As the robotics has progressed, the industry has consistently adopted more robots to automate many types of grunt work. More than 540,000 new industrial robots settled worldwide in 2023, receiving the number of total industrial robots active above 4 million, per IFR.
Industrial robots usually excel in repetitive tasks, but they find it difficult to perform precise tasks, handle sensitive materials and adapt to changing conditions – a robot in the kitchen of a restaurant would make it more useful, for example. That is why many industrial processes are still manual.
Start of South Korea Rlwrld Its purpose is to solve this problem with a fundamental AI model that has made robotics specifically by combining large linguistic models with traditional robotics software. The company says this model will allow robots to make quick and agile movements and execute some “reasonable reasoning”.
“The use of the RLWRLD foundation model, the processes that require a lot of manual work can be fully automated by learning and copying human expertise, making the work environments more effectively,” said Jung-Hee Ryu, founder and CEO of RLWrld.
The start comes from Stealth with 21 billion KRW (about $ 14.8 million) in seed funding. The round was driven by the business capital company, and Mirae Asset Venture Investment and Global Brain also invested.
Specifically, Rlwrld has attracted a long list of major strategic investors – Ana Group, Pksha, Mitsui Chemical, Shimadzu and KDDi from Japan. LG Electronics and Sk Telecom from Korea. and Amber Manufacturing from India.
RLWRLD said seed funding would be used to finance the idea of proof projects with its strategic investors. Safe computing infrastructure such as GPU, market robots and devices to collect extensive data. and hire top research talent. The start will also use the new money to develop advanced hand movements that include five fingers-a capacity that has not yet been proven by its competitors such as Tesla, Figure AI and 1X, Ryu said.
Ryu said Rlwrld is also working with its strategic investors to explore ways of automating differently Work Flows with Human using the AI model. They are preparing a demonstration of autonomous humanoid -based action, scheduled for later this year, Ryu said. In addition, the company is working to develop a platform that can support various types of robots, including industrial, collaborative, autonomous mobile robots and anthropoids.
Founded in 2024, Rlwrld is Ryu’s third start. Its second start, Olaworks, acquired by Intel in 2012And finally it became the Korean R&D center in the Computer Vision. And 2015 founded A boot accelerator, future gameThis focuses on deep technology companies.
When asked what inspired him to start a new company, Ryu said he noticed how quickly the newly formed AI companies were increasing in the US, Europe and China, while comparable newly formed AI companies in Korea and Japan were absent.
He spoke with more than 30 AI teachers from Korea and Japan about their challenges – everything from lack of infrastructure such as data and GPUs and obstacles that discouraged them to start a business – and opportunities available.
“We have determined that it would be strategically beneficial to prioritize the Robotics Foundation (RFM) models above the technologically saturated LLMS field, utilizing Korea and the notable global advantages of Japan in construction,” he said.
Shortly afterwards, he brought six teachers from top institutions in South Korea, including Kaist, Snu and Postech, along with their research teams, to launch the RLWRLD.
RLWRLD is not alone in dealing with this problem. The newly established companies such as Skild AI and Natural intelligence They create similar fundamental models for robotics, as are the largest businesses such as Tesla, Google Deepmindand nvidia.
But Ryu believes that his start is a good start, as he already has AI and robotics experts he needs to develop fundamental models for robotics, as well as humanoid robots with high degree of freedom (DOF).
“Additional, [such companies] They are usually based on low dof robot, such as predators with two fingers. The RLWRLD has already secured a high DOF reference robot and therefore expects higher performance results, “he said.
Ryu also said that thanks to its strategic investors, the RLWRLD can quickly collect valuable data from nearby construction. In 2024, a reference It showed that Japan and South Korea collectively represented 9.2% of world production production.
The RLWRLD aims to generate revenue from this year through proof of proof projects (POC) and demonstrations of cooperation with strategic partners.
The long -term goal of starting is to serve factories, logistics and retail stores, and even robots that can be used in domestic environments to help homework. In the meantime, the priority is to target industrial data, as they are willing to pay for most and have a strong demand for automation.
The start has 13 employees.
