A US F-22 Raptor controlled an MQ-20 Avenger drone over Nevada, a key step towards future Collaborative Combat Aircraft and swarm warfare.
Summary
On November 18, 2025, the US Air Force confirmed that an F-22 Raptor pilot had controlled an MQ-20 Avenger, a jet-powered combat drone, during an exercise over Nevada. The demonstration, conducted in mid-October at the Nevada Test and Training Range with General Atomics, Lockheed Martin, and L3Harris, used secure data links and a tablet interface in the cockpit to fly the MQ-20 as a “loyal wingman.”
This test marks a concrete step towards Collaborative Combat Aircraft, drones that accompany fighters to extend their sensors, saturate enemy defenses, and accept risks that would not be taken with a manned aircraft. In a context of strategic competition with China and Russia, the US Air Force is seeking to increase the number of platforms while controlling costs and potential losses. The F-22/MQ-20 test shows that the technology now works in actual flight and that the questions are no longer about feasibility, but about industrialization, tactics, and the limits we are willing to set for these autonomous systems.
The strategic framework of Collaborative Combat Aircraft
The F-22/MQ-20 demonstration is part of the US Collaborative Combat Aircraft program, the core of the next generation of air power. These CCAs are designed to fly in patrol with fighters, share their sensors, carry weapons, and sometimes expose themselves on the front line. The US Air Force is considering purchasing 100 to 150 drones for “Increment 1” alone, with hundreds more units in the longer term, for operational service by the end of the decade.
In this scenario, the F-22 is no longer just an air superiority fighter, but a “pack leader” coordinating several drones. The aim is to compensate for the relative numerical weakness of stealth fighter fleets with a mass of less expensive autonomous platforms. These CCAs are designed to be air superiority multipliers: they can carry air-to-air missiles, electronic warfare pods, or long-range sensors, while accepting mission profiles that are too dangerous for a manned aircraft.
The demonstration over Nevada
The test flight took place on October 21, 2025, at the Nevada Test and Training Range, a vast test area covering tens of thousands of square kilometers, used for high-intensity scenarios. The flight was financed with industry funds, without a specific contract from the US Air Force, underscoring the competitive stakes surrounding CCAs.
An F-22 flew in a simulated mission configuration, while an MQ-20 Avenger played the role of autonomous wingman. Both aircraft were equipped with Pantera software radios and connected by the BANSHEE datalink provided by L3Harris, integrated into an open radio architecture developed by Lockheed Martin. The F-22 was equipped with a GRACE (Government Reference Architecture Compute Environment) module, designed to quickly accommodate new software, as well as a tablet-based control interface, the Pilot Vehicle Interface (PVI).
During the flight, the pilot was able to send commands to the MQ-20, modify its trajectory, assign it a simulated mission profile, and check feedback, all from a single-seat cockpit that was already very busy. The Avenger flew with mature “mission” autonomy software, capable of performing complex tasks once the intention was defined.
The role of the F-22 Raptor as mission leader
The choice of the F-22 Raptor as the first airborne CCA controller is not insignificant. The US Air Force designated it as the “platform threshold” in a report to Congress, i.e., the reference platform for the first integrations of collaborative drones.
The F-22 remains one of the world’s most capable fighters, with a very low radar signature and a flight capacity of over 15,000 m (50,000 ft). Its modernized avionics, with open modules such as GRACE, make it an ideal test bed for future manned-unmanned teaming architectures. Its presence in the most demanding scenarios—air defense of contested areas, deep penetration, long-range interception—corresponds precisely to the type of missions where CCA would make a decisive contribution.
Ultimately, the US Air Force plans to extend this role to other platforms: the F-35A, F-16, F-15E, and F-15EX are cited as candidates. But the F-22 serves as an operational laboratory: it will be used to test the first tactics, cockpit workload, human-machine interfaces, and rules of engagement for a mixed swarm.
The MQ-20 Avenger platform, a flying laboratory for cooperation
On the drone side, the MQ-20 Avenger is a jet-powered UCAV, approximately 13 m (44 ft) long with a wingspan of 20 m (66 ft). It can reach speeds of around 740 km/h (460 mph), climb to over 15,000 m (50,000 ft) and stay in the air for up to 20 hours, with a range of around 5,800 km (3,600 mi). It carries its weapons in an internal bay, which reduces its radar signature, and can carry nearly 3 tons of payload (sensors, weapons, pods).
Above all, the Avenger has been used for years as a test bed for advanced autonomy software, notably with Shield AI’s Hivemind AI, or in “Avenger ER” configuration for flights lasting more than 23 hours. Within the CCA framework, it serves as a “surrogate,” i.e., a representative platform, pending the arrival of the dedicated YFQ-42A (General Atomics) and YFQ-44A (Anduril) prototypes, which are already in test flights.
This choice is pragmatic: the Avenger exists, flies, and can be quickly instrumented. It allows for testing of links, autonomy, and the interface with fighters, without waiting for the maturity of future production CCAs.
The technological challenges of control from a single-seat cockpit
The demonstration validates the technical chain—radios, datalink, GRACE software, tablet—but it opens up a central debate: how much can we load onto a pilot who is already saturated with information? Officials at General Atomics and Lockheed Martin themselves acknowledge that flying a single-seat fighter while supervising a drone via a tablet is complex, especially in a combat environment.
In the scenario tested, the pilot does not “remotely pilot” the MQ-20 in detail. Instead, he defines objectives and areas, with the onboard autonomy managing the trajectory and micro-decisions. The real challenge is therefore the level of abstraction of the interface: giving enough control to exploit the drone’s potential without turning the pilot into an overwhelmed RPA operator.
The demonstration also highlighted the use of open, non-proprietary, government-owned data links. This allows the US Air Force to avoid being locked into a manufacturer’s systems and, ultimately, to connect several families of CCA, surface sensors, or allied platforms on the same architecture.
Expected operational gains in air superiority
Tactically, this first F-22/MQ-20 cooperation opens up many options. A combat drone such as the Avenger can fly ahead of the patrol, illuminate the airspace with radar or an infrared sensor, and transmit the tactical situation to the F-22. It can also act as bait, forcing the enemy to reveal its ground-to-air defenses or fighters.
With a payload of several thousand kilograms, the MQ-20 can carry long-range air-to-air missiles, guided bombs, or electronic warfare pods. Coupled with a low-signature F-22, it enables unprecedented combinations: for example, an Avenger generating electromagnetic noise to saturate radars, while the Raptor slips through a window of stealth to engage a key target.
In a future swarm of Collaborative Combat Aircraft, a single F-22 could coordinate several drones, each optimized for a specific task: sensors, jamming, strike, communications relay. The effect on air superiority would be twofold: multiplying the axes of attack and drastically complicating the enemy’s defense.
Gray areas and questions ahead for the US Air Force
Technical success does not mean that all issues have been resolved. Manufacturers remain cautious about the level of autonomy actually used during flight and the rules of engagement envisaged for these systems. The US Air Force will have to decide on sensitive issues: how much autonomy to give the drone, how to divide responsibility between the pilot and algorithms, and how to certify these behaviors for real-world operations.
There are also very concrete budgetary and industrial challenges. Producing dozens, then hundreds, of CCAs capable of tracking stealth fighters will require new standards for maintenance, logistics, and training. The first production decisions are expected in 2026 for Increment 1, but the exact trajectory—one or more types of drones, delivery schedule, integration with allies—remains confidential.
Finally, the F-22/MQ-20 demonstration sends a clear signal to China, Russia, and regional powers: the next stage of aerial warfare will no longer be fought solely between manned fighters, but between networks of autonomous platforms, linked by open architectures and remotely piloted by a few highly skilled crews. The question is no longer whether this model will prevail, but at what pace, and with what degree of human control over the decision to fire.
Sources
- Air & Space Forces Magazine, “F-22 Pilot Controls ‘CCA Surrogate’ Drone from Cockpit,” November 17, 2025.
- Breaking Defense, “In a first, F-22 pilot controls wingman drone from cockpit,” November 17, 2025.
- The War Zone, “F-22 Pilot Controls MQ-20 Drone From The Cockpit In Mock Combat Mission,” November 17, 2025.
- FlightGlobal, “F-22 successfully controls MQ-20 uncrewed jet in demonstration flight,” November 18, 2025.
- MQ-20 Avenger technical specifications: Designation-systems, Missile Defense Advocacy Alliance, General Atomics.
War Wings Daily is an independant magazine.