At roughly 1 p.m. Pacific on April 6, one of General Atomics’ prototype YFQ-42A drones lifted off from a company airfield in the California desert, climbed briefly — and then came back down the hard way. The aircraft, a production-representative model of the Air Force’s Collaborative Combat Aircraft programme, crashed shortly after takeoff. No one was hurt. Every subsequent test flight has been grounded pending an investigation.

The timing could not be worse. The Air Force is preparing to make a production decision on its CCA programme within the next six months — the single biggest procurement question in unmanned combat aviation. General Atomics is competing head-to-head with Anduril, and a crash like this, even if the cause turns out to be minor, hands the competition a talking point.

The CCA concept is deceptively simple: build semi-autonomous drone “wingmen” cheap enough to lose in combat, smart enough to fly alongside manned fighters, and lethal enough to matter. The Air Force wants them at under $25 million apiece — a fraction of a manned jet — so they can be deployed in numbers that overwhelm enemy air defences.

A Billion-Dollar Bet on Unmanned Wingmen

Budget documents released days before the crash show the scale of the Pentagon’s ambition. The FY2027 request includes $996.5 million in CCA procurement funding — nearly a billion dollars — plus another $150 million in advance procurement for 2028. Add in $1.37 billion for continued research and development, and the total 2027 ask tops $2.3 billion.

Since fiscal year 2024, the service has poured almost $1.91 billion into developing the CCA concept. If the per-unit cost stays below the $25 million target, the 2027 procurement funding alone could buy roughly 40 aircraft. The Air Force has not disclosed how many it plans to order.

General Atomics calls its CCA family “Gambit.” The YFQ-42A involved in Sunday’s crash was one of several production-representative prototypes built specifically for Air Force evaluation — not a rough proof-of-concept, but an aircraft intended to demonstrate what a production model would look and fly like.

What Happens Next

General Atomics issued a brief statement confirming the mishap and noting that “established procedures and safeguards worked as intended.” The company opened a formal investigation and paused all YFQ-42A flight testing until the cause is identified. No timeline has been given for resuming flights.

The crash comes at a particularly intense moment for the programme. In February, General Atomics completed what it called a landmark demonstration — teaming an MQ-20 Avenger with an F-22 Raptor at Edwards Air Force Base, proving that a drone wingman could coordinate in real time with America’s most advanced air-superiority fighter. That success generated genuine momentum.

Whether this setback slows the programme depends entirely on what investigators find. A software glitch or a one-off mechanical failure can be fixed. A design flaw in the airframe or flight-control system would be a different story — one that could shift the balance of the competition toward Anduril or open the door for Northrop Grumman’s Increment 2 bid.

The Bigger Picture

The CCA programme is not just another procurement line item. It represents a fundamental rethinking of how the Air Force fights. Instead of depending entirely on small fleets of expensive, exquisitely capable manned platforms — the F-35, the coming F-47 — the idea is to surround those jets with swarms of cheaper, expendable drones that can scout, jam, and strike.

If it works, it changes the math of air combat. An adversary that shoots down a $25 million drone instead of a $100 million fighter with a pilot inside has won a battle but lost the exchange. Multiply that across dozens of drones per sortie, and the calculus of air defence shifts dramatically.

One crash will not kill that vision. But it is a reminder that building autonomous combat aircraft — ones reliable enough to trust in contested airspace — remains one of the hardest engineering challenges in modern aviation. The desert test range just proved it.

Sources: Air & Space Forces Magazine, Breaking Defense, Defense One, General Atomics