On May 2, 2024, Air Force Secretary Frank Kendall buckled into the rear seat of a modified F-16 at Edwards Air Force Base in California, folded his hands in his lap, and let a machine do the flying. The X-62A VISTA climbed, banked, and maneuvered through a test profile over the Mojave Desert with no human hand on the stick. Algorithms developed by the Johns Hopkins University Applied Physics Laboratory controlled the jet from takeoff through landing, making Kendall the first Cabinet-rank defense official to ride in a fighter aircraft piloted entirely by artificial intelligence.
The flight was not a stunt. It was the most visible milestone yet in a Pentagon campaign to field a new class of autonomous warplanes that would fly alongside crewed fighters in combat. The Air Force calls the effort the Collaborative Combat Aircraft program, or CCA, and the service’s stated goal is to eventually put more than 1,000 AI-driven drones into operational squadrons.
From DARPA dogfights to a secretary in the seat
The X-62A VISTA has been the Air Force’s primary testbed for autonomous flight research since the early stages of the DARPA Air Combat Evolution program, known as ACE. Throughout 2023, the jet flew AI-versus-AI and AI-versus-human dogfight scenarios at Edwards, gradually expanding the envelope of what machine pilots could handle in a high-performance airframe. By the time Kendall flew, the system had already logged dozens of autonomous sorties.
What changed on May 2 was the stakes. Putting the service’s top civilian leader aboard was a deliberate signal of confidence. An official Defense Department photo page documents Kendall in the cockpit, timestamped and credited to photographer Richard Gonzales. The APL technical release describes the software that made it possible: a framework called the ACE Distributed Operations Manager, or ADOM, which coordinates multiple AI decision-making agents simultaneously aboard a single airframe. Rather than following a preset route, ADOM managed competing AI processes in real time, a capability the Air Force considers essential for future scenarios in which autonomous wingmen must react to threats without waiting for human commands.
The Collaborative Combat Aircraft program takes shape
The X-62A work feeds directly into the CCA program, which aims to pair relatively low-cost unmanned jets with crewed fighters like the F-35 and the forthcoming B-21 Raider. In April 2024, the Air Force awarded Increment 1 contracts to two companies: Anduril Industries and General Atomics Aeronautical Systems. Each firm is building a different CCA prototype designed to carry sensors, weapons, or electronic-warfare payloads while flying autonomously in formation with piloted aircraft.
A separate prototype, the XQ-67A Off-Board Sensing Station built by General Atomics, completed its first flight in February 2024, weeks before Kendall’s ride. That aircraft is intended to serve as a sensor node that extends the awareness of crewed fighters far beyond their own radar range. Together, these programs represent the industrial backbone of the Air Force’s push toward a mixed fleet of human and machine pilots.
Kendall himself has repeatedly cited the 1,000-drone figure in public remarks and budget testimony, framing CCA as a force-structure necessity rather than a technology experiment. The logic is straightforward: autonomous wingmen cost a fraction of a crewed fighter, can be risked in missions too dangerous for human pilots, and can multiply the combat power of a shrinking fleet of manned jets. Air Force budget documents for fiscal years 2025 and 2026 include dedicated CCA funding lines, though the service has not published a binding acquisition schedule or a specific year by which all 1,000 aircraft must be operational.
What the public still does not know
For all the symbolism of Kendall’s flight, several critical details remain outside the public record as of mid-2026. No primary Department of Defense flight logs or performance telemetry from the May 2 sortie have been released. Without that data, independent analysts cannot assess how the AI performed under specific conditions, what altitude or speed regimes it operated in, or whether any safety interlocks activated during the flight.
The safety architecture is only partially described. How quickly a human pilot could retake control, and under what failure conditions that handoff would trigger, is not spelled out in the APL release or the Defense Department’s documentation. The absence of that information does not mean safeguards were lacking; it means the public cannot independently evaluate them.
Cost is another open question. Building, certifying, and sustaining a fleet of AI-piloted combat drones at the scale the Air Force envisions would rank among the service’s largest procurement programs in decades. Early estimates from defense analysts have placed the per-unit cost of a CCA airframe in the range of $10 million to $25 million, a fraction of the roughly $80 million price tag for an F-35A, but the total program cost across more than 1,000 airframes, plus software development, ground infrastructure, and maintenance, has not been officially disclosed.
The strategic pressure behind the timeline
The urgency driving CCA is not abstract. Pentagon officials have pointed repeatedly to China’s own investments in autonomous combat aircraft and drone swarms as a primary motivation. The People’s Liberation Army Air Force has tested several unmanned combat air vehicle prototypes, and Chinese defense firms have displayed AI-enabled drone concepts at international arms shows. For U.S. military planners, the concern is that a future conflict in the Western Pacific could pit American pilots against adversaries who can afford to lose unmanned jets by the dozen.
That competitive pressure helps explain why the Air Force moved so quickly from DARPA-funded research to formal acquisition contracts. It also explains why Kendall chose to climb into the X-62A himself. “The secretary flew in it” is a trust indicator, not a technical benchmark, but in the bureaucratic politics of defense spending, a Cabinet official willing to strap into an AI-controlled jet carries real persuasive weight with lawmakers who control the budget.
Where the program goes from here
The next measurable milestones will move beyond a single airframe. The Air Force needs to demonstrate that ADOM or its successors can coordinate multiple unmanned aircraft flying in formation, not just multiple AI agents running on one jet. Anduril and General Atomics are expected to fly their Increment 1 prototypes through a series of increasingly complex test campaigns over the next several years, with initial operational capability targeted before the end of the decade.
Additional transparency about safety mechanisms, performance metrics, and acquisition timelines would allow Congress and the public to judge whether the 1,000-drone ambition is keeping pace with engineering reality. Until those demonstrations and disclosures arrive, Kendall’s ride in the X-62A VISTA stands as a verified, symbolically powerful early step: a real AI system flew a real fighter jet with a real senior official aboard, under controlled conditions, and brought him home safely. Whether that moment becomes the foundation of a new era in military aviation or a high-profile waypoint on a longer, harder road depends on what the Air Force can prove next.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
