Daily Japan updates

Japan's AI market pushes from documents toward hands-on work

Two company announcements place AI at opposite ends of implementation: a ¥5 billion physical-AI bet and a retrieval-based assistant for teachers.

Developments

Japan’s latest AI announcements point in two directions at once. One company is funding systems that act in the physical world; another is narrowing a chatbot around a defined body of professional material.

1. Algomatic Dynamics directs ¥5 billion toward physical AI

Algomatic Dynamics says it has raised ¥5 billion from DMM.com to expand robotics and physical-AI research, including hardware and computing resources. Its announced model stretches from collecting and structuring motion data to training, deploying, tuning, and maintaining robotic hands. The company aims to release an AI multi-finger hand platform in Japan during 2026, although the timetable and technical performance remain company claims rather than independently verified results. (Read the announcement)

2. A publisher grounds a teacher assistant in curriculum materials

SELF says Nihon Bunkyo Shuppan has adopted SELFBOT for a free chat service supporting teachers of moral education. The system is described as using retrieval-augmented generation to consult curriculum commentary and the publisher’s teaching guides when answering questions about lesson goals, classroom prompts, student discussion, and reflection. The deployment illustrates a narrower business case for generative AI while leaving claims about accuracy and educational value to future evaluation. (Read the announcement)

What the two developments suggest

The practical divide is no longer simply between companies that use AI and those that do not. The more useful question is where the system is anchored: in costly hardware and real-world motion, or in bounded professional documents and workflows.