FastLabel Selected for NEDO National Project “GENIAC”

FastLabel announced that it had been selected for “Research and Development of Robot Foundation Models (GENIAC),” a national project run by the Ministry of Economy, Trade and Industry and NEDO. FastLabel made the announcement and is the selected company. In addition to confirming the selection, the statement outlined plans to develop locomotion and manipulation technology for quadruped robots and a VLA model to improve patrol and inspection work in the plant industry.

FastLabel’s selection for a NEDO national project shows that its business is entering the field of robot foundation model research and development, beyond data support. According to the announcement, the project is “Research and Development of Robot Foundation Models (GENIAC),” making the company’s role in data preparation and model development a point of interest.

The new development in this announcement is FastLabel’s entry into a national project framework run by the Ministry of Economy, Trade and Industry and NEDO. Alongside the selection, the statement outlined the development of locomotion and manipulation technology for quadruped robots and a VLA model, as well as field testing at a Kyushu Electric Power thermal power plant. The news therefore centers less on what the company will build and more on FastLabel being selected for a public project focused on “robot foundation models.” In an earlier article, FastLabel launches “Data Factory Operations,” a service supporting the launch and operation of data production factories, the company highlighted support for launching and operating data production. The GENIAC selection can be seen as an expansion from data preparation and operational support into robot foundation model research and development. The previous article focused on service delivery, while this announcement concerns participation in a national project, shifting the company’s positioning from “support services” to a role in a public research program. From an industry perspective, connecting real-world robot data, training data preparation and model development is important for robot foundation models. When a data-related company such as FastLabel joins a national project, a key question is what role it will play in collecting, preparing and managing the quality of training data, beyond model development itself. The announcement names Kyushu Electric Power as a research partner and says the companies will conduct field testing at one of its thermal power plants to automate and improve patrol and inspection work.

The published press release says FastLabel was selected for “Research and Development of Robot Foundation Models (GENIAC).” GENIAC is a project run by the Ministry of Economy, Trade and Industry and NEDO. FastLabel says it will handle data design, training data preparation, model development and evaluation, and advance technology development in collaboration with Kyushu Electric Power.

In Japan, companies responsible for preparing and managing training data are expected to play an important role in the development of robot foundation models, alongside developers of the models themselves. The involvement of companies such as FastLabel, which promotes data production support, in national projects is likely to bring greater attention to the role of domestic players in data preparation.