China has clear ambitions for collaborative combat aircraft (CCAs), but while there are indications of frenetic activity, progress and actual capabilities are difficult to gauge.
The Peoples’ Liberation Army parade of 3 September 2025 offered a brief yet informative glimpse of China’s CCAs, with several assets – or mock-ups – appearing on trailers towed behind trucks.
In the lead position was the tailless, flying wing GJ-21, understood to be the aircraft carrier-capable version of the GJ-11. There were two CCAs with distinct dorsal intakes, one of which featured a dramatic delta profile. Finally, there were two previously unknown tailless, single-engined CCAs roughly the size of the Chengdu J-10 fighter.
These add to other Chinese CCAs seen over the years, including the FH-97, which resembles the Kratos XQ-58A Valkyrie. An aircraft that looked like the original FH-97 – with a distinct dorsal intake – was part of parade rehearsals, but for some reason did not appear in the parade itself.
There is also the FH-97A, models of which featured at China’s Zhuhai air show in 2022 and 2024. Unlike the FH-97 it has engine intakes on the sides and closely resembles the Boeing MQ-28 Ghost Bat.
The development of CCAs is consistent with China’s broader push towards automation across all aspects of industry and life. As with western air forces, CCAs offer the Peoples’ Liberation Army Air Force (PLAAF) ‘affordable mass’ in future conflicts without the necessity of investing in larger numbers of valuable manned platforms and training large numbers of aircrew – likely a growing challenge given China’s deteriorating demographic outlook.
Moreover, the PLA has spelled out the importance of CCA’s directly. The US Air Force’s (USAF’s) China Aerospace Studies Institute recently translated a Liberation Army News article discussing the PLA’s view on the history and future of air warfare.
The report declares that “intelligentised unmanned combat systems” using artificial intelligence will operate independently, collaborating to engage targets.
“Unmanned aircraft will serve as wingmen to manned aircraft and will execute aerial combat missions by the form of manned-unmanned teaming,” says the report.
Like western CCA concepts, the report sees Chinese CCAs acting as offboard sensor and weapons platforms for manned assets, with the ability to conduct “limited autonomous combat under the command of manned aircraft”.
It adds that this form of CCA employment is starting to “germinate” and has been “verified”.
Two Chinese manned aircraft appear optimised for managing CCAs. The J-20S is the world’s first known two-seat stealth fighter and appeared in the skies over Chengdu in November 2021. It also showed up in the flying segment of the September 2025 parade and is believed to be in its early days of PLAAF service.
Speculation suggested that a key role for the J-20S is acting as a command centre of sorts for CCAs. Indeed, AVIC issued a statement about the type at the Zhuhai air show in November 2024 that appears to confirm a CCA-management role.
“[The J-20S] possesses excellent medium- and long-range air superiority capabilities, as well as precision strike capabilities against land and sea targets,” says the Chinese airframer.
“It also stands out for its advanced situational awareness, electronic warfare, and tactical command and control capabilities. The aircraft can conduct coordinated operations between manned aircraft and drones.”
The other Chinese type that online speculation closely associates with CCAs is the three-engined J-36, which first appeared over Chengdu in late 2024. As with the J-20S, the J-36 has a second crew member. It is possible, even likely, that the second crew member’s role is orchestrating CCAs.
Indeed, it could well be that both J-36 crew members, sitting side-by-side for easier communication, are tasked with managing other assets in the battlespace, leaving artificial intelligence to fly the aircraft itself. Similarly, it is believed that the pilot of the USAF’s developmental Boeing F-47 will not be a “pilot” in the traditional sense, but rather a systems manager with far broader responsibilities than just their own aircraft.
Longer term, according to the PLA, unmanned systems will operate largely autonomously, becoming “smart swarms” that adapt to battle conditions with little to no human oversight. Unlike traditional fighter aircraft, which tend to be exquisite standalone systems with unique characteristics and capabilities, such CCAs will feature a high degree of commonality, particularly with their software and autonomy systems.
Chinese planners appear to understand the value of commonality for the country’s array of CCA efforts. Chinese journals have referred to a mission system that can be used on different platforms. This work has largely been driven by China Electronics Technology Group and airframer AVIC.
China is also working on robust datalinks that will connect CCAs to each other and manned aircraft. Moreover, China is understood to be working on technology whereby if a CCA’s connection to a manned aircraft or leader CCA is cut, then it can automatically establish a new network with surviving CCAs, determine the leader, and then continue the mission.
For its part, the USAF has its Autonomy Government Reference Architecture (A-GRA) that can be used on different CCAs. The system was recently introduced by USAF autonomy vendors RTX Collins and Shield AI into the two different CCA test assets: the General Atomics Aeronautical Systems YFQ-42A and the Anduril Industries YFQ-44A.
This approach offers a universal standard for US and allied CCAs, into which software and algorithms can be ported in from various vendors.
Of China’s various CCA efforts, it appears that the most advanced is the GJ-11 Sharp Sword. In November 2025, to celebrate its 70th anniversary, the PLAAF released footage showing the type in flight. The brief clip showed the tailless aircraft flying in formation with a J-20 and a Shenyang J-16D electronic warfare jet.
The presence of the J-16D in the shot could hint that the GJ-11 has a role in the suppression of enemy air defences/destruction of enemy air defences (SEAD/DEAD) mission. Indeed, an early role envisaged for CCAs in the West is SEAD/DEAD.
The clip also shows the arrowhead-shaped GJ-11 being towed from a hangar, as well as the aircraft just after taking off, landing gear still down. Still, the level of ground controller involvement in GJ-11 operations remains unknown. As to whether it is in service or testing, in 2024 satellite images showed what appeared to be three GJ-11s on the apron at a Chinese airbase in Tibet, suggesting at least limited operations.
More intriguing is the GJ-21, which is believed to be in testing. In 2025 images appeared to show the aircraft – or more likely a mock up – on the deck of the new Type 076 amphibious assault ship CNS Sichuan.
With its twin island, the type 076 is unlike any western amphibious assault ship and features a single electromagnetic catapult system for launching fixed-wing aircraft. The Institute for the Study of War suggests that up to six GJ-21s could operate from type 076 vessels.
Photos that purport to be of the GJ-21 with its landing gear extended also show that it has a tailhook. This could mean it is also intended to serve aboard China’s new CATOBAR – catapult assisted take-off but arrested landing – carrier CNS Fujian.
What appear to be mock-ups of a twin-tailed, dorsal intake CCA also appeared recently on the deck of the freighter Zhong Da. It seems that the concept calls for the aircraft to be launched with a series of trailers joined together to create a catapult system. It is possible that the images are not even genuine. If the catapult were to stay horizontal, as depicted in the images, a launching CCA would probably slam into the bow of the ship – possibly the intention is to angle the catapult upwards.
It is also not clear how a second CCA, sitting on the deck below the catapult, will be lifted into launch position. The ship also bristles with vertical missile launch tubes that can be hidden in containers, has a naval rotary cannon turret to destroy incoming missiles, as well as a radar and a satellite communications system.
As for the FH-97, which has appeared at Zhuhai, little is known about its status, although there is speculation it may have entered early service. The type seems to have gone through a design change, with the updated FH-97A using side-by-side intakes as opposed to the dorsal intake of the FH-97.
In November 2024 at Zhuhai, an FH-97A mock-up appeared in the static park, accompanied by several loitering munitions. A pair of air-to-air missiles were visible in the aircraft’s central weapons bay.
As for the two large tailless CCAs in the September 2025 parade, there is only speculation. One CCA incorporates a cranked-kite lambda wing design. On-line speculationn suggests that it will be used for intelligence, surveillance, reconnaissance, and electronic warfare operations. Conversely, the larger, diamond-shaped CCA may be intended for use in both air-to-air battles and strike operations.
Despite the variety of platforms, observers should avoid focusing too much on specific airframes and their features. Indeed, though they represent the future CCAs are somewhat of poor cousins to advanced manned platforms. If truly attritable, they feature less capable engines, sensors, and overall kinematic performance than the advanced manned jets they will complement.
Their real strength lies in the capabilities of the software or AI that governs them and how they complement the overall network. A single, robust software or AI function that can be seamlessly adapted to many different CCAs is more valuable than a collection of distinct aircraft.
Ultimately, China’s work on CCAs is best understood not as a collection of individual aircraft but as an evolving ecosystem — one in which software, autonomy, and connectivity matter more than any particular airframe. The parade glimpses, the satellite images, the air show mock-ups: these are data points in a much larger and still-unfolding story.
What is clear is that China is pursuing CCA development with characteristic breadth, fielding multiple platforms across naval and land-based roles while investing in the common architecture that will tie them together. Whether its autonomous systems can match the ambition of PLA doctrine remains the central unanswered question – but the pace of development suggests that question may not take long to answer.
