Lockheed Martin Integrated AI Into X-62 VISTA Flight Tests

The X-62 VISTA completed 27 AI-directed intercepts, testing autonomous capabilities for future combat aircraft.

Updated on Oct. 5, 2026 in Artificial Intelligence

Bold flat-color editorial illustration showing a stylized modular sensor pod, representing military AI integration systems. — Lockheed Martin successfully integrated the Supermassive AI framework into its X-62 VISTA research aircraft to conduct 27 autonomous intercepts. AI Illustration. Upload story photo >

Live Poll

Do you trust the use of artificial intelligence in military defense systems?

On August 4, 2026, the X-62 VISTA conducted 27 AI-directed intercepts of a T-38 target across eight separate flights at Edwards Air Force Base. This research-stage development utilized the Supermassive framework to integrate autonomous flight agents.

Why it matters

The Defense Department is prioritizing the development of collaborative combat aircraft and advanced human-machine teaming. These flight tests provide data on the integration of large language models into real-time airborne sensor processing.

The X-62 VISTA performed 27 intercepts using real-time data from the Legion Pod infrared sensor. As of October 2026, Skunk Works engineering teams have integrated 55 large language models to support these autonomous operations.

The players

Lockheed Martin

An aerospace and defense contractor known for the F-35 Lightning II and advanced prototyping through its Skunk Works division.

OpenAI

A research organization focused on developing large language models and artificial intelligence safety.

The details

The Supermassive framework acts as the integration and testing architecture for the X-62 AI agent. It processes inputs from the Legion Pod — an external, modular infrared search-and-track sensor — to execute intercept maneuvers. The engineering process also relied on wind-tunnel testing completed in under two months for the upcoming Vectis platform design.

August 4, 2026: X-62 VISTA conducted AI-directed intercepts.
September 14, 2026: Lockheed announced plans for four Vectis prototypes.
October 2026: Skunk Works utilized 55 large language models.
End of 2027: Expected first flight of Vectis prototype.

The Tech Race

This effort advances the integration of large language models into autonomous systems to support the Department of Defense's Collaborative Combat Aircraft program. The test represents a shift toward field-testing AI agents in high-speed, multi-sensor intercept scenarios.

These results currently inform military engineering research rather than commercial or civilian aviation workflows. Progress toward the 2027 Vectis flight will indicate the pace at which these autonomous frameworks move from experimental research to flight-ready prototypes.

The takeaway

The successful integration of 55 large language models into the X-62 testing cycle highlights a move toward LLM-driven decision-making in high-speed tactical environments. Readers should watch for the scheduled 2027 flight of the first Vectis prototype as a key indicator of program maturity.