USAF Awards CCA Production: General Atomics & Anduril Build AI Wingmen
The US Air Force has made one of the most consequential decisions in modern military aviation history. In June 2026, the service awarded production contracts for its Collaborative Combat Aircraft (CCA) program to two companies: General Atomics, building the FQ-42, and Anduril Industries, building the FQ-44. These aren’t just new drones — they’re autonomous AI wingmen designed to fly alongside human pilots, absorb risk, and multiply combat power in ways that were science fiction just a decade ago.
What makes this announcement even more remarkable is the speed at which it happened. The contracts arrived months ahead of schedule, and Anduril claims this represents the fastest path from prototype to production for a fighter aircraft in over 50 years. The US Air Force isn’t just buying new hardware — it’s signaling a fundamental transformation in how America intends to dominate the skies in an era of peer-to-peer conflict.
This piece breaks down everything that matters: who won, what they’re building, how the AI actually works, and why the Pentagon’s unconventional “software sold separately” strategy could reshape defense procurement for generations. Whether you’ve followed the CCA program since its inception or you’re just catching up, the details here are genuinely extraordinary.
—
What Are Collaborative Combat Aircraft — And Why Do They Matter?
Collaborative Combat Aircraft are semi-autonomous, uncrewed platforms designed to operate in coordinated formations with manned fighters. Think of them as force multipliers — wingmen that never tire, never panic, and can be sent into the most dangerous corners of a contested airspace without risking a human life.
The concept addresses a core tension in modern air warfare. Advanced adversary air defense systems have made penetrating contested airspace extraordinarily costly. Rather than sending human pilots on the most dangerous ingress routes, CCAs can scout ahead, suppress enemy defenses, carry additional weapons, or simply saturate an adversary’s targeting systems with more threats than they can handle simultaneously.
Key Capabilities of Increment 1
The first increment of CCA is built around a clear set of performance requirements:
– 700 nautical mile combat radius — giving them reach comparable to manned tactical aircraft
– Air superiority mission focus — built to survive and fight in contested environments, not just loiter over battlefields
– Semi-autonomous operation — capable of executing complex mission tasks with minimal human input
Secretary of the Air Force Troy Meink has stated the goal is fielding over 150 combat-capable CCAs by the end of the decade. That’s not a research program — that’s a warfighting capability with a hard deadline attached to it.
—
The Hardware Winners: FQ-42 and FQ-44
The April 2024 design, build, and test contracts set the competition in motion. Two years later, General Atomics and Anduril emerged as the production winners from what was a highly competitive field.
General Atomics and the FQ-42
General Atomics built its reputation on programs like the MQ-1 Predator and MQ-9 Reaper — uncrewed systems that redefined how the US military conducts persistent surveillance and strike operations. The FQ-42 represents the company’s evolution into the high-end combat arena, moving from the relatively permissive airspace of counterterrorism operations into the contested skies of great power competition.
General Atomics brings decades of unmanned systems integration experience to the FQ-42, and its existing relationships with the Air Force’s logistics and sustainment infrastructure give it a credible path to the kind of scale the service is demanding.
Anduril and the FQ-44
Anduril is a different kind of defense company — founded in 2017 by Palmer Luckey, it operates on a software-first philosophy that has disrupted traditional acquisition timelines. The FQ-44 reportedly reflects that philosophy: a platform designed from the ground up to be upgraded rapidly as autonomous software capabilities evolve.
The production numbers Anduril is advertising are striking. The company claims it can deliver up to 150 FQ-44 aircraft per year — a production rate that, if sustained, would allow the Air Force to build serious fleet depth quickly. For context, that’s roughly the annual production rate the US has achieved for the F-35 across its entire global customer base, achieved here for a single aircraft type from a single facility.
—
The “Software Sold Separately” Strategy: A Procurement Revolution
Here’s where the CCA program departs most sharply from how the Pentagon has traditionally bought weapons systems — and it’s arguably the most strategically significant aspect of the entire program.
The Air Force has deliberately decoupled the mission autonomy software from the airframe procurement. You buy the drone. The brain that flies it comes from a separate, competing contract.
Why This Matters
Traditional defense procurement tends to lock a service into a prime contractor’s entire ecosystem. You buy an F-35, you’re buying Lockheed Martin’s avionics, Lockheed Martin’s software updates, Lockheed Martin’s upgrade roadmap. The leverage sits with the contractor.
The “software sold separately” model breaks that dependency entirely. It creates a persistent competition among software vendors, which produces several concrete benefits:
– Faster iteration cycles — Software companies can push updates without waiting for hardware contract modifications
– Genuine price competition — Multiple vendors bidding keeps costs from inflating unchecked
– Resilience — If one software vendor’s approach proves inadequate or is compromised, the Air Force can switch without scrapping the entire airframe
– Specialization — Companies can focus on what they’re best at, whether that’s building airframes or writing autonomous decision-making code
The Six Software Contenders
The Air Force has established a pool of six companies competing for mission autonomy software contracts:
1. Anduril Industries
2. General Atomics
3. Lockheed Martin
4. Northrop Grumman
5. RTX Collins Aerospace
6. Shield AI
Note that Anduril and General Atomics appear on both lists — they’re competing as hardware providers and software providers simultaneously. This creates a fascinating dynamic where a company could theoretically fly its own airframe with its own software, or see a competitor’s software running on its hardware.
Shield AI, which has built its reputation specifically on autonomous flight software with its “Hivemind” pilot AI, is a particularly interesting contender. The company has no airframe to protect in this competition — it’s purely focused on winning the software fight.
—
Cost, Funding, and the Numbers Behind the Program
The financial architecture of the CCA program reflects the Air Force’s awareness that it needs mass — not just capability — to deter and defeat peer adversaries.
The Unit Cost Target
The Air Force is targeting a unit cost of roughly one-third the price of an F-35A. The Lot 17 F-35A was priced at approximately $82.5 million, which puts the CCA target at under $30 million per aircraft. That’s not cheap by commercial standards, but in the context of tactical aviation, it’s transformative.
At $30 million per unit, the Air Force can field three CCAs for every manned fighter it would otherwise buy. In a conflict where attrition is a real factor — and where losing a $80+ million aircraft means losing years of pilot training — that math matters enormously.
The FY27 Funding Request
The budget picture for the first production lot is substantial:
– $1.4 billion requested for CCA development in FY27
– Nearly $1 billion requested for CCA procurement in FY27
– First production lot expected to execute with appropriated FY27 funding
This is real money committed to a real program — not a research line item dressed up as a capability.
—
How AI Wingmen Actually Work in Combat
The “loyal wingman” concept sounds intuitive, but the operational mechanics deserve a closer look. These aren’t remotely piloted vehicles in the traditional sense, where a human operator is joysticking the drone in real time. The latency alone would make that unworkable in a high-speed, contested environment.
Instead, CCAs are designed to operate with a significant degree of autonomous decision-making authority within mission parameters set by a human commander. A manned pilot flying an F-35 or the new F-47 Next Generation Air Dominance platform could theoretically manage a flight of CCA wingmen the way a quarterback manages receivers — setting plays, adjusting assignments, and making high-level calls while the individual players execute autonomously.
The Manned-Unmanned Team
In practical terms, this manned-unmanned teaming concept could work across several mission profiles:
– Suppression of Enemy Air Defenses (SEAD): CCAs could prosecute radar and missile battery targets while manned aircraft remain outside the most lethal threat envelopes
– Weapons carriage: A single manned F-35 accompanied by two CCAs effectively triples its weapons payload without tripling the number of pilots at risk
– Deception and saturation: Multiple CCAs appearing on enemy radar force threat prioritization decisions that open windows for manned aircraft to exploit
The AI autonomy software is what makes all of this possible. The six competing software vendors aren’t just writing autopilots — they’re building tactical decision-making systems capable of responding to dynamic combat environments, identifying threats, choosing engagement solutions, and coordinating behavior across a distributed formation.
—
The Road Ahead: Challenges and Increment 2
The optimism surrounding the CCA program is well-earned, but the harder problems are still ahead.
Increment 1 vs. Increment 2
The current production awards cover Increment 1 — a defined set of capabilities centered on the air superiority mission with a 700 nautical mile combat radius. Increment 2 is already in its early stages, with nine vendors engaged in what will likely be a more complex capability set. The specific requirements for Increment 2 haven’t been fully publicly defined, but the implication is expanded missions, enhanced autonomy, and potentially different performance envelopes.
The Challenges the Air Force Must Navigate
Several significant hurdles remain:
Cybersecurity: An autonomous aircraft that can be spoofed, jammed, or hacked represents a liability rather than an asset. The Air Force must ensure the autonomy software is hardened against adversary electronic warfare and cyber intrusion at a level that manned aircraft simply haven’t required.
Rules of engagement and AI ethics: As these systems gain autonomy, defining where human decision-making authority must be preserved — particularly around lethal force decisions — is both a legal and ethical imperative. International laws of armed conflict don’t map cleanly onto fully autonomous weapons systems.
Integration complexity: Getting FQ-42s and FQ-44s to communicate seamlessly with F-35s, F-47s, and each other across a contested electromagnetic spectrum is an engineering challenge of the first order.
Sustainment at scale: The Air Force has a mixed record with sustainment costs on complex programs. Keeping 150+ CCAs mission-ready will require supply chains, maintenance concepts, and logistics infrastructure that don’t yet fully exist.
The Long-Term Vision
The CCA program is best understood not as a product but as a platform — a foundation on which the Air Force will layer increasingly capable autonomous systems over time. As autonomy software matures through the competitive process, the same airframes could become dramatically more capable without requiring new production contracts. That’s the promise of the “software sold separately” model playing out over a decade-long horizon.
—
Conclusion: A New Era of Air Combat Takes Shape
The USAF’s decision to award CCA production contracts to General Atomics and Anduril — producing the FQ-42 and FQ-44 respectively — is more than a procurement milestone. It’s the opening chapter of a fundamentally different kind of air warfare, one built around human-machine teaming, distributed combat power, and software-driven adaptability.
The key takeaways from this program are significant enough to make any list of transformative defense decisions in the past half-century:
– Two proven companies will deliver AI wingmen at scale, with Anduril claiming capacity for up to 150 FQ-44s annually
– The “software sold separately” model creates persistent competition among six vendors, including Shield AI and Lockheed Martin, preventing stagnation
– The $30 million cost target and $2.4 billion in FY27 funding requests signal serious institutional commitment
– Over 150 combat-capable CCAs by the end of the decade is the immediate strategic goal, feeding into the broader Next Generation Air Dominance architecture
The hardest work — proving autonomous AI can perform reliably in a lethal, contested, electromagnetically denied combat environment — still lies ahead. But the fact that these contracts arrived months ahead of schedule, using what Anduril calls the fastest fighter development timeline in 50 years, suggests the Air Force has found a procurement approach that actually works.
The era of AI wingmen isn’t coming. It’s already in production.
—
Frequently Asked Questions
What is a Collaborative Combat Aircraft (CCA)?
A CCA is a semi-autonomous, uncrewed aircraft designed to fly in coordinated formations alongside manned fighter jets. It operates with AI-driven autonomy to complete mission tasks, absorb risk, carry weapons, and multiply the combat effectiveness of manned platforms like the F-35.
Who won the CCA production contracts?
General Atomics won the contract to produce the FQ-42, and Anduril Industries won the contract to produce the FQ-44. Both contracts were awarded in June 2026, months ahead of the originally planned schedule.
What does “software sold separately” mean in this context?
The Air Force deliberately separated the mission autonomy software from the airframe purchase. Six companies — Anduril, General Atomics, Lockheed Martin, Northrop Grumman, RTX Collins Aerospace, and Shield AI — compete independently for contracts to supply the AI software that will control the CCAs, regardless of which company built the airframe.
How much do CCA drones cost?
The Air Force is targeting a unit cost of under $30 million — roughly one-third the price of an F-35A, which was approximately $82.5 million per aircraft in its most recent production lot.
How many CCAs does the Air Force plan to field?
Secretary of the Air Force Troy Meink stated the goal is to field over 150 combat-capable CCAs by the end of the decade, with the first production lot funded through FY27 budget requests totaling approximately $2.4 billion.
How will CCA drones work with manned aircraft?
CCAs operate under the command authority of manned pilots flying platforms like the F-35 or the F-47 NGAD aircraft. Rather than being remotely piloted in real time, they use AI autonomy software to execute mission tasks — such as weapons delivery, threat suppression, or electronic warfare — within parameters set by the human commander.