Selvas AI (CEO Kim Jae-hyun) has expanded the application of its voice recognition and speech synthesis technologies to Incheon International Airport guide robots, following their deployment in healthcare and financial kiosks. As of 2:53 p.m. on the 3rd, Selvas AI's stock price surged by 19.18% from the previous day to 9,630 won.

■ 'Selvas Trigger' Installed in About 20 Incheon Airport 'Airstar' Robots

According to the Electronic Times, Selvas AI partnered with robotics specialist company Clobot to install its AI voice interface technology, 'Selvas Trigger,' in approximately 20 'Airstar' guide robots at Incheon International Airport. Airstar combines autonomous driving with voice recognition to guide passengers to boarding gates and key airport facilities, and provides navigation to those locations when necessary.

According to official Selvas AI materials, the applied technology is an interface that initiates voice conversations by recognizing the user's wake word. The company explained that the relevant technology was provided to distinguish the user's voice and execute guidance functions even in the noisy airport environment.

However, the technology deployment does not directly translate into large-scale revenue. Currently, the officially confirmed fact is that the voice interface has been applied to about 20 Airstar robots, with contract amounts and revenue contributions undisclosed. The significance of this deployment lies in Selvas AI expanding its scope from its existing software-centric business to the 'Physical AI' domain, which involves controlling robots and equipment via voice.

■ Healthcare AI Business Also Expanding... Collaborations with InterMD, Mediana, and EZCareTech

The application scope of voice AI is expanding beyond airports into healthcare settings. According to Selvas AI's official materials, the company partnered in June with InterMD, a platform for medical professionals with approximately 50,000 doctor members, to expand the service channels for its medical voice recognition solution, 'Selvas Medivoice.'

Selvas Medivoice is a solution that converts spoken medical findings into text to support the creation of medical records. Unlike general voice recognition, it is a product focused on processing specialized and medical terminology specific to different departments.

Additionally, Selvas AI is jointly developing an AI-based smart ward integrated solution with Mediana and EZCareTech. According to the company's announcement, the three companies plan to develop a system that links biosignals collected from patient monitors with electronic medical records in real time, and uses AI to analyze patient conditions to provide early warnings. The automation of medical records using voice recognition and pilot hospital demonstrations are also included in the scope of collaboration.

This suggests that Selvas AI's healthcare business is evolving beyond simple voice-to-text conversion into an integrated structure that connects patient monitors, electronic medical records, and early warning systems. However, the smart ward agreement is currently at the joint development and commercialization promotion stage; actual product completion, hospital supply, and revenue generation must be confirmed through future demonstration results and contracts.

■ Mediana Accounts for a Large Share of Consolidated Performance... Limitations in Interpreting Voice AI Alone

According to a report by Hankyung (Korea Economic Daily) citing the company's announcement, Selvas AI recorded consolidated revenue of 114.8 billion won and an operating profit of 1.28 billion won in 2025. Revenue increased by 2% year-on-year, while operating profit surged by 60%. Hankyung reported that the expansion of overseas sales by subsidiary Mediana and improvements in the cost structure drove the performance improvement. North American sales of patient monitors increased by over 60% year-on-year, and European sales of automated external defibrillators grew by more than 65%.

Therefore, it is difficult to interpret Selvas AI's corporate performance solely based on voice recognition software revenue. The consolidated financial statements include not only Selvas AI's AI solutions but also Mediana's medical device businesses, such as patient monitors and automated external defibrillators.

■ Holds Over 99 Fundamental Patents... Reflected Based on Latest Disclosure

Selvas AI supplies voice recognition, speech synthesis, handwriting recognition, and optical character recognition technologies to enterprises and public institutions. Its official website introduces key products such as the voice recording solution 'Selvas Note,' the medical voice recognition solution 'Selvas Medivoice,' and the disease onset probability prediction solution 'Selvy Check-up.'

According to the quarterly report submitted to the Korea Exchange for the first quarter of 2026, the business divisions are categorized into the HCI division, which supports human-computer interaction, and the AI Digital Transformation division, which supports AI transition across industries. The quarterly report includes the company's statement that Selvas AI has secured a recognition rate of over 98% in healthcare-specific voice recognition and holds more than 99 fundamental technology patents. This figure is higher than the 74 patents displayed on some corporate information websites, and it is appropriate to use the latest disclosure standards when writing articles.

■ Challenge: Proving a 'Recurring Revenue Structure' Rather Than the Number of Use Cases

What Selvas AI must prove going forward is not the number of use cases, but a recurring revenue structure. The deployment in Incheon Airport robots is an example of voice technology being used in a Physical AI environment, but the contract scale and maintenance revenues have not been disclosed. For healthcare AI as well, the trends in the number of hospitals using the technology, paid users, and contract renewals must be monitored following deployment in hospitals and platforms.

Voice recognition technology can be utilized as an input method for various devices, including robots, kiosks, automobiles, and medical equipment. However, for use cases to lead to improved performance, a recurring revenue stream must be established through software licensing fees, maintenance, and additional equipment supply, going beyond individual project deliveries.

Future evaluation criteria are expected to focus less on short-term stock price movements driven by generative AI themes and more on the following: whether additional robots will be supplied following Airstar, whether the smart ward collaboration will translate into actual hospital contracts, and whether AI software revenue will grow alongside medical device revenue.

[This article was written with AI assistance. This article is a reference material for investment decisions and does not recommend the purchase or sale of any specific stock. Investors should verify company disclosures and exchange data before investing.]