AIST and ZenmuTech are pursuing real-world implementation centered on “secure computation,” which enables confidential data to be analyzed without disclosing it externally. The report follows the collaboration from its inception through the development and commercialization of practical tools that enable data sharing.
“Secure computation” is attracting attention as a technology that enables data to be used while protecting confidential information, and the National Institute of Advanced Industrial Science and Technology (AIST) and ZenmuTech are pursuing a collaborative initiative aimed at real-world implementation. The approach is based on dividing data into multiple fragments, processing them on separate computers, and extracting only the final result.
The collaboration dates back to 2018, when ZenmuTech’s Representative Director, COO and CTO Kunii Shimpei (國井 晋平) visited Hanaoka Goichiro (花岡 悟一郎) and built a relationship of trust. At an early stage, the parties recognized the need for third-party proof of security and decided to collaborate after obtaining verification from AIST.
Technology development progressed toward practical use in a two-computer configuration, and in 2021 produced results that accelerated the shuffling process for encrypted data by more than 100 times. This prompted the development of a tool called QueryAhead to accelerate the identification of on-site needs and market applications.
AIST and ZenmuTech continued studying ways to make data sharing between companies a reality, focusing on applications in manufacturing supply chains, and built a system that enables analysis while keeping data confidential. The formation of a consortium of around 40 companies is also under way, and implementation and demonstrations are being accelerated.
In 2023, AIST established AISol and promoted joint research, while ZenmuTech received an investment from AISol and listed on the TSE Growth Market in 2025. A framework for expanding the use of secure computation across society is gradually taking shape.
The path toward real-world implementation of data use while keeping the data hidden involves many challenges, but as data becomes increasingly important in the AI era, practical applications of secure computation are advancing through collaboration between companies.