Japan plans AI-powered satellites for faster military counterstrike decisions

Japan’s Ministry of Defense has a problem that most militaries would recognize: modern threats move faster than traditional decision-making chains. Hypersonic missiles, maneuvering warships, and time-sensitive targets don’t wait for analysts to sift through satellite feeds. Tokyo’s proposed fix is to move the analysis into orbit.

On September 4, 2026, the Ministry of Defense unveiled plans to develop domestically produced tactical AI satellites designed to compress the gap between detecting a target and authorizing a long-range missile strike. The satellites would process data from multiple surveillance platforms in near real-time, cutting out the lag that currently sits between observation and decision.

How the program works

The tactical AI layer sits inside the satellite itself, running analysis before the data ever touches the ground. The result is a system that can track moving targets and feed guidance data to long-range missiles with dramatically reduced latency.

A prototype development contract was awarded to NTT Data on January 8, 2026. Demonstrator satellites are scheduled for construction across fiscal years 2027 through 2029, with ground testing and the first orbital validation runs targeted for fiscal year 2030. The entire AI stack will be Japanese-made, a deliberate choice that reflects both national security concerns about foreign technology dependencies and Tokyo’s broader push to build a domestic defense technology base.

Japan’s space defense spending tells its own story. Investment in that category climbed from ¥79 billion in fiscal year 2022 to ¥540 billion in fiscal year 2025.

The strategic context behind the hardware

Japan’s 2022 National Security Strategy revision formally endorsed counterstrike capability, meaning Japan would no longer rely entirely on the United States to hit back at an aggressor. Without fast, reliable targeting data, a counterstrike capability is mostly theoretical. Missiles need to know where to go, and moving targets, ships, mobile launchers, don’t sit still. An AI system in orbit that can track, classify, and hand off targeting coordinates in near real-time is what bridges the gap between policy intent and operational reality.

The program fits inside a record defense budget request of ¥8.89 trillion for fiscal year 2027, roughly $55.6 billion. That figure prioritizes AI development, unmanned systems, and expanded strike capacity alongside the satellite initiative.

Japan’s MOD has been building out a broader constellation architecture, working with Mitsubishi Electric and SKY Perfect JSAT on surveillance and strike-support satellite networks. Mitsubishi Heavy Industries previously demonstrated an in-orbit AI system called AIRIS, designed for ship detection, which served as an early proof-of-concept for the kind of orbital machine learning now being scaled into this program.

What to watch as the program develops

NTT Data’s role as prototype contractor is notable. The company is primarily known as an IT services and data infrastructure firm rather than a traditional defense prime, which suggests Japan is deliberately pulling commercial technology expertise into its defense industrial base rather than leaning entirely on established aerospace contractors.

The harder question the program raises is doctrinal rather than technical. Faster targeting decisions supported by AI recommendations compress the time available for human oversight at exactly the moment when the stakes are highest. Japan’s MOD has not detailed publicly how human authorization fits into an AI-accelerated counterstrike loop, and that design choice will matter considerably more than the orbital insertion date.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.