it_chosunhace 15d
Company Name : CUBIGCEO : Ho Bae, Minchan ChungYear of Establishment : 2021Investment Stage : Series AKey Product : DTS, X-PERT, LLM Capsule, AzooIndustry : AI+Healthcare, AI+Retail/CommerceTechnology : AI Model·Architecture / Large Language Model (LLM), Vision-Language Model (VLM), etc. AI Development·Operation Technology / MLOps, Data Engineering, etc.Website : cubig.aiCubig's advanced technology opens new possibilities for safe and scalable data use, achieving 99% performance similarity without accessing original data. X-PERT. / CUBIGSolving Data Challenges Without Accessing Original DataFounded in 2021, Cubig is a startup developing next-generation data solutions that allow high-performance AI training without personal information, using synthetic data technology. By addressing data challenges without accessing original datasets, Cubig provides a safe and effective AI training environment for high-risk industries such as finance, healthcare, public sector, and defense, drawing attention both in Korea and abroad.Cubig's core technology consists of three main solutions. "Data Transform System (DTS)" is a differential privacy (DP)-based synthetic data generation platform that creates multi-modal data, including tables, text, images, and time-series, without accessing original datasets. This technology creates data without personal information based on differential privacy (DP) to safely supply datasets needed for AI training and analysis.It is suitable for high-risk sectors such as finance, healthcare, public services, and defense, offering an integrated platform that covers data generation, validation, and analysis.X-PERT is a synthetic data-based AI agent collaboration platform that allows enterprises to safely clean and use sensitive or confidential data across departments without real data. X-PERT supports data merging, analysis, and sharing, and its no-code interface makes it accessible to non-developers. Finally, LLM Capsule is a generative AI security gateway that detects and anonymizes personal and confidential information in real time when using public LLMs. Operating locally, it prevents external leaks and can be used in air-gapped environments, enabling safe AI deployment in public, financial, and healthcare sectors.Establishing Itself as a Core Solution for Data InnovationCubig has demonstrated rapid growth. The company holds eight domestic patents (one granted) and has filed two international patents, alongside numerous publications and AI research projects that validate its technology. The company collaborates with a variety of organizations, including SKT, Hana Bank, Woori Bank, Naver Cloud, the military, and university hospitals. Some of these partnerships have completed proof-of-concept (PoC) stages and are moving into contractual agreements.From a business perspective, Cubig operates both SaaS (subscription-based) and on-premise (deployment-based) solutions, establishing diversified revenue streams through AI analytics solutions and a synthetic data distribution platform. Its data marketplace, Azoo, allows anyone to trade synthetic data, supporting safe synthesis, purchase, and sale of datasets, and ensuring scalability and sustainability for data-driven businesses.Azoo is a marketplace for trading and distributing synthetic data, allowing companies to safely synthesize their own data or buy and sell datasets, helping them quickly access the data they need.Globally, Cubig is targeting European and North American markets to become the world's top private synthetic data platform provider by promoting overseas sales using Google SEO, securing global data buyers, and expanding B2B partnerships.In this era where AI is driven by data, Cubig delivers "high-performance AI analytics without accessing original data," setting new standards for AI ethics and personal data protection. Beyond serving as a mere alternative, its synthetic data technology has become a core solution driving the next wave of data innovation.jyhong@chosunbiz.comThis article was written in 2025 and reflects information available at the time of writing.