By: Houston Cantwell (Brig Gen USAF Ret.), Autonomy Global Ambassador – Defense
The Air Force’s vision for Collaborative Combat Aircraft (CCA), autonomous jets linked to fighter jets to multiply combat power and lower risk in the highest threat environments, once seemed like science fiction. At this year’s Air Force Association Warfare Symposium, three Air Force experts (Brigadier General David Epperson, commander of the U.S. Air Force Warfare Center at Nellis Air Force Base, Colonel Timothy Helfrich, Program Acquisition Executive For Fighters And Advanced Aircraft at AFLCMC, and Lieutenant Colonel Matthew Jensen, Commander of Nellis’ Experimental Operations Unit (EOU)) all described the program as very real, moving on parallel tracks of fast acquisition, early operational integration and deliberate trust-building. CCA isn’t just an aircraft, but rather a new weapon system that includes software, mission planning, debrief tools, sustainment and training. The harder task will be operationalizing the concept with operators who can trust autonomy and an enterprise that can field and sustain them.
Flying Now But Not Yet Fielded
Increment One continues to advance with two contracted air vehicle companies, General Atomics and Anduril, who moved from contract award to first flight in less than 18 months. Helfrich framed this as a “success story” of acquisition speed and a template for acquisition reform. He also stressed that Air Force Research Laboratory (AFRL) work dating back to Skyborg enabled the Air Force to move so quickly into Increment One by accelerating autonomy maturation and setting conditions for rapid prototyping.
In parallel, the Air Force is executing a “software-sold-separately” approach, integrating two different mission autonomy vendors into the program. General Atomics flies with Collins Aerospace mission autonomy software and Anduril with Shield AI software.
While Increment One aircraft are actively flying and testing, panelists noted that flight tests alone do not equal readiness. CCA will succeed only if the Air Force builds the surrounding ecosystem as quickly as it builds the air vehicles. That will require robust testing and the not-so-glamorous fundamentals behind any new capability like doctrine, organization, training, maintenance, modernization, facilities and policy.
Near-term priorities involve progressing toward weapons delivery, continuing “quarterback” demonstrations (including an F-22 controlling a surrogate CCA with field-representative equipment), integrating actual CCAs with crewed platforms this year and completing source selections for production of both air vehicles and mission autonomy by the end of 2026.
Bringing Warfighters “Left” To Compress Timelines
Jensen explained how his unit, the EOU, is designed to solve the historic Air Force problem of developing systems in a long, linear sequence that delays progress. The EOU’s mission, he said, is to discover how to employ and integrate collaborative combat capability on “relevant timelines,” less than the traditional 10–15 years and as short as two to three.
The key difference is timing and feedback. Rather than waiting for developmental and operational test phases to conclude before operators start shaping employment, the EOU embeds warfighters early, side-by-side with vendors and the Air Dominance Combined Test Force. This enables operational concepts, requirements, and tactics to iterate while the hardware and software mature.
Epperson also noted that CCA pushes the Air Force away from platform-centric thinking toward a true “system of systems” approach. Multiple types of collaborative aircraft, with different mission sets, team with manned fighters to generate mass and create dilemmas for an adversary. The Warfare Center, he said, is wargaming force mixes and employment concepts now so they can be validated through testing and exercises as capability grows.
Flight Autonomy Vs. Mission Autonomy: Modularity Matters
One of the panel’s clearest explanations focused on autonomy itself. Helfrich distinguished between flight autonomy—tightly coupled, safety-critical software that keeps an aircraft flying safely—and mission autonomy, which directs the aircraft as a weapon system executing tactics and mission tasks.
That split is central to speed. With government reference architectures and modular design, the Air Force can iterate on mission autonomy rapidly—updating, improving, and even swapping it—without touching safety-critical flight controls. Jensen said the EOU treats mission autonomy like training a student pilot: master basic behaviors first (hold position, traffic avoidance), then build toward complex tactical performance.
Sets & Reps For Trust, The Real Requirement
Trust underpinned nearly every technical discussion. The panel argued that trust must extend beyond the pilot “quarterbacking” CCAs to everyone sharing airspace with them. In a large strike package, every flight must trust the autonomous aircraft to stay in assigned blocks, maintain separation and behave predictably.
That kind of trust is built through repetition, “sets and reps,” where the pilots can see autonomy do what it’s supposed to do, over and over, across various scenarios. Early pilot feedback demanded maximum information and granular control. As pilots gained confidence through repetition, especially in virtual environments, they demanded less data which reduced overall cognitive burden. The goal is an interface intuitive enough that it feels like using a smartphone, not like learning a new aircraft manual.
Continuous Software Updates: MVP On Day One, Then Improve
Will mission autonomy be “ready” on day one? The Air Force, Helfrich said, is building a model to deliver an initial minimum viable product (MVP), which it will then improve continuously based on operator feedback. Updated contracting incentives and modular architectures are now designed to support rapid iteration.
Jensen described the operational aspiration: fly, critique, send fixes and show up the next day with an improved software load. Helfrich said demonstrations have proven that modular swaps can happen quickly. Day one capability is only the starting point.
Fielding And The Enterprise: Squadrons, Basing, Planning, Debrief
As the discussion turned to fielding, Jensen described early work on force structure, including the concept of “autonomous fighter squadrons,” and the challenge of determining what mix of existing personnel specialties can support operations versus what might require new training pipelines.
Epperson and Helfrich emphasized flexibility in basing and employment. CCAs may not need to take off and land where manned aircraft do. This can enable commanders to distribute them across an area of responsibility and rendezvous airborne, an extension of distributed operations the Air Force is already pursuing.
Mission planning and debrief will be essential parts of the weapon system. Jensen said traditional engineer-centric post-flight analysis won’t work operationally. Autonomy must be debriefable in “pilot language,” meaning it will need to explain what it did and why. Efforts are underway with external partners to build debrief tools that can deliver such transparency and accelerate learning.
A Template, Not Just A Program
In closing, Jensen described the EOU’s culture as “force integration left.” Bring warfighters into development early so the enterprise is ready when aircraft arrive. Helfrich called CCA “acquisition transformation in action,” enabled by sustained partnership among operators, acquisition professionals, industry and the science-and-technology community. Epperson framed the broader goal as bridging “industry innovation into warfighter reality.”
Clearly, Increment One CCA has moved well beyond conceptual. The real test now is whether the Air Force can scale this new kind of capability involving modular autonomy, rapid updates, trusted teaming and enterprise-wide integration…well before the next fight demands it.