
Lockheed Martin and the U.S. Navy have successfully demonstrated an artificial intelligence-enabled anti-submarine warfare (ASW) system during the Rim of the Pacific (RIMPAC) 2026 exercise, showcasing a rapidly retrainable AI “sonar expert” capable of identifying new underwater threats in minutes.
The demonstration featured SensorMAX, an AI and machine learning-based system that analyzes subsurface acoustic data collected by sonobuoys to assist MH-60R Seahawk helicopter crews in detecting and tracking submarines. The company said the technology is designed to improve situational awareness by continuously monitoring multiple acoustic sources and rapidly adapting to emerging threats.
During the exercise off Hawaii, two MH-60R helicopters equipped with SensorMAX deployed sonobuoys that transmitted underwater acoustic data to the AI system. SensorMAX monitored the incoming information, helping aircrews identify potential submarine contacts while simultaneously supporting up to eight sonobuoy surveillance feeds—twice the capacity of legacy systems, according to Lockheed Martin.
One of the key objectives of the demonstration was to validate the system’s ability to rapidly update its threat-recognition models. When operators identified a previously unknown acoustic signature, they labeled the data, retrained the AI model in less than five minutes and transmitted the updated software over an encrypted data link back to the aircraft and the wider fleet.
The company described the process as a continuous “capture, label, retrain and deploy” workflow that enables AI models to evolve during ongoing operations without interrupting missions.
Lockheed Martin said the successful demonstration validated the concept of combining AI-enabled acoustic analysis with crewed anti-submarine warfare platforms, supporting the U.S. Navy’s efforts to expand the use of artificial intelligence and crewed-uncrewed teaming in future maritime operations.
According to the company, SensorMAX is platform-agnostic and can process data from multiple types of acoustic sensors across air, sea and land domains. The lightweight software architecture also allows AI models and mission software to be updated over the air without requiring aircraft to return to base.
The demonstration was supported by Lockheed Martin’s collaboration with software company FUSE Inc., which provides secure over-the-air software delivery capabilities for deployed platforms.
Lockheed Martin said the ability to rapidly update AI classifiers during flight could significantly shorten the detection-to-engagement timeline against evolving submarine threats while expanding the protective coverage around naval task groups and other high-value assets.