Neuromeka

Korea funds physical AI through at least four ministry programmes and several public-private alliances, and the largest awards disclosed so far have gone to established manufacturers rather than startups. Of KRW 444.9 billion in project budgets that MEKAWIRE has verified from company and institution releases, Daedong leads KRW 297.7 billion, Doosan Robotics KRW 98.9 billion and Neuromeka KRW 17.5 billion, and no project is led by a venture-backed startup; startups enter as consortium members, chip suppliers or through regulatory sandbox approvals.

Korea's Ministry of Science and ICT said on Oct 14 that its R&D policy director Kim Sung-soo met Nature editors, including Nature Sensors chief editor Olga Bubnova, KAIST president Bae Choong-sik and KAIST robotics professors Myung Hyun and Shim Hyun-chul over lunch on the sidelines of the 2026 Nature Conference on autonomous robotics, held by KAIST and Nature in Daejeon from Oct 13 to 15, Robot Newspaper (irobotnews) reported from the ministry's announcement. According to the report, the group discussed the core technologies and unsolved problems standing between humanoids and commercial use, how Korea's manufacturing and research base could be turned into a competitive edge, and ways to expand international joint research. Kim said the ministry would push a humanoid ecosystem through its three mega-projects and the K-Moonshot programme and would use the views gathered at the meeting in future humanoid R&D policy; first vice minister Koo Hyuk-chae, in a video address to the conference, described humanoids as central to physical AI. No new budget was announced. The ministry's largest live commitment in MEKAWIRE's records is the KRW 1.41 trillion national physical-AI R&D programme for North Jeolla and South Gyeongsang from 2026 to 2030, under which Neuromeka leads a KRW 17.5bn shipyard welding project (reported Sept 28).

Korean servo-drive and motion-control maker RS Automation is building an agentic-robotics platform meant to link domestic robot component makers through a common interface and test bench, Robot Newspaper (irobotnews) reported on Oct 14, citing the company. According to the report, a higher-level AI would use prior knowledge of motor and reducer characteristics, load inertia, friction, stiffness and transmission type, correct it with data gathered in operation, and pass the result to RS Automation's real-time servo control, so that a drive does not have to learn every condition by trial and error; motor, reducer, encoder, servo-drive, actuator and robot-hand makers would plug in their products and compare performance by combination. The company named humanoid robots and AI factories as the main targets. No timeline, partners, budget or product release were given, and the plan does not appear on RS Automation's own press page. MEKAWIRE context: Korea still imports most precision reducers and high-end servo parts for humanoid joints, and in MEKAWIRE's coverage the government money this autumn has gone to system builders such as Doosan Robotics and Neuromeka rather than to component integration, which is the gap RS Automation is pitching for.

Korean cobot maker Neuromeka said on Oct 2 that it and the research group of Professor Kim Min-jun at KAIST's School of Electrical Engineering developed SCDP (Spatially Conditioned Diffusion Policy), a control method that lets its EIR humanoid do fine two-handed work using only a single RGB camera in its head, Robot Newspaper (irobotnews) reported; the paper was accepted at the Conference on Robot Learning (CoRL) 2026, and MEKAWIRE has not yet seen the paper or a company release, so the figures below are as reported. SCDP predicts where the robot's fingertips will travel, projects that path onto the camera image and pulls visual features from several resolutions around the expected contact area, re-selecting them as the motion is refined. According to the report, EIR separated a stack of nested paper cups with both hands without any wrist camera, and in Meta-World simulation with 20 demonstrations per task SCDP averaged 82.5% success on the six hard tasks against 74.2% for a standard diffusion policy that also had a wrist camera; in real cup-grasping trials on a Franka Panda arm with visual clutter it scored 85% (handle) and 80% (wall) against 30% and 65% for the baseline, over 20 trials each. Neuromeka CTO Heo Young-jin said removing wrist cameras and cabling frees space around the hand for work with human tools in tight cells. The research follows Neuromeka's role as lead organisation in a KRW 17.5bn science-ministry project on autonomous shipyard welding, which MEKAWIRE reported on Sept 28, and comes five days before Tomorrow Robotics published its Habilis Brain 0 manipulation model.

Korean cobot maker Neuromeka said on Sept 28, in a statement posted to its newsroom, that it will lead a project under the Ministry of Science and ICT's programme to develop and test human-AI collaborative Large Action Models, focused on welding the curved outer-hull blocks used in shipbuilding. The project runs from August 2026 to December 2030 with a total budget of KRW 17.5 billion, of which KRW 10.5 billion is government R&D funding, and includes Neuromeka's subsidiary Robolution and other industry, academic and research partners. Neuromeka plans to build combined datasets of block geometry, weld-seam positions, robot motion, welding conditions and quality readings, then use them to train a model that lets the robot find seams, plan its own path and correct trajectory and settings in real time as parts distort under heat. Chief executive Park Jong-hoon said the company aims to turn the results into a standard autonomous welding platform for shipyards and later offshore plants, steel structures and defence. The ministry's wider physical-AI programme commits KRW 1.4131 trillion in North Jeolla and South Gyeongsang from 2026 to 2030. Earlier in September, Pohang City said Neuromeka signed an investment MOU for a KRW 180 billion robot foundry and data centre in the city, as reported by EBN and reposted on Neuromeka's site; on Oct 6 Doosan Robotics said it had won two government projects worth about KRW 98.9 billion, one of them for intelligent cobot welding.

Neuromeka (KOSDAQ: 348340) said on July 22 that automotive parts maker DIC (KOSDAQ: 092200) has picked it as supply partner for a gear-machining automation project, according to an Edaily report the company reposted on its newsroom. DIC makes core transmission parts and reducers for electric vehicles and is using the investment to move towards a smart factory. Neuromeka will deliver what it calls a hybrid automation architecture that adds AI-driven cobots to DIC's existing gantry loader system, with AI situational awareness, real-time collision avoidance and learning-based teaching meant to cut set-up time when the line switches between products, so that several part types can run on one line. The equipment will also be networked over Ethernet so that quality data, machine status and maintenance history feed into DIC's manufacturing execution system. The contract value was not disclosed. The two companies had earlier signed an MOU with autonomous-robot maker Twinny to co-develop an AI mobile humanoid cobot. Chief executive Park Jong-hoon said the selection shows the company's process design and system integration skills in precision auto-parts machining. In MEKAWIRE's records this is one of several Neuromeka industrial deals in 2026, after a rights offering in April and before the KRW 17.5 billion shipbuilding welding project it won in September.

Korean cobot maker Neuromeka said on July 21 that it has signed a strategic memorandum of understanding with Samsung C&T to put robot services into commercial buildings, according to an iNews24 report the company reposted on its newsroom. Samsung C&T's Bynd is a cloud building platform built on AI and digital twins that links a building's main equipment and services, so occupants can talk to an AI assistant to adjust lighting and air conditioning or call a robot. Neuromeka will supply robot management and control, tying its robot platform and AI-based operations software to Bynd's spatial data, digital twin, smart access, parking and energy functions. The two plan joint research on physical-AI building technology, a proof of concept linking the two platforms, joint bids in Korea and abroad and a longer-term roadmap, and want to win references through real commercial building projects before expanding. Chief executive Park Jong-hoon called the deal a step that takes Neuromeka's physical AI from the factory floor into smart buildings. No financial terms were disclosed. Building service is a new market for Neuromeka, whose recent wins in MEKAWIRE's records are industrial: a KRW 17.5 billion national project for autonomous ship-hull welding in September and a KRW 180 billion robot foundry MOU with Pohang.

The International Federation of Robotics said on 25 September 2025 that 542,000 industrial robots were installed worldwide in 2024, the second-highest annual total on record and 2% below 2023. Japan (44,500 units, down 4%), the Republic of Korea (30,600, down 3%) and Germany (26,982, down 5%) all contracted, while China installed 295,000 units and domestic Chinese makers took 57% of their home market for the first time.

South Korea's Ministry of Trade, Industry and Energy launched the K-Humanoid Alliance in Seoul on 10 April 2025, with about 40 companies, universities and other organisations signing a cooperation agreement aimed at making Korea the world's leading humanoid nation by 2030. The ministry expects more than 1 trillion won of public and private investment by 2030 and set 2028 targets for a shared robot AI foundation model and for humanoids under 60kg with more than 50 degrees of freedom.