Ravn X Program Continues Progress

Inside Ravn X SN0, the purpose built test vehicle that is proving autonomy before the flight capable airframes actually liftoff. Ravn X has a length of 80ft, a wingspan of 60ft, a height of 18ft, and a gross takeoff weight of 55,000lb.

For years many have seen Aevum’s Ravn X as a small corner in the aerospace industry. The design is currently in the testing stage with two prototypes in production.

As for the test, founder and CEO Jay Skylus described it as an attempt, not a victory lap. They were able to achieve most of their objectives and the autonomy was “learning the whole time” as described by Jay. This was an incredible milestone for the company supported by the U.S. Space Force and investors.

I was with my buddy Ricky and we were shocked to see the testing going on! For a small local airport, Long Beach serves as the future of space development and technologies. With Aevum, Relativity Space, JetZero, and in the future Anduril, the technology made from the airport will be tremendous. 

Long Beach serves as Aevum’s Autonomous Flight Research Center. It’s mainly a research and developmental place for the company. 

The Vehicle That Everyone Got Wrong

There has been a misunderstanding since Aevum's 2020 rollout: the airframe undergoing testing, known internally as SN0, isn't a mockup, and it isn't meant to fly. Skylus has said this going back to the original unveiling. Aevum built three airframes from the start. The first is a composite ground test article, structurally real but never intended to leave the runway. The second and third are the flight-capable vehicles, and according to Skylus those are now in production.

That distinction is the whole ballgame for understanding what just happened. SN0 exists to take on the riskiest, least glamorous work in the program: proving out autonomous taxi behavior, ground handling, sensor fusion, and the software that has to make thousands of small decisions before a Ravn X vehicle is ever cleared to add an engine light-off, let alone a wing airborne.

Why Build Such a Controversial Yet Difficult Vehicle?

Aerospace has a long tradition of exactly this kind of hardware. Airliner programs build static and fatigue-test airframes that spend their whole service life bolted to a test rig, absorbing decades of simulated flight cycles so engineers can find failure points long before a passenger ever boards the real thing. Full motion "iron bird" rigs, another staple of flight test programs, let engineers exercise flight control software and hydraulics on the ground under stable and controllable conditions. 

Ravn X's ground test article is a variation on that same logic, adapted to an autonomous vehicle instead of a piloted one. Before Aevum's autonomy software ever has to fly, land, and taxi a real rocket-carrying aircraft with no pilot backup, it has to prove it can taxi one without incident, again and again, in front of the FAA, airport operations, and the company's own engineers. Doing that on a dedicated ground article means a taxiway excursion, a sensor fault, or a software hiccup costs a composite structure built for exactly that risk, not one of the two production airframes meant to carry an actual rocket to altitude. Quite frankly, this vehicle looks pretty sturdy. Im excited to see what the future holds.

It also means the flight vehicles, in theory, can skip straight to flight test points instead of relearning ground behavior from scratch, since the taxi, brake, and situational-awareness logic will already have been exercised on SN0. Of course they can test that with the first few movements with power but that knowledge will have already been learned without propulsion of any sort.

Why This Buys The Flight Vehicles?

The payoff of this approach shows up downstream. Every taxi run, every edge case the autonomy stack encounters on SN0, becomes training data and validated logic that carries straight over to airframes two and three. Those are the vehicles Skylus says are already in production. Rather than the flight capable Ravn X discovering ground handling quirks for the first time with a rocket payload aboard, that vehicle inherits a system that's already logged real world data. It's the same reason airlines trust fatigue tested airframes and pilots trust simulator hours: the learning happens where a mistake is cheapest, so the real vehicle shows up already ahead.

Skylus has said the team will coordinate its next test opportunity soon, and each one adds to the same growing dataset. For a program built around full autonomy from taxi to orbit, that's exactly the order of operations you'd want: teach the system to walk, thoroughly, before asking it to fly. I am excited for whenever the next test is and I look forward to hopefully capturing it when the day comes. I would like to thank the Aevum team for being such a great group of folks and I wish them the best in their future endevours. If they ever need a photographer to help out with their program and social media pages I would be more than willing to help!