[AI Leaders] SKAI Intelligence: "This Is the First Revenue Year for Robotic Synthetic Data"

[ZDNet Korea]

"Synthetic data for industrial robots has started generating revenue this year. At present, projects are carried out individually through a closed-loop process. However, as we have verified continuous improvements in efficiency, we expect revenue to grow rapidly once data production reaches mass scale."

Lee Jae-cheol, CEO of SKAI Intelligence, made the remarks during an interview with ZDNet Korea at the company's headquarters in Gangnam, Seoul, late last month.

SKAI Intelligence is a digital twin and synthetic data infrastructure company that generates synthetic data representing real-world physical environments. It is the only Korean company listed as an official NVIDIA Retail Partner. In the advertising sector, the company has secured AI content production revenue from brands under LVMH through its 3D content automation platform, B.THREE.

The synthetic data developed by SKAI Intelligence serves as both a training dataset for robots before deployment and a "failure notebook" that records situations where robots may malfunction or make incorrect decisions. The data includes object positions, poses, labels, camera and lighting conditions, and material properties, enabling immediate AI training while also allowing engineers to trace under which conditions an AI model produced errors.

The company's core competitiveness lies in its Real2Sim2Real closed-loop framework. Using CAD models, material characteristics, and tolerance information from real products, SKAI Intelligence builds digital twins and generates synthetic data within the simulation environment. Errors and failures observed in real production are then fed back into the system for continuous refinement. Rather than creating data once, the platform continuously improves through real-world feedback. This enables customers to identify vision-related risks before production lines are fully operational, significantly reducing on-site debugging costs.

Based on these capabilities, SKAI Intelligence signed a Strategic Cooperation Framework Agreement (CFA) with ABB Robotics, one of the world's four largest industrial robotics companies, last month. According to Lee, although several global synthetic data companies competed for the project, they were ultimately not selected. While individual technologies can be replicated, the integrated industrial data pipeline—combining manufacturing expertise, high-fidelity digital twins, automated labeling, and validation experience with robotics companies—creates a significant barrier to entry.

"You need to understand both industrial production environments and simulation systems to implement the entire mechanism. Very few companies truly possess both capabilities, and that is precisely why ABB Robotics selected us," Lee said.

The two companies are conducting proof-of-concept (PoC) projects every month, accumulating quantitative performance metrics. Their immediate priority is to extend the closed-loop framework into mass production and demonstrate measurable cost savings for manufacturers.

"The outcomes of these projects will be announced progressively during the second half of the year," Lee said. "Once validation is completed on ABB Robotics' digital twin infrastructure, we plan to expand into other precision applications, such as medical robotics, where robots recognize objects through optical systems and perform highly accurate operations."

After completing validation with ABB Robotics and other global partners, SKAI Intelligence plans to accelerate its expansion into the Korean market. Korea represents an attractive opportunity due to its concentration of precision manufacturing industries, where assembly tasks involving components as small as 0.5 to 3 millimeters are commonplace.

"As early as the second half of this year—or next year at the latest—we intend to enter Korea's precision manufacturing market and begin commercialization," Lee said. "Our target applications range from automotive component assembly to automated cabling for data centers—essentially any process requiring highly precise joining operations."

The company is also accelerating its global expansion. SKAI Intelligence currently operates wholly owned subsidiaries in Singapore, Shanghai, and Paris, and plans to establish a U.S. subsidiary as early as next year to strengthen its presence in one of the world's leading manufacturing markets. At the same time, the company has begun preparations for an initial public offering (IPO), working toward meeting listing requirements.

Lee's long-term vision is to establish the global standard for manufacturing automation data. He believes that in a market where standards have yet to emerge, defining those standards will allow the company to lead the industry's future direction.

"Our short-term goal is to become the world's No. 1 synthetic data infrastructure company for industrial manufacturing automation," Lee said. "Over the longer term, we aim to collaborate with global industry leaders to establish international standards for manufacturing data."