Siemens differentiates itself from Palantir by handling large-scale industrial data and integrating with existing manufacturing software, leveraging a knowledge graph to improve AI context.

Pilot projects are underway with leading Korean electrical/electronics, shipbuilding, and semiconductor companies, with active collaboration from major industry groups including HD Hyundai; full-scale AI projects are expected to begin next year.

Trust, governance, and appropriate authority controls are emphasized as determinants of an AI agent's level of autonomy in manufacturing environments.

In a demonstrated use case, an AI agent detects a high-bandwidth memory (HBM) yield abnormality, traces the cause across manufacturing data, drafts an isolation plan, and auto-updates the MES after human approval, cutting response time from days to about six minutes.

Siemens is expanding into physical AI, pursuing autonomous factory capabilities with partners including Nvidia, Boston Dynamics, and domestic robotics firms, with collaborations spanning virtual factories and robot training, exemplified by the Foxconn Mexico AI factory project.

Siemens Digital Industries Software unveiled Intelligence Center X (ICX), an enterprise AI agent platform, in Seoul, with ambitions to scale AI adoption in Korea starting next year.

Siemens identified six target Korean industries—semiconductors and memory, shipbuilding, batteries, automobiles, displays, and steel—to boost production efficiency and automate workflows.

ICX consolidates data across PLM, ERP, and MES without copying data, using Graph Studio, AI Studio, and Mendix to enable AI agents to analyze and act on manufacturing information.

Siemens CTO Simon Guerin cites data context as a major barrier to AI ROI and advocates for knowledge graphs to better capture relationships within industrial data.