Lockheed Martin Skunk Works, in partnership with the U.S. Air Force Test Pilot School (TPS) and industry partners, has successfully executed sensor-driven artificial intelligence (AI) fighter intercepts aboard the X-62 Variable In-flight Simulation Test Aircraft (VISTA). The test series validated the integration of real-time operational sensor data directly into an onboard AI agent’s decision-making loop to execute tactical manoeuvres against live airborne targets.

Conducted across eight test flights at Edwards Air Force Base, the X-62 VISTA completed 27 AI-controlled intercepts. Equipped with a Lockheed Martin Legion Pod, the aircraft tracked a live T-38 target jet, streaming classified Infrared Search and Track (IRST) data directly to the AI agent. The system processed the real-time target feeds to autonomously manoeuvre the fighter into tactical intercept positions, marking a transition from simulated target data to live operational sensor streams.

The integration leveraged Skunk Works’ “Supermassive” AI agent generation process, achieving full system integration and ground testing with the X-62 architecture in three months. By delegating complex targeting and manoeuvring tasks to autonomous software, defence engineers aim to reduce pilot cognitive workload, enhancing situational awareness and overall survivability in contested environments.

Looking ahead, the test architecture will support the X-62 Mission Systems Upgrade. The planned enhancement will enable the aircraft to test advanced combat systems, multi-domain sensors and autonomous AI agents operating within a next-generation mesh network framework.