Air Force and Lockheed Martin Achieve First Live AI-Driven Fighter Intercepts

The U.S. Air Force and Lockheed Martin have demonstrated an AI agent autonomously flying a fighter jet using live infrared sensor data. This milestone marks a critical transition from simulated environments to real-time, closed-loop combat cycles using operational hardware.

During a test campaign titled "HAVE HEAT" at Edwards Air Force Base, the X-62 Variable In-flight Simulation Test Aircraft (VISTA) performed 27 autonomous air intercepts against a live T-38 Talon target across eight flights. The AI agent relied on Lockheed Martin's Legion Pod, an infrared search and track (IRST) system. This hardware allows the aircraft to detect targets passively, avoiding the radar emissions that typically alert an adversary to a fighter's presence.

The Path to Autonomy

The AI agent was developed via Lockheed Martin Skunk Works' "Supermassive" platform, a specialized system for generating AI agents. Integration and ground testing were completed in just three months. To ensure operational safety, the Air Force conducted a parallel experiment known as "HAVE HOLIDAYS," which focused on the integration of modular software and hardware while testing safety constraints designed to prevent the AI from exceeding limits defined by human operators.

Implications for Air Superiority

This demonstration proves that AI can reliably close the "sensor-to-action" loop in real-time without human intervention. By delegating complex tactical calculations and intercept geometries to algorithms, the Air Force can significantly reduce pilot workload and increase situational bandwidth. This capability is essential for maintaining air superiority in high-threat, complex environments where the speed of decision-making is a primary advantage.

Ron Fehlen, vice president and general manager at Lockheed Martin Skunk Works, stated that the agents consumed classified infrared feeds to execute combat-critical maneuvers in real time, describing the achievement as a "decisive advance" toward AI-augmented air dominance. Lt. Col. Joshua Strafaccia, dean of faculty for research at the Air Force Test Pilot School, added that the tests represent a meaningful expansion toward the integrated, AI-driven control of multiple sensors and vehicles.

Future Outlook

The X-62 VISTA continues to serve as a "flying laboratory" for the U.S. Air Force, providing a low-risk environment to research data fusion and autonomous flight. These results are expected to inform the development of Collaborative Combat Aircraft (CCA) and the broader strategy of crewed-uncrewed teaming. While the current tests focused on passive infrared sensing, the Air Force continues to evaluate how these autonomous agents will integrate with broader mission systems in future operational deployments.