ISTANBUL

China has unveiled its first full-process demonstration of the Atlas drone swarm operations system, state media reported Wednesday, showcasing a platform designed to coordinate large numbers of drones for reconnaissance and strike missions.

The domestically developed system includes a Swarm-2 ground combat vehicle, a command vehicle and a support vehicle, according to the military channel of CCTV News.

Footage showed the launch vehicle carrying the logo of state-owned China Electronics Technology Group Corp, the Global Times reported.

The Swarm-2, introduced in 2024, can carry and launch 48 fixed-wing drones, while a single command vehicle can control up to 96 drones at once.

Each drone can carry different payloads, including electro-optical reconnaissance systems, strike munitions and relay communications equipment. They can be flexibly grouped into multifunctional swarms for complex missions, according to a previous CCTV News report.

The latest report Wednesday said the system conducted coordinated reconnaissance during testing, autonomously identified the command vehicle among targets and launched drones at a test range with three visually similar targets.

The Swarm-2 vehicle released one drone every three seconds, the report said.

The type and order of drones launched can be adjusted based on mission needs. Reconnaissance drones may be deployed first to gather intelligence, while electronic warfare drones can be sent ahead of attack drones to disrupt enemy systems.

The drones can autonomously adjust to environmental factors such as airflow disturbances and create dense, precise formations in a short time.

The report said a single operator can manage large-scale aerial operations, controlling up to 96 drones.

The system also includes drones of different sizes.

Chinese military affairs expert Wang Yunfei told the Global Times the system could greatly expand battlefield use, including saturation attacks on enemy air defenses, precision strikes and deep-strike missions.

He added that with AI-enabled pre-training and embedded algorithms, the drones can independently carry out tasks such as target recognition, task allocation and route planning while adapting to changing battlefield conditions.