cokada

Five firms—don’t ask who—will vie to make brains for USAF robot wingmen

The Air Force, which is keeping the names under wraps, will pick a winner in time for 2026 production.

DAYTON, Ohio—Five companies will get money to develop autonomous control systems for the Air Force’s planned robot wingmen—and the service is keeping the list a secret to safeguard the designs from spies.

One or more will eventually be chosen to produce the systems, which will be integrated onto collaborative combat aircraft built by General Atomics and Anduril, the Air Force’s program executive officer for fighters and advanced aircraft said.

“There will be a downselect process based on knowledge gained over time as a function of performance of those autonomy vendors, and we will get down to a lower number of those vendors and they will create an autonomy implementation that will be applied to CCA increment one,” Brig. Gen. Jason Voorheis told reporters on Monday at the Air Force’s Life Cycle Industry Days conference.

But the identities of the autonomy companies will stay secret in a bid to keep their technology hidden from prying eyes, officials said. 

“If you can think of the mission autonomy as the brain of your autonomous vehicle, it is important that many protections are put around that, which drove the different classification posture,” said Col. Timothy Helfrich, senior material leader for the advanced aircraft division. But he did say the companies are a mix of “both traditional and non-traditional vendors.”  

Several companies that have been working to develop an AI pilot could be in the running—like EpiSci and ShieldAI, both of which had their AI agents flying on Air Force Secretary Frank Kendall’s AI-piloted F-16 ride earlier this year. 

The autonomy must be ready in time for the Air Force to make a production decision on increment one in 2026. General Atomics and Anduril have to fly their CCA offerings before the service makes its decision. 

Whether the service decides to carry both of the physical aircraft vendors through production depends on the performance of their prototypes, Voorheis said. But, he said, he doesn’t anticipate either falling short because both are doing “very well,” and the service went through a rigorous selection process to choose companies that would be successful on increment one.

“I would say that our full intent is to have a competition in the 2026 timeframe. It is equally likely that we will carry both of those vendors through to production,” Voorheis said. 

“We're not going into it with any notion of an equal split or a specific 60-40 split. It will be wholly dependent on the performance of each of those vendors, and then the capacity that [the] department, our joint partners, or international partners need,” Voorheis said. 

The program also will kick off “increment two” next year, Voorheis said, and award contracts for the concept design phase in fiscal 2025. The service is still figuring out the requirements for the next tranche of CCAs, and may bring in international partners for increment two development, Voorheis said. 

“The aperture opens significantly on CCA increment two. As early as right now, and certainly as we go into the formal concept refinement contracts in fiscal [20]25, our intent would be to bring in interested international partners to help us shape what increment two looks like, and we would even be open to increment two co-development, co-production type of arrangements if that's what's of interest to foreign partners,” Voorheis said.