DARPA Flies AI-Controlled F-16 Converted From the Operational Fleet

DARPA Flies AI-Controlled F-16 Converted From the Fleet

An AI agent has already flown a fighter jet in a dogfight. That was DARPA's X-62A VISTA, a one-of-a-kind testbed built under the Air Combat Evolution program to prove the concept. This month at Eglin Air Force Base, DARPA and the U.S. Air Force did something with a different scale of ambition: they took a standard F-16 from the operational fleet, modified it with a bolt-on autonomy kit, and flew it under AI control 1.

The distinction matters. A purpose-built test platform tells you whether a technology can work. A conversion process applied to aircraft already in the inventory tells you whether you can field it broadly. DARPA described the flights as proof that standard operational aircraft can be converted for AI use, not merely tested in a lab 1.

The modification is called the VENOM Autonomy Kit, or VAK. It interfaces with the F-16's existing flight controls and mission systems through a design that lets a pilot toggle between human and AI control mid-flight using a switch. Multiple performers built and integrated the kit under DARPA's ACE program. A human pilot remains in the cockpit to monitor the AI agent throughout 1.

Brig. Gen. James "Fangs" Valpiani, the DARPA program manager, said the team "automated flight controls and sensors on a standard F-16 without changing the jet's core software" 1.

That sentence is the story. No core software changes. No platform redesign. A kit that bolts onto production aircraft. For anyone building or investing in defense technology, the implication is concrete: military AI is becoming a software-upgrade problem, not a hardware-procurement problem.

The VENOM aircraft will anchor the next phase under DARPA's Artificial Intelligence Reinforcements program, known as AIR. That program will run multiple AI agents through live-flight testing on the VENOM fleet. The stated goal is to prepare human pilots to coordinate with autonomous, uncrewed aircraft in combat, feeding into the Air Force's Collaborative Combat Aircraft initiative 1.

Lt. Col. Patrick "Dice" Highland, who takes over as AIR program manager this month, said the flights offer an early look at how AI agents could actively reshape air warfare 1.

Valpiani was measured about what remains unsolved. He called the aerial combat threat environment increasingly complex and acknowledged that hard questions persist about combat AI performance and trustworthiness under the conditions of real warfare 1.

Terry Wilson, who directs AI Development and Transition for the Air Force Research Laboratory's Autonomy Capability Team and serves as deputy program manager for both VENOM and AIR, described the milestone as moving these capabilities from simulation into actual flight 1.

The shift from a single purpose-built testbed to a fleet-convertible platform changes the economics of military AI development. When every experiment required a one-of-a-kind aircraft, testing was a bottleneck: one plane, one mission at a time. The VENOM approach turns converted F-16s into reusable airborne testbeds. Multiple aircraft running multiple agents in live flight, iterated on a schedule.

DARPA calls this an efficient pipeline for rapid AI development 1.

For the defense industry, the structural implication follows from the kit design. If autonomy is a bolt-on layer rather than a clean-sheet platform, the companies that develop the strongest AI agents are positioned to capture the next layer of value in aerial combat. The airframes are already fielded. The interface is proven. What iterates fastest is the software, and software moves on a timeline that traditional defense procurement was never built to match.