The Pentagon’s vision of cheap robot fighters flying shotgun for crewed jets took two big steps this week. One went forward. The other went straight into the Utah desert.

Kratos Defense & Security Solutions has begun building the first XQ-58A Valkyries fitted with landing gear — a version that can take off and land on an ordinary runway instead of being flung skyward on a rocket booster and floated back down under a parachute. It is the configuration the U.S. Marine Corps has been asking for, and the company says several more airframes are already taking shape on the line.

Then, on 4 August, one of them crashed.

A wingman you can afford to lose

The Valkyrie is the poster child for a whole new category the U.S. calls the Collaborative Combat Aircraft — an uncrewed jet cheap enough to send into places you would not risk a piloted fighter, and smart enough to fight while it is there. Stealthy, high-subsonic and long-legged, it is built to fly alongside a F-35 Lightning II or an F-22 and do the dirty, dangerous work: scouting ahead, carrying extra missiles, soaking up the first shot.

Kratos has been flying the type since 2019, when the AFRL demonstrator first left the ground. What has changed is the price of admission. The whole point of the Valkyrie is that it is attritable — a word the military uses to mean “we would rather not lose it, but we can.” That philosophy is exactly why this week’s two headlines are not the contradiction they look like.

Wheels change everything

Until now the Valkyrie has launched from a rail with a rocket booster and come home under a parachute — no runway required, but also no easy way to turn it around and fly it again the same afternoon. The new landing-gear version fixes that. It can roll down a runway like any other jet, or still be booster-launched and then land conventionally, giving planners the best of both worlds.

That flexibility is what the Marine Corps wants for its MUX TACAIR effort, and Kratos has reportedly launched production of a second batch of two dozen airframes to meet the interest.

Then one went down at Dugway

On 4 August, a Valkyrie flying in what Kratos called a “test configuration” crashed shortly after takeoff at Dugway Proving Ground in Utah. Nobody was hurt. The company described the flight as an internally funded research effort — “part of continually adding configurations and capabilities” — and said flight-test operations had been “paused temporarily in an abundance of caution” while it works out what went wrong.

Here is the thing: for a program built around expendability, losing a test article is a bad afternoon, not a crisis. The airframe is cheap by fighter standards, the crash hurt no one, and the production line for the version the Marines actually want keeps moving. A decade ago a mishap like this could have frozen a drone program for a year. For the CCA generation, it is the cost of learning fast.

Watch an earlier Valkyrie test flight to see why the concept turns heads — a jet-shaped drone rocketing off a rail with no pilot aboard:

Sources: The War Zone, FlightGlobal, Defence Industry Europe, Kratos Defense.