A room full of Air Force combat crew members watched the YFQ-44A Fury taxi out of its parking spot and accelerate down the runway at Edwards Air Force Base in mid-April without a single engineer in a control booth overriding their inputs. That fact — that warfighters were the ones making the aircraft go — is the entire story. Not the drone's dimensions or its weapon hardpoints or the name of the command and control software. It is that the Air Force's Experimental Operations Unit took full operational control of an autonomous strike aircraft for the first time, and the handoff worked.
This matters because the Air Force's Loyal Wingman program, formally called Collaborative Combat Aircraft or CCA, is split between exactly two companies: Anduril Industries and General Atomics. The service wants 1,000 of these semi-autonomous drones to fly alongside F-22s and F-35s as weapons platforms, electronic warfare assets, and decoys. But it may only fund one design into production. The decision is due this summer. What happened at Edwards in the week of April 10 to 17 will almost certainly weigh that decision. Anduril just proved its aircraft works with the operators who will actually use it in war. General Atomics' competing YFQ-42A Dark Merlin had a takeoff accident recently. That asymmetry matters.
Here is what actually occurred. The Experimental Operations Unit, drawn from Air Combat Command, conducted a series of daily sorties at Edwards under the watch of the 412th Test Wing. The sorties covered the full cycle: pre-flight checks, weapons loading and unloading (using inert AIM-120 air-to-air missiles), autonomous taxi and takeoff initiation, in-flight tasking, and post-flight data management. Every input came from EOU operators on the ground using Anduril's Menace-T command and control system, a ruggedized laptop-based interface. Anduril personnel, who had previously flown all test missions, handed the aircraft off entirely. No stick and throttle operator hiding in the background. No engineers overriding autonomous decisions in real time. Lt. Col. Matthew Jensen, the EOU commander, said it bluntly: 'Every sortie generated and flown was done with a warfighter, not an engineer or test pilot, kicking the tires and controlling the prototypes.' The aircraft measures 6.1 meters in length with a 5.2-meter wingspan, roughly half the size of an F-16. Maximum takeoff weight is 2,270 kilograms. It has swept trapezoidal wings, a chin-mounted air intake, and a cruciform tail. It looks like a light fighter because that is essentially what it is.
The reason this happened as quickly as it did traces back to a single architectural decision Anduril made years ago. Instead of building a human-piloted drone and bolting autonomy onto it later, the company embedded autonomous taxi, takeoff, and flight planning into the airframe from the first prototype flight. That October 2025 maiden flight was semi-autonomous from day one. That decision ate up engineering time at the front end of the program but collapsed training time at the back end. EOU operators needed days of instruction to fly the Fury, not weeks. Col. Timothy Helfrich, the Air Force's portfolio acquisition executive for fighters and advanced aircraft, described the operational benefit explicitly: 'By embedding the operators from the EOU with our acquisition professionals, we create a tight feedback loop that lets us trade operational risk with acquisition risk in real-time.' Translation: the Air Force gets to see whether the aircraft actually works with its combat crews before committing a billion dollars to production. That is not how most defense acquisition happens.
The Edwards exercise also tested whether the Fury could sustain operations from a simulated forward site with minimal support infrastructure and high sortie rates. That is the actual combat scenario the drone needs to support. The Air Force's Agile Combat Employment doctrine calls for dispersing small numbers of aircraft across many forward locations rather than concentrating them on a few big bases. A drone that requires extensive logistics infrastructure fails that requirement. A drone that can turn around between sorties with a skeleton crew of maintainers wins. The Fury apparently cleared that bar. That is not a theoretical advantage. That is operational capability the Air Force told Anduril it needed to demonstrate.
General Atomics' competing aircraft, the YFQ-42A Dark Merlin, has not yet had an equivalent exercise with combat crews. The company experienced a takeoff accident with the Dark Merlin earlier this year. The War Zone reported that Edwards AFB was originally expected to host testing of both aircraft but did not, possibly due to that accident. That is the competitive picture: Anduril just ran a clean, operator-driven evaluation that proved the Fury works with the people who will actually use it in combat. General Atomics is recovering from a crash. In a downselect decision, that asymmetry is the entire ballgame. The Air Force might choose both programs to move into low-rate production. It might choose only one. If the service chooses one, Anduril has just handed itself a material advantage. The Fury has been proven in the hands of combat crews. The Dark Merlin has not.
This is the moment where the Pentagon's acquisition timeline actually collided with operational reality and produced something real. Not a prototype that looks impressive on a runway. Not a contract announcement. An actual combat crew taking full control of an autonomous strike aircraft, flying it, and not breaking it. The Air Force has been talking about Loyal Wingmen since 2019. Anduril has been building toward this single moment since it entered the CCA program in 2024. General Atomics is now behind on proving the same capability. The production decision is due this summer. Anduril walks in with demonstrated operator confidence and a clean test record. General Atomics walks in with a recent crash and no equivalent operator evaluation. That is not a feature of the program. That is the difference between winning and losing a billion-dollar contract.
The specific milestone to watch is whether the Air Force publicly announces the production downselect by the end of Q3 2026, as expected. If Anduril is chosen, look for an announcement that frames operator feedback from the EOU exercise as a decision factor. If both companies are selected for low-rate initial production, watch for contract values and quantities — the Air Force's way of signaling which design it actually prefers. Third, monitor whether the Air Force conducts a similar operator evaluation with the Dark Merlin before the decision. If it does not, that silence is its own answer about the competitive posture.