Three separate Korean industrial moves in a single week point to a coordinated push to own the full robotics value chain for manufacturing, from on-device AI silicon to the software layer that directs heterogeneous robot fleets across factory floors.
Why it matters
Factory automation is entering a phase where general-purpose robots, not fixed-function arms, are expected to handle variable tasks. The companies positioning to supply, program, and orchestrate those robots will shape how manufacturing scales in the next decade. ANALYSIS The near-simultaneous announcements from Samsung SDS, LG Electronics with its chip partners, and Doosan Robotics suggest Korea's largest industrial groups are each staking out a distinct layer of that stack rather than competing head-on.
The big picture
Samsung SDS officially declared its entry into what it calls the Robot Transformation (RX) business at Real Summit 2026 in Seoul on September 84. The company plans to launch a Robot Orchestration Platform in 2027 that integrates production equipment and robots for unified operations. The platform will feature collision prevention between heterogeneous robots and automatic rerouting instructions in the event of malfunctions. To populate that platform with hardware, Samsung SDS formed the Team REX partnership with 10 physical AI companies including Rainbow Robotics and Walden Robotics.
Walden, spun off from the Toyota Research Institute in January, is the linchpin of the Samsung SDS hardware strategy1. Samsung SDS made a strategic investment in Walden through Samsung Venture Investment in June. The two companies are now pursuing a "robot orchestration" business that directs different types of robots through a single control network. Samsung SDS has already run technical validation at production sites operated by affiliates including Samsung Electro-Mechanics, SEMES, Samsung Heavy Industries, and Samsung E&A. Walden raised $300 million in a funding round in July, co-led by Toyota and Deviation Capital, with Nvidia, Boeing, and Samsung Venture Investment among the main investors.
Walden's robot pairs a two-armed upper body with a wheeled mobile base and is designed to handle a range of tasks in manufacturing and logistics settings. The company applies large behavior models to its robots and began deploying general-purpose robots on production lines at Toyota's North American plants in February.
One layer down the stack, LG Electronics, Nota, and Mobilint will jointly develop a humanoid robot powered by a domestically produced neural processing unit3. Nota was selected as the lead institution for the humanoid category of the K On-Device AI Semiconductor Technology Development Program, run by the Ministry of Trade, Industry and Energy. Mobilint will supply the NPU; Nota will optimize the vision-language-action model embedded in the humanoid to the chip's specifications using token compression, quantization, pruning, and knowledge distillation. A humanoid with an embedded VLA model can recognize its surroundings and act without relying on external servers, minimizing communication latency.
Doosan Robotics, meanwhile, is building toward industrial humanoids on a different timeline. CEO Kim Minpyo stated at the 2026 Global Machinery Technology Forum, "The issue is not whether we will pursue humanoids or not, but the order in which we do so. We will first define task intelligence and then build bodies that fit those needs"2. Doosan's roadmap runs from collaborative robots to intelligent robot solutions to industrial humanoids, with a proof-of-concept target of 2028.
ANALYSIS Each Korean player is occupying a different altitude. Samsung SDS is positioning as the integration and orchestration layer, the fleet-management brain that sits above any single robot vendor. LG and its partners are tackling the inference hardware problem, aiming to run VLA models on Korean-made silicon inside the robot itself. Doosan is taking a data-first path, collecting operational intelligence from collaborative robots already deployed before scaling up to humanoid form factors.
Samsung SDS CEO Lee Jun-hee emphasized that the company's AX strategy "requires fundamental business redesign beyond simple AI adoption" and announced plans to train approximately 200 field deployment engineers. ◆ That language frames robotics not as a product sale but as a systems-integration engagement, the kind of recurring-revenue model that IT services firms prize.
Walden's investor roster is telling. Nvidia, Boeing, Toyota, and Samsung Venture Investment all participated in the $300 million July round. ◆ The overlap between Walden's backers and Samsung SDS's affiliate network creates a tight feedback loop: Toyota supplies the robotics research lineage, Samsung SDS supplies the factory floor, and Nvidia supplies the compute substrate.
What's next
Samsung SDS targets its Robot Orchestration Platform launch in 2027. The company also plans to expand AI infrastructure through the National AI Computing Center in 2028 and the Gumi Center in 2029. Doosan Robotics aims to conduct proof-of-concept testing for industrial humanoids by 2028. Kim Minpyo offered a framing for what comes after: "In the future, the primary operator of robots will be agents, not humans".