' Teamcenter PLM software. Teamcenter integrates and manages data across the product lifecycle, from planning and design to development and manufacturing. According to Siemens, existing Teamcenter customers can more easily pursue AI transformation (AX) because they do not need to reprocess their data separately. The company is therefore seeking to expand ICX adoption among its existing Teamcenter customers, including HD Korea Shipbuilding & Offshore Engineering, Hyundai Motor, and Kia.
The ICX platform consists of three layers: Graph Studio, AI Studio, and Mendix. Graph Studio links data distributed across departments and systems to establish context. AI Studio uses models trained on company-specific data to deliver AI capabilities based on that context. The third layer, Mendix, enables frontline workers to develop and operate business applications using the resulting outputs. The ultimate goal is to enable AI agents to understand corporate data and business rules, analyze problems, and carry out subsequent tasks.
Siemens cites context as the key differentiator between ICX and other AX platforms. Consider a semiconductor manufacturing facility experiencing a decline in high-bandwidth memory (HBM) yields. Under conventional workflows, employees would need to move between PLM, MES, and other systems to identify the cause, a process that could take several days. Siemens says ICX could reduce the time needed to resolve the issue to just six minutes, as its AI can more readily understand the problem by connecting data across multiple systems.
The platform also incorporates defenses against prompt injection to strengthen operational technology (OT) security. It can restrict abnormal AI behavior by defining agents' permitted actions and accessible tools in advance, while granting conditional system access based on user roles. Siemens said decision paths can also be traced and corrected when problems arise.
ICX was first unveiled at an event in Detroit in June. Siemens said it has since secured customers in industries including aerospace, defense, and financial services.
Siemens Digital Industries Software's South Korean subsidiary reported revenue of 185.9 billion won and operating profit of 7.9 billion won last year. Revenue increased 11.4% and operating profit rose 17.9% from a year earlier.
來源:The Elec · thelec.net