[NoFilterRobot] "From Human Hands, Not 100 Robots"... WhatsLab Targets 'Robot Data'

Interview with Sim Donghyun, CEO of WhatsLab

Data Accumulates as Workers Wear Gloves on the Job

Participating in NVIDIA Inception and Collaborating in Robot Development

"Will Expand to Sensory Data Such as Tactile Information"

The core of competition in the robotics industry is shifting from hardware to data. This is because for robots to move like humans in actual industrial settings, they require an advanced learning process for sophisticated movements. WhatsLab, a robot learning data solution company, has targeted this niche. While countless companies focus on deploying robots to industrial sites, WhatsLab has taken a different approach by converting the very process of human work into data.

Sim Donghyun, CEO of WhatsLab, met with The Asia Business Daily on the 22nd at the "Humanoid Summit Seoul 2026", a global Robot and AI (Artificial Intelligence) conference held at COEX in Seoul, and said, "While robot hardware and models are important, the biggest bottleneck in the industry right now is data. Our goal is to quickly resolve the data bottleneck."

Shim Donghyun, CEO of WhatsLab, is being interviewed by The Asia Business Daily at the "Humanoid Summit Seoul 2026," a global robot and AI (Artificial Intelligence) conference held on the 22nd at COEX, Seoul. Photo by Jooni Park.

The "Air Glove Atlas" developed by WhatsLab is a glove-type controller equipped with sensors that measure the movements of fingers and wrists. When a worker wears the Atlas during actual work, hand motion data is collected. This data is then converted through dedicated software called "Spine" and can be used to control real robots remotely or serve as training data for robots.

One key differentiator of WhatsLab's product is that it allows data collection without the need for a robot, unlike other products on the market. Traditional methods require both robots and human operators or auxiliary devices.

Sim emphasized, "If you want to collect data using 100 robots, you need to purchase all 100 robots and arrange space to install them. In contrast, we allow workers to wear the device while doing their actual tasks, and the data naturally accumulates. This significantly reduces both the costs and time required to collect data."

WattsLab, a robot learning data solution company, is showcasing its glove-type device product line at its exhibition booth at the global robot and AI conference "Humanoid Summit Seoul 2026," held on the 22nd at COEX in Seoul. Photo by Junyi Park.

To collect data in real industrial environments, it is also important to ensure the comfort and convenience of the worker wearing the equipment. WhatsLab is improving the size and shape of its products to reduce discomfort for workers in the process of gathering robot learning data.

Sim noted, "Field operators have to wear the product and collect data, and in a situation where they already have to gather data for robots that could potentially replace their own jobs, any discomfort with the devices could make them even more resistant. That's why we've been developing products that are smaller in size and less uncomfortable to use."

From Virtual Training to Robot Learning

Actual demonstration video of the virtual training content presented by WhatsLab at 'MWC Barcelona 2026'. Screenshot from WhatsLab's YouTube channel.

WhatsLab began as an extended reality (XR) virtual training business, expanding into the robotics sector in January this year through a partnership with Robotis. It has also joined NVIDIA's startup incubation program "Inception," receiving support necessary for robot learning data processing and validation.

To enable the use of hand motion data for robot development, WhatsLab has integrated its products with major development software. It has collaborated on development with Robotis and Rovros, while Samsung Electronics and Wonik Robotics have purchased the products for use in robotic research.

Going forward, WhatsLab plans to collect not only hand movement data but also sensory information such as tactile feedback. Sim stated, "In the long term, I would like to collect all types of sensory information humans can perceive, not just tactile feedback but also pain and temperature."

The upcoming product "Axis" will feature functions that collect both hand movement and tactile sensory information. Starting next month, WhatsLab plans to attract new investment to secure funding for further technology development.

This content was produced with the assistance of AI translation services.