Every Company Faces Unique AI Challenges—Custom Solutions Are the Key
The number of companies looking to adopt artificial intelligence (AI) services is rapidly increasing.
According to the Korea Chamber of Commerce and Industry, 78.4% of domestic companies believe that AI adoption is necessary. However, only 30.6% have actually implemented AI in their operations—less than half. In the manufacturing industry, the adoption rate is just 23.8%. The main barriers are a lack of technology (34.6%) and cost burdens (23.1%). As a result, more AI solution providers are emerging to help companies easily build the AI services they need.
Superb AI is one such company. Founded in 2018, it has secured major domestic clients and overseas customers such as Toyota and Nippon Steel (Japan Steel Works) and has recently expanded into the video surveillance solution sector. Client acquisitions have also driven investment. In September last year, the company secured a Series C investment of 13.5 billion won, bringing its cumulative investment to 49 billion won. Superb AI plans to leverage this growth momentum to pursue an initial public offering (IPO) by 2026.
Five Co-Founders Led by CEO Hyun-soo Kim
Superb AI is led by CEO Hyun-soo Kim, born in 1990, who has followed an elite academic and professional path. He graduated at the top of his class from Duke University with degrees in electrical engineering and biotechnology, and later pursued a doctorate in AI and robotics. In 2016, he joined SK Telecom after being recruited, working on AI speaker development and reducing errors in Tmap usage time at the SK Telecom Research Institute.
While conducting AI research, Kim became frustrated that too much time was being spent on data preparation rather than research itself, which led him to start his own company. “One of the most necessary technologies I recognized while working on AI was data labeling automation. I wanted to try it myself,” he said.
In 2018, Kim founded Superb AI with Chief Technology Officer Cha Moon-soo, Vice President Lee Hyun-dong, 3D Vision Team Lead Lee Jung-kwon, and Video Analysis TF Lead Lee Jong-hyuk. The company entered the AI market with image data labeling automation.
Data labeling—often referred to as the “eye-popping” task in the AI industry—can be thought of as tutoring for AI. It involves creating data that AI can learn from. For example, a dog photo might be labeled with “This is a dog” for AI training. Traditionally, humans performed this work manually. Superb AI changed this model so that AI assists in labeling, with humans only reviewing parts where AI cannot provide accurate results.
The approach quickly produced results. Soon after its founding, Superb AI began receiving client inquiries, and there was M&A interest in its data labeling technology. “We have five co-founders, and we agreed to accept an acquisition offer only with unanimous consent. All five decided not to sell,” Kim said. “I still feel the same way. The AI market is huge, and we have only taken a small step forward. This is just the beginning.”
Expanding with AI All-in-One Services
Building on its data labeling success, Superb AI has expanded its services. In addition to its core labeling automation model, it launched the all-in-one Superb Platform, covering the full AI development cycle—from selecting training data and automating data processing to AI model training, diagnosis, application, and deployment.
“We aimed for an end-to-end model with the Superb Platform,” Kim explained. “Most companies seeking to apply AI tend to fragment their contracts—one for data design, another for algorithm development. We saw an opportunity in providing all of this through a single platform.” While customers can still purchase individual services, Kim noted that “expanding the service scope has significantly increased demand.”
Superb AI has secured over 100 corporate customers, including Hyundai Motor Company, Kakao, and LG Electronics. It has also launched a video surveillance solution for the manufacturing, construction, and logistics industries. This solution uses AI to analyze multiple video streams with technologies such as object detection and deep learning-based image classification to detect smoke, vehicles, faces, falls, and abnormal behavior. It can also track specific objects over time, even as conditions change.
“Existing video surveillance solutions often cannot accurately detect the objects customers want, resulting in low satisfaction,” Kim said. “Our solution overcomes these limitations, greatly improving safety and efficiency at industrial sites.”
Overseas Expansion and Global Clients
Superb AI has established subsidiaries in the U.S. and Japan and provides services to major companies in both markets. In the U.S., clients include Qualcomm and Amazon Web Services (AWS); in Japan, the company works with Toyota and others. In 2023, it partnered with global IT distributor TD Synnex to expand sales. The company now provides products and services to more than 150,000 customers in over 100 countries.
“Both the U.S. and Japan are growing at about twice the annual rate in sales and orders,” Kim said, noting that sales volumes are similar in both countries.
2026 IPO Target and Technology Listing
Superb AI aims for an IPO in 2026, with Samsung Securities as the lead underwriter. Most of the Series C funding was invested in securing talent to meet growing demand and ensure a successful IPO. Some have raised concerns that the IPO timeline is aggressive, but Kim explained, “If we raise additional external funds such as Series D, the number of PEF investors will inevitably increase. We decided it would be better to focus on a technology-based listing rather than take that path.”
Looking ahead, Kim is considering market strategy for the on-premises version of the Superb Platform, launched last year alongside its cloud-based SaaS offering. “We are deciding whether to target a few industries most suited for this version and focus marketing there,” he said.
AI solutions are generally delivered as either on-premises or cloud-based products. While cloud solutions dominate due to ease of deployment, South Korea’s regulatory environment—particularly in sectors such as defense, where foreign software cannot be used—creates strong demand for on-premises AI systems installed directly on a company’s internal servers.
댓글
댓글 쓰기