Every argument about autonomous combat aircraft eventually arrives at the same question, and it is not a technical one. It is whether the person in the cockpit will trust the thing flying off their wing.

Frank Kendall, who ran the US Air Force as its Secretary from 2021 to 2025, spent an hour with the Mitchell Institute on 16 July arguing that this is now the critical variable in the Collaborative Combat Aircraft programme. Not thrust. Not sensor fusion. Trust. And the way you build it, he suggested, is to put the drones into Red Flag and see what happens.

He also had an answer for anyone expecting fighter pilots to resist being handed robot wingmen.

The quote that matters

The expectation, endlessly repeated in commentary, is that fighter pilots will fight this. They are a famously proud community whose entire professional identity is bound up in being the person in the aeroplane. Hand them an uncrewed wingman and you are, in theory, handing them their own replacement.

Kendall put that question to an Air Combat Command pilot working on CCA tactics, and reported the answer.

That is the whole programme in one sentence. A CCA is not competing with the pilot for glory. It is competing with the pilot for the job of being first into the surface-to-air missile envelope. Framed that way, the cultural resistance everyone predicted largely evaporates.

The full Aerospace Nation session is above if you want Kendall unfiltered rather than filtered through a blog post.

Why Red Flag specifically

Red Flag, run out of Nellis since 1975, exists because of a grim statistical discovery from Vietnam: a combat pilot's odds of survival climb sharply after roughly ten missions. The fix was to manufacture those first ten missions in the Nevada desert, against a professional aggressor force, before anybody shot back for real.

That is precisely the problem CCAs now present, with one adjustment. It is not the drone that needs the ten missions. It is the human learning to work with it.

Kendall's specific interest is a number nobody yet knows: how many drones one pilot can actually direct while also flying a fighter and fighting an engagement. The marketing answer has always been vague and optimistic. The real answer depends on cockpit workload, interface design, how much autonomy the aircraft has, and how tired the human is. You do not find that number in a laboratory. You find it in an exercise, repeatedly, with people who are trying to win.

Convenient, then, that the Experimental Operations Unit evaluating both CCA types already sits at Nellis, which is where Red Flag happens. The aircraft and the exercise are on the same ramp.

How far along this actually is

Further than most people realise. Both Increment 1 aircraft have flown, the YFQ-42A in August 2025 and the YFQ-44A in October 2025. In June 2026 the Air Force declined to pick a winner and instead ordered both, contracting General Atomics and Anduril for the first 150 aircraft, with deliveries expected before the end of the decade. A readiness unit has been earmarked for Beale.

And on 15 July, days before Kendall spoke, a YFQ-44A fired a live AIM-120 AMRAAM at a target over the Mojave, the first air-to-air missile shot by an American collaborative combat aircraft. A human held weapons release. But the aeroplane that carried it, tracked with it and launched it had nobody aboard.

So the sequence is roughly: it flies, it shoots, it gets ordered in quantity, and the remaining question is whether squadrons can actually fight with it. That last step is the one Red Flag answers, and the Air Force has not yet said when the drones will show up in Nevada.

On current form, it will not be long.

Sources: Air & Space Forces Magazine, Mitchell Institute for Aerospace Studies, TWZ, Airforce Technology, Congressional Research Service, US Air Force.