LG CNS expands role in robot, AI training for manufacturers

South Korea’s LG CNS, an IT solutions firm, is expanding its role in training and managing physical AI, including humanoid and industrial robots, for manufacturing companies.

The Seoul-based company does not build robots or develop AI models but advises on selecting suitable robots, trains them for specific factory needs, and manages workflows such as task assignment and maintenance.

LG CNS works with various AI providers, including Google Gemini and OpenAI’s ChatGPT, and hardware from US, Chinese, and South Korean firms.

The company aims to serve manufacturers that do not develop their own robotics or AI systems, positioning itself as a bridge in the growing automation market.

The sector is expected to expand as robot prices decrease and companies seek to improve efficiency and safety, with projections estimating the humanoid robot market could reach US$38 billion by 2035.

LG CNS also announced a partnership with a biotech firm to develop AI-powered health services.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

The company is developing specialized robot foundation models and management tools

LG CNS does not build robot hardware or create general-purpose AI models. It is building a specialized Robot Foundation Model (RFM), described as the AI brain of robots, through its Future Robotics Lab 1.
The platform lets manufacturing teams run their own RFMs. It covers data collection and cleaning, simulation-based fine-tuning, plus day-to-day operation and monitoring of trained models 1.
The work fits into an LG Group push into “physical AI.” Affiliates such as LG Electronics provide hardware such as robotic actuators (the components that move a robot’s joints and parts) 2.
The “AX” AI transformation business is bringing in more revenue. The AI and cloud divisions generated ₩3.5872 trillion last year, up 7% 3.

The firm’s model shifts value from building robots to putting them to work

LG CNS argues that as robot hardware and general AI models become standardized, more value will sit in integration work that gets robots running on factory floors 4.
CEO Hyun Shin-gyoon says the differentiator will be applying and running AI in real industrial settings. He calls this role the “on-site application specialist” 4.
The approach adds a layer between robot makers such as Agility Robotics (a US company that builds humanoid-style robots for warehouses and factories) and the day-to-day constraints of factory operations 4.
Hyun estimates the move from proof of concept (PoC) to robots working on production lines can take about two years. He ties that timeline to building mass production systems, not to uncertainty about the technology 4.