South Korea's government has unveiled an ambitious initiative to integrate physical AI into the nation's manufacturing sector, committing 1.4131 trillion Korean won (approximately $1.01 billion) over the next five years. The program aims to expand AI's role beyond defect detection to encompass production control, equipment coordination, and overall factory management, with plans to deploy mature technologies across domestic manufacturing sites and export the concept as a "K-Factory" model.

The Ministry of Science and ICT held a briefing on September 16 in Seoul to outline the 2026-2030 roadmap for physical AI research projects in North Jeolla and South Gyeongsang provinces, as reported by local media. A significant aspect of this initiative is the integration of tacit knowledge and process expertise accumulated by manufacturers, data that has traditionally remained within factory walls, into the national R&D framework, enabling AI to move from detecting abnormalities to understanding workflows and directly controlling equipment.

Under the plan, the two provincial projects will undergo initial technology validation, with successful outcomes to be scaled to manufacturing bases nationwide. The scope of technology development spans neural processors, digital twins, robotic collaboration, and industrial data infrastructure, among other critical components.

1.4 Trillion Won Investment Targets Precision Manufacturing and Factory Orchestration

The total R&D budget of 1.4131 trillion won allocates 736.8 billion won to the North Jeolla project, which focuses on factory platforms, while 676.3 billion won is directed to South Gyeongsang for precision manufacturing. Each region approaches the challenge differently: one tackles how factories operate collaboratively, while the other explores how AI can comprehend and control production processes.

The South Gyeongsang project is developing so-called "large action models" that aggregate motion data from workers and robots, physical laws, and spatial information to enable AI to understand specific production flows and operate real equipment. Digital twin technology, data pipelines, and neural processor infrastructure will connect data training, field application, and result feedback loops.

Meanwhile, the North Jeolla project centers on "factory orchestration," aiming to enable robots and equipment from different manufacturers to work together, managing the entire facility as a unified system. AI will handle factory layout design and logistics robot pathing, validating solutions through three-dimensional digital twins before deploying them to physical equipment, ultimately progressing toward fully automated "lights-out" factories where AI oversees design, construction, and operations.

During a proof-of-concept trial last year, AI completed a robotic factory design in roughly three hours, a task that previously required three to four specialists working for nearly a month, according to project officials.

13 Companies Share Tacit Knowledge, Shifting AI from Detection to Production Control

A key milestone of this project is the first-time inclusion of tacit knowledge from production floors in a national R&D program. Cha Suk-won, a Seoul National University professor leading the South Gyeongsang project, noted that 13 companies have agreed to share tacit knowledge data from their factories. The collected data will be processed and used to build physical models for different process stages, which can then be applied to other manufacturers with similar production workflows.

While manufacturing AI has previously focused on defect detection and anomaly identification, this initiative seeks to leverage accumulated factory data and experience to help AI grasp specific operational procedures and process rules, enabling it to participate in equipment control decisions.

Chang Young-jae, a KAIST professor directing the North Jeolla project, proposes viewing a factory as "a giant robot composed of various machines," coordinating diverse equipment through unified scheduling to achieve holistic plant synergy. Shinsung E&G, a semiconductor and display materials supplier, has committed 20 billion won to participate in building the related physical AI industrial ecosystem.

From Pilot Facilities to Global Exports, South Korea Pursues an Integrated Manufacturing AI Solution

The government's ultimate objective extends beyond the two regional demonstration projects, with plans to gradually spread proven technologies to all manufacturing bases nationwide and create an exportable "K-Factory" proposition. Between 2026 and 2030, R&D and validation will concentrate in North Jeolla and South Gyeongsang, followed by deployment of physical AI capabilities across additional manufacturing facilities.

The strategy combines South Korea's decades of accumulated process data and production expertise with AI, robotics, and digital twin technologies to deliver comprehensive solutions covering factory design, equipment coordination, and operational management. This signals a shift from single-point AI applications to systemic factory-level transformation: AI is no longer simply identifying anomalies on production lines, but increasingly shaping factory design, equipment control, and production scheduling, while positioning South Korea to export complete "AI factory" packages to international markets.