🛩️ The race to build AI-powered "wingman" drones is reshaping military aviation. The US has already flown two CCA prototypes. China unveiled four new stealth drone designs at a single parade. Japan plans its first AI drone test flight in late 2025 — but how far behind has it fallen, and what structural challenges does its defense industry face?
What Is a "Loyal Wingman"? The Future of Air Combat
One of the most closely watched concepts in modern military aviation is the "loyal wingman" — an AI-powered unmanned aircraft that flies in formation with manned fighters, autonomously performing reconnaissance, electronic warfare, and attack missions. The US calls them Collaborative Combat Aircraft (CCA), while Japan uses the term "combat support unmanned aircraft."
Why does this technology matter so much? First, drones can be sent into dangerous missions without risking pilots' lives. Second, they cost far less than manned fighters, enabling a "mass over exquisite" strategy. Third, AI-driven autonomous decision-making can coordinate multiple aircraft simultaneously while reducing pilot workload.
The Ukraine conflict demonstrated that inexpensive drones can destroy tanks and warships costing millions. That lesson has sent shockwaves through military doctrines worldwide, accelerating investment in manned-unmanned teaming technology.
United States: Over 1,000 CCAs to Escort 6th-Gen Fighters
The US leads the unmanned combat aircraft race by a significant margin. Under its Next Generation Air Dominance (NGAD) program, the US Air Force is developing CCA drones to operate alongside Boeing's sixth-generation F-47 fighter.
In 2024, the USAF awarded development contracts to five companies — Anduril, Boeing, General Atomics, Lockheed Martin, and Northrop Grumman — then selected Anduril and General Atomics for Increment 1 production. General Atomics' YFQ-42A "Gambit 2" completed its maiden flight in August 2025, followed by Anduril's YFQ-44A "Fury" in October 2025.
The scale is staggering. Increment 1 envisions 100–150 aircraft, while Increment 2 could add up to 2,350 more. The USAF has stated plans to field at least 1,000 CCAs. The FY2026 budget allocates over $4 billion for the F-47 and CCA programs combined, with approximately $800 million for CCA alone. Initial operational capability is targeted before 2030.
In December 2025, the Air Force selected nine vendors for Increment 2, expanding the program to cover air-to-air combat, ground attack, electronic warfare, and ISR missions with multiple CCA types.
China: Four New Stealth Drones Unveiled in a Single Day
China is widely regarded as neck-and-neck with, or in some areas ahead of, the United States.
The Zhuhai Airshow in November 2024 showcased the Feihong-97A, a loyal wingman drone designed to fly alongside manned fighters. Even more dramatic was China's September 2025 military parade, where four entirely new stealth drone designs debuted alongside the established GJ-11 attack drone.
Two of the new designs are smaller CCA-class aircraft, one resembling the US XQ-58. The other two are significantly larger — roughly the size of a J-10 fighter — powered by what analysts believe are derivatives of the WS-10 or WS-15 turbofan engines. This suggests China is building a tiered unmanned force: affordable mass-production drones alongside high-performance platforms.
China has also developed the J-20S, the world's first twin-seat fifth-generation stealth fighter, specifically designed so the rear-seat pilot can manage a swarm of companion drones. In November 2025, the PLA Air Force released footage of a J-20 flying in formation with GJ-11 drones and J-16D electronic warfare aircraft.
Europe, Australia, and South Korea Join the Race
The competition extends far beyond the US and China.
The United Kingdom introduced its "StormShroud" loyal wingman demonstrator in May 2025, with an initial order for 24 platforms. Developed as an Autonomous Collaborative Platform (ACP) separate from the GCAP fighter program, StormShroud is designed to team with existing 4th and 5th-generation fighters as well as future aircraft.
Australia has been a CCA pioneer through its MQ-28 Ghost Bat program with Boeing. The Ghost Bat features a modular nose section that can be swapped to reconfigure the drone for reconnaissance, electronic warfare, or attack missions. An AIM-120 missile firing test was planned for late 2025 or early 2026.
South Korea is developing multiple loyal wingman types in parallel with its KF-21 fighter program, with Korean Air having unveiled a UCAV model for the Korean Air Force.
Japan's Current Position: The Road to GCAP-Linked Drones
Japan's unmanned combat aircraft development centers on the GCAP (Global Combat Air Programme) next-generation fighter being jointly developed with the UK and Italy.
Mitsubishi Heavy Industries (MHI) unveiled two combat support drone concepts at the 2024 International Aerospace Exhibition in Tokyo — a medium-sized CCA airframe and a missile-shaped expendable drone designated "ARMDC-20X." MHI has been working on AI-equipped drone technology since 2022 under contracts from Japan's Acquisition, Technology and Logistics Agency (ATLA).
Kawasaki Heavy Industries (KHI) debuted its own AI-powered combat support drone concept at DSEI Japan 2025, emphasizing electronic warfare capability and power generation from its proprietary KJ300 compact engine. KHI envisions its drones supporting not only the GCAP fighter but also the P-1 maritime patrol aircraft.
SUBARU's aerospace division is also involved, having manufactured test drones displayed at ATLA's Technology Symposium 2025. Flight testing in Hokkaido successfully demonstrated five-drone formation flying using GPS navigation with voice commands from a helicopter-mounted control system.
ATLA planned the first flight of an AI-equipped unmanned demonstrator for November 2025. In December 2025, it also issued a tender for a study on "risks and challenges in introducing drones that cooperate with the next-generation fighter" — directly addressing how much autonomy AI systems should be granted in combat situations.
The FY2026 defense budget includes ¥4.8 billion ($32 million) for conceptual design of GCAP-linked drones, and ¥100.1 billion ($670 million) across all unmanned asset programs, with a target of establishing a "multi-layered coastal defense system using unmanned assets" by FY2027.
Structural Challenges Facing Japan's Defense Industry
Despite steady progress, Japan's defense industry faces several deep-rooted challenges compared to its US and Chinese competitors.
The budget gap is enormous. The US plans to spend $8.9 billion on CCA programs from FY2025–2029. Japan's linked-drone budget stands at ¥4.8 billion ($32 million) — less than one-hundredth of the US figure. MHI's entire aerospace, defense, and space division generates roughly ¥1 trillion in annual revenue, less than one-tenth of Lockheed Martin's $71 billion.
No combat data for AI training. AI performance depends on the quality and quantity of training data. Japan has no recent combat experience, making it reliant on simulations. The US, by contrast, has already tested AI-flown fighter aircraft (the X-62A, a modified F-16) in simulated dogfights against human pilots.
A narrow industrial base. Japan's defense companies have historically depended entirely on domestic demand due to strict export controls. While restrictions have been gradually relaxed — the 2014 defense equipment transfer principles and 2024 authorization of third-country exports — overseas sales remain limited. Few new entrants are joining the defense sector, constraining innovation.
Ethical and legal uncertainty over AI autonomy. How much autonomous decision-making should combat AI be allowed? ATLA's December 2025 tender explicitly seeks answers to this question. The debate over "how much should humans remain in the loop" is not merely technical — it is a policy and philosophical challenge that will shape procurement decisions.
Japan's Strengths and Future Outlook
Japan does possess genuine advantages, however. International collaboration through GCAP with the UK and Italy offers both technology sharing and cost distribution. The competitive dynamic between MHI and KHI produces diverse concepts. Japan's dual-use technology ecosystem — AI, sensors, semiconductors — creates opportunities for startups and cross-industry entrants.
Japan is also pragmatically evaluating foreign unmanned systems, including Turkey's Bayraktar TB2S and Israel's Heron Mk II, for potential adoption alongside domestic development. This realistic approach of combining proven foreign technology with indigenous R&D could help narrow the gap.
The unmanned combat aircraft race is no longer a distant future scenario. The US and China are in flight-test stages, targeting operational deployment before the end of this decade. Whether Japan can establish its presence depends on sustained budget commitment, AI technology investment, industrial base reinforcement, and deeper international cooperation.
In Japan, debate swirls between those who see unmanned combat aircraft as essential for national security and those who question whether AI should ever be trusted with lethal decisions. How is your country approaching the development of unmanned combat aircraft and the military use of AI? We'd love to hear your perspective.
References
- https://trafficnews.jp/post/617947
- https://j-defense.ikaros.jp/docs/mod/004199.html
- https://j-defense.ikaros.jp/docs/commentary/004463.html
- https://newswitch.jp/p/39402
- https://instituteofgeoeconomics.org/research/2025052801/
- https://www.airforce-technology.com/projects/collaborative-combat-aircraft-cca-usa/
- https://euro-sd.com/2025/11/articles/technology/47629/ccas-rcs-loyal-wingmen-and-effectors-developing-unmanned-systems-for-the-future-air-superiority-team/
Reactions in Japan
Saw the actual drones at ATLA's symposium. Smaller than I expected, but hearing they flew a 5-drone formation got me hyped. Japan can do this after all.
CCA budget: US over ¥800 billion, Japan ¥4.8 billion... Two orders of magnitude difference. Are they seriously trying to catch up?
Good to see MHI and KHI competing with different concepts. My worry is that in Japan, instead of competition producing a winner, they'll just merge things in a compromised way.
Former JASDF here. Pilot shortage is real, so loyal wingmen are welcome. But how do you guarantee AI reliability for autonomous decisions under jamming with zero combat data? That worries me.
Seeing China unveil 4 new stealth drone types at their parade honestly gives me chills. Japan can't afford to leisurely say 'we're at the concept stage.'
The debate about whether AI should make attack decisions is important. But while we debate, China deploys. How do you balance ethics and security?
The root cause of Japan's defense tech startup drought is VCs avoiding defense investments. Anduril won a CCA contract within a few years of founding.
The UK developing StormShroud separately within the GCAP framework is actually a chance for Japan to learn. That's the real benefit of joint development.
The expendable missile-drone ARMDC-20X is an interesting cost-conscious approach. Cheap and disposable might be more practical in real combat than expensive and irreplaceable.
Testing both Turkey's Bayraktar and Israel's Heron is pragmatic. Better to import proven systems while developing domestically than insist on going native and fall 10 years behind.
I understand people opposing AI weapons. But if the other side uses AI, do we fight back with humans only? The gap between ideals and reality hurts.
Had no idea SUBARU makes drones. Only knew them for cars, but their 'Aerospace Company' division name means they're serious.
Defense budget over ¥9 trillion sounds huge, but only about ¥100 billion for drones. Need a bigger share or 'fundamental strengthening' is just a slogan.
Same pattern as semiconductors — 'world-class basic research, laps behind in commercialization.' Worried the drone field follows the same trajectory.
In Ukraine, drones costing tens of thousands of yen destroy tanks. Clearly the era of expensive manned aircraft is ending, yet Japan still buys F-35s at ¥18.7 billion each.
Drone development gets attention, but how does MoD secure AI engineers? Private sector pays better, plus security clearance barriers. The talent problem might be more critical.
The insight that 'combat AI can only see what its sensors detect' stuck with me. Better sensors = higher cost = can't mass-produce. 'Lots of cheap drones' is easier said than done.
GCAP targets 2035 completion, but by then AI drones will have evolved so much that manned fighters themselves might be questioned. Should we design for future battlefields with today's thinking?
A friend working on the USAF CCA program says Anduril's speed is insane. For Japan to produce a company that goes from founding to prototype first flight in a few years, it needs a fundamental shift in how VCs engage with defense.
Honestly, Japan's drone development isn't a threat to China. ¥4.8 billion budget? China spends more on a single airshow. But the Japan-UK-Italy framework deserves attention.
German defense industry analyst here. Japan and Germany share remarkably similar challenges — history of export restrictions, domestic market dependency, public caution on AI weapons. There's significant room for cooperation.
As a Brit, I'm bullish on GCAP. But as StormShroud shows, the UK decided to develop loyal wingmen without waiting for the manned fighter. Japan should consider decoupling drone development from GCAP timelines too.
Korea is also developing loyal wingmen for KF-21, but Japan's two-company competition between MHI and KHI seems like a good structure. Korea feels too concentrated on Korean Air.
India's HAL is also developing the CATS Warrior wingman drone, completing engine ground runs in 2025. Like Japan, we lack combat data, but we have cooperation with Israel as an advantage.
As an Italian, rooting for our GCAP partner Japan. Leonardo has drone tech too — integrating all three nations' technologies has potential against US/China. The problem is always slow political decision-making.
Living in Taiwan. Japan's drone development matters for us too. Stronger Japanese defense capability raises China's invasion cost in a Taiwan Strait scenario. Honestly wish they'd move faster.
Polish military blogger here. The biggest Ukraine lesson: in drone warfare, '100 good-enough units' beats 'one perfect unit.' Can Japan's defense industry make this conceptual shift?
Singapore defense researcher. Interesting that Japan is evaluating Bayraktar TB2S and Heron — both popular in Southeast Asia. Japanese adoption could deepen interoperability with ASEAN nations.
Aussie here. An engineer from the MQ-28 Ghost Bat program told me the hardest part of modular design is interface standardization. Japan needs to nail this early or it'll cause headaches later.
Aviation enthusiast in the UAE. I respect Japan's tech capability, but prioritizing 'ethical debate' over development speed is hard to understand from a Middle Eastern perspective. Real threats don't wait for philosophical conclusions.
I'm Ukrainian. We live the reality of drone warfare every day. My advice to Japan: don't aim for perfection. Deploy at 80% completion, improve with field feedback. That's how you survive.
Work in French defense. Europe's FCAS (Franco-German-Spanish next-gen fighter) also includes CCA development, but Dassault vs Airbus leadership disputes cause delays. Japan's GCAP governance looks clearer.
As a Brazilian military commentator, what catches my attention is Japan testing foreign drones. For countries with immature defense industries, 'learn while buying' is the most efficient strategy.
From a Vietnamese perspective, I hope Japan establishes drone technology and eventually transfers it to friendly nations. In South China Sea security, Japan's tech capability is a vital regional asset.