[Interview] Flitto Inc. CEO Simon Lee: “On-Prem AI Market to Expand as Security Tightens ... Double Growth Next Year”
- Input
- 2026-11-16 15:44:37
- Updated
- 2026-11-16 15:44:37

[Financial News] As companies increasingly adopt artificial intelligence (AI), data security and the ability to optimize AI models are emerging as new sources of competitiveness. Demand is growing, especially in industries that handle sensitive information and critical technologies, such as semiconductors, defense and finance, for “on-premises AI”—AI operated within a company without routing data through external servers. This means that, alongside AI model performance, technologies for refining data to suit each company’s working environment while maintaining security are becoming increasingly important.
In an interview with Financial News at Flitto Inc.’s headquarters in Gangnam-gu, Seoul, on the 16th, Flitto Inc. CEO Simon Lee said, “Companies for which security is important feel uneasy about using cloud-based software as a service (SaaS).” He added, “As language-related data is also directly tied to security, demand is growing for building AI systems on-premises.”
On-premises AI is a way for companies to operate AI on their own servers and internal networks. Because key corporate data is not sent to external clouds, it can reduce the risk of information leaks and increase control over data. It is particularly useful in advanced industries that handle critical technologies and confidential information, as AI can be built to meet each company’s security requirements.
Flitto Inc. recently signed a contract worth about 3.9 billion won with a major global semiconductor company to supply an on-premises AI interpreting and translation solution. The contract runs for three months and involves building an AI system on the client’s internal network. The system provides speech recognition, summarization, interpreting and translation for internal messaging platforms and online meetings.
Lee emphasized that the contract goes beyond simply supplying interpreting and translation services: it is a project to build an AI system inside a company. Since early this year, Flitto Inc. has been stepping up its efforts to expand AI model lightening and on-premises deployment for companies with significant security needs.
Flitto Inc.’s key competitive strength is its data-refinement technology. In industries such as semiconductors, where specialized terminology is abundant, it is difficult to obtain accurate results using only general-purpose AI models. Drawing on its expertise in processing language data, Flitto Inc. trains models on industry-specific terminology and optimizes them for each client’s computing environment.
Lee stressed, “To implement on-premises AI, a company needs its own AI model, the technology to train that model, and data-refinement capabilities. Very few companies have all of these capabilities.”
Flitto Inc.’s data-refinement capabilities are a particularly important competitive strength in on-premises AI deployment. That is because a client’s existing materials must be processed into a form suitable for AI training, while industry-specific terminology and work practices must also be incorporated. Flitto Inc. is advancing its technology by linking data refinement, model lightening and deployment on companies’ internal networks.
The ability to tailor deployments to each client is another strength. During clients’ technical evaluations, Flitto Inc. uses its own graphics processing unit (GPU) servers to build AI systems directly and verify their performance. During actual deployment, the personnel in charge stay near the client for at least 3–6 months to build the system.
On-premises AI is becoming a new growth driver for Flitto Inc. Lee forecast that the scale of the company’s related business—including on-premises AI revenue and the number of clients—would grow roughly twofold next year compared with this year. To that end, it is making the United States, the Middle East and Japan key hubs for its on-premises AI business. In Japan, interest in adopting on-premises AI is high, particularly in security-sensitive industries such as finance. The United States has significant growth potential because its market is large and project prices are high. In the Middle East, government-led AI projects are active, and many clients require AI to run on local servers rather than external clouds. Lee said, “Next year, we will accelerate the growth of our US business while increasing our operating margin.”
[email protected] Jang Min-kwon Reporter