US Defence firm Northrop Grumman conducted flight tests showing multiple AI systems can seamlessly transfer control of a single aircraft mid-flight. The trials involved three AI firms – Shield AI, Accelint, and Applied Intuition. The tests are part of Northrop’s Talon IQ initiative, formerly known as Beacon.

A manned Vanguard Model 437 demonstrator was converted into a test platform integrating Northrop’s Prism autonomy system alongside partner AI software for evaluation. “Talon IQ’s open, modular testbed accelerates autonomy software development, delivering scalable, mission-ready solutions at lower risk and cost,” said Northrop Grumman in a statement.

In April, Northrop Grumman released a concept video revealing its proposed F/A-XX next-generation carrier fighter for the US Navy.

Modular flight autonomy

Northrop Grumman has advanced its push into autonomous aviation with a series of flight demonstrations under its Talon IQ testbed. It highlighted the maturity and flexibility of its Prism Mission Autonomy software.

According to the firm, recent trials showed that autonomy systems from multiple partners can be dynamically integrated during flight and swapped during flight without disrupting performance. This marks progress toward modular, mission-ready autonomous ecosystems for future aircraft.

During a recent test done on April 17, the aircraft initially operated under Northrop’s Prism autonomy system before transitioning mid-flight to software developed by Applied Intuition and Accelint. The transition occurred seamlessly, with each system taking control at the level of individual mission “skills,” demonstrating that autonomy can be shared and reconfigured in real time. The aircraft maintained stable and continuous operation throughout, underscoring the reliability of the underlying architecture.

Grumman highlights that the capability is enabled by a plug-and-play design that allows third-party autonomy software to integrate into the Talon IQ ecosystem while aligning with US government reference frameworks. The approach allows both specialized and general-purpose AI systems to operate within the same platform, offering the flexibility to tailor capabilities for different mission needs. By enabling autonomy at the skill level, the system supports rapid switching between tasks such as navigation, coordination, and mission execution.

Scalable flight AI

The Talon IQ platform is built on the Model 437 Vanguard, a manned demonstrator converted into an autonomous testbed. The setup provides a lower-cost and lower-risk environment compared to fully uncrewed systems, allowing companies to test and refine their autonomy software in real-world flight conditions. The platform forms part of Northrop’s broader Project Talon portfolio, which focuses on accelerating the development of next-generation autonomous capabilities.

Prism Mission Autonomy software serves as the core of this ecosystem, offering the speed and adaptability needed for complex operations. It supports coordination between crewed and uncrewed assets, dynamic route planning, and real-time deconfliction with other vehicles. These capabilities are critical for future combat and operational scenarios where multiple autonomous systems must operate together in contested environments.

According to Grumman, by opening the Talon IQ ecosystem to external partners, it is fostering a collaborative development model aimed at accelerating innovation. The flight-proven platform enables rapid iteration, validation, and deployment of mission autonomy solutions, positioning Talon IQ as a key stepping stone toward the company’s next-generation Talon Blue platform and broader autonomous aviation goals.