An artificial intelligence agent demonstrated advanced combat capabilities by flying a modified F-16 aircraft through 27 intercepts against a live target, according to AeroTime. Announced by Lockheed Martin on August 4, 2026, the series of eight sorties utilized real-time data from an operational sensor to maneuver the aircraft into tactical intercept positions. The testing was conducted at Edwards Air Force Base in California through a collaboration between Lockheed Martin Skunk Works, the US Air Force Test Pilot School, and industry partners.
Technical Testing Data
The campaign featured the X-62 Variable In-flight Simulation Test Aircraft (VISTA), a modified F-16D Block 30 platform equipped with Block 40 avionics. The aircraft carried a Lockheed Martin Legion Pod, which utilizes an IRST21 sensor to passively detect targets via heat signatures. Unlike previous AI tests that relied on simulated data, this series required the AI agent to process live infrared search and track feeds to execute maneuvers in real time.
| Feature | Specification |
|---|---|
| Aircraft Model | F-16D Block 30 (X-62 VISTA) |
| Mission Count | 8 Sorties |
| Intercept Count | 27 |
| Target Aircraft | T-38 |
| Sensor System | Legion Pod (IRST21) |
Lockheed Martin highlighted the effectiveness of its “Supermassive” agent generation capability, which facilitated the integration and ground testing of the AI software with the X-62 in a three-month timeframe. Ron Fehlen, vice president and general manager at Lockheed Martin Skunk Works, noted that the flights proved the system’s ability to close the sensor-to-action loop in an operational environment.
Operational Context
The X-62, part of the 412th Test Wing, has a history of pioneering autonomous flight, including its participation in DARPA’s Air Combat Evolution program and the first AI-versus-human dogfights in September 2023. Following a redesignation from NF-16D in June 2021, the aircraft underwent a Mission Systems Upgrade in December 2025. This upgrade, funded by the Pentagon’s Test Resource Management Center, is intended to support future integration of combat systems and sensors within a mesh network.
Why It Matters
This milestone signifies a transition from controlled, simulated environments to complex, dynamic operational scenarios for autonomous systems. By integrating live sensor feeds into the decision-making loop of a legacy platform, the defense sector is validating the feasibility of "loyal wingman" concepts. This advancement reduces the lead time for fielding autonomous capabilities and suggests that future combat aircraft may prioritize modular, AI-ready avionics architectures. As the US Air Force moves forward with the Collaborative Combat Aircraft program, these tests provide a empirical baseline for how AI agents will handle the uncertainty of modern aerial engagement.
Reader Discussion & Insights