NIKKEI THE PITCH会議室 第25回 政府が動いた、民間が動いた——知っておきたい 防衛×スタートアップ

Why Defense × Startups, and Why Now

Ukraine changed the fundamentals of defense. The old model—major heavy-industry conglomerates spending years developing tanks, fighter jets, and missiles—cannot keep up with a conflict where drones dominate and combatants iterate on weaknesses within weeks. The Showa-era industrial structure, built for stability, scale, and long development cycles, simply lacks the speed and cost discipline that modern warfare now demands.

Drones are rooted in consumer technology to begin with. Many of the drones sold as children's toys for ¥5,000–6,000 are made by Chinese manufacturers, and the concept of "dual use"—repurposing civilian technology for defense—has spread rapidly. The same logic applies to AI and robotics: motion-sensing technology used in factory automation can be adapted into targeting systems. The recognition that failing to cultivate domestic technology creates a national-security vulnerability has sharpened in both government and industry.

Government Moves — Three Pillars

**A new investment vehicle**: The Defense Ministry and METI are planning a dedicated joint organization, to be established in fiscal 2027, that will invest in startups via domestic VCs. The government's role is to act as the first major customer—priming the pump in a sector that private capital has historically avoided. The associated unmanned-vehicle budget has been set at ¥277.3 billion, up 2.4x year-on-year.

**The early intercept-drone acquisition program**: This initiative radically compresses the traditionally slow defense-procurement timeline. Moving from open solicitation to contract in roughly three months—the pace startups take for granted—is now being applied to defense acquisition.

**Export regulations reformed**: The April 2026 regulatory revision lifted restrictions on exporting completed defense equipment. Lethal weapons still require government review and face limits on eligible buyer countries, but the framework for taking Japanese defense products to overseas markets is taking shape—part of a broader effort to build a government-to-government sales mechanism analogous to the U.S. Foreign Military Sales (FMS) system.

Private Capital Moves — Banks, VCs, and Corporates

Following the government's policy shift, private money has begun to move. Japan Development Bank changed its lending policy toward weapons-related companies from "proceed with caution" to "open to consideration." MUFG Bank announced plans to expand lending to defense-related firms. The shift in European attitudes toward the defense industry triggered by Russia's invasion of Ukraine has now reached Japan.

On the VC side, Crosstech Ventures announced that its Fund III will invest seriously in defense startups. A managing partner at Globis Capital Partners stated plainly that government procurement creates investment opportunities that otherwise would not exist. True to the market maxim "don't fight policy," the government's direction is translating directly into investment decisions.

Open innovation with large corporates is advancing as well. A major domestic heavy-industry manufacturer has taken a stake in Preferred Networks to co-develop AI for defense equipment. Rakuten has signaled a tie-up with German defense-AI startup Helsing. Globally, VC investment in defense startups reached $40.8 billion (approximately ¥6.4 trillion) in the first half of 2026—up 56% year-on-year according to PitchBook data—and Japan is at the entry stage of that wave.

Startups to Watch — and the Hurdles That Remain

**Terra Drone** saw its domestically developed intercept drone, the Terra B1, pass the early-acquisition program's verification trial as a sole qualifier, and signed a mass-production procurement contract with the Acquisition, Technology & Logistics Agency in August 2026, with delivery scheduled for September. The company is also moving to acquire a Ukrainian intercept-drone firm as a subsidiary, aiming to absorb technology refined in live combat.

**Sakana AI** raised more than ¥31 billion in the first half of this year to accelerate its defense capabilities. It was commissioned by an ATLA research institute to analyze data collected from drones and other assets to support commanders' decisions, and in July 2026 it signed a research and demonstration contract with the Ministry of Defense on AI utilization—a model close to what Palantir does in the United States for AI-assisted operational decision-making.

Serious challenges remain, however. Payment on contracts can come two years after signing, and a cash deposit equal to 10% of the contract value must be posted upfront—a direct hit to the finances of capital-light startups. The government has begun introducing remedies such as advance-payment recognition, but the legacy procurement framework, designed around large weapons systems, has not changed at its roots. Mass-production capacity is an equally large obstacle: the government's target of 80,000 domestically produced unmanned vehicles per year by 2030 is ambitious, and the factories and companies needed to hit that number are still to be built. And just as the internet itself began as military technology before migrating to civilian life, the broader question of how society governs the way defense-derived technology is used cannot be avoided.