Lockheed Martin, Verizon and others showcase NetSense airspace protection using AI on existing 5G networks

Lockheed Martin said it has demonstrated a commercial technology approach to counter unmanned aerial systems, or UAS, by using artificial intelligence applied to radio indicates already carried over existing 5G networks. In a live test in the Miami area, the partners showed how their NetSense system could identify, track and monitor drones and then issue alerts while maintaining “track custody” through flight.

The companies described NetSense as an Airspace Awareness-as-a-Service offering, positioning it for customers that need persistent airspace monitoring but do not want to replace large parts of cellular infrastructure to do it. Lockheed said the technology combines warning and tracking systems that read radio-frequency, or RF, disturbances, use AI to predict and track a UAS flight path, and automatically send real-time alerts.

During the July demonstration, the partners used a Verizon 5G network spectrum as the sensing foundation. Lockheed said ODC contributed AI-native Radio Access Network software, NVIDIA provided an AI Aerial platform, and Keysight supplied an RF simulation platform. The companies said the setup was able to detect drones and alert the team in real time in a “critical, high-traffic, urban area.”

Lockheed’s release framed the approach as a way to achieve at-scale sensing without “recapitalization” of cellular infrastructure, and said it is intended to be available ahead of the timeline for 6G Integrated Sensing and Communication, or ISAC. ISAC is the concept of blending sensing and communications in next-generation networks. Lockheed said the NetSense algorithms are also intended to be complementary to 6G ISAC when those systems become available.

The company said NetSense can be delivered as a subscription that interfaces with a customer’s existing security operations. It also said pilot deployments are planned for the second half of 2026 and early 2027, with commercial general availability planned for 2027. Early deployments, the release added, will be limited to select customers with urgent need for the capability, and interested organizations can request briefings via Astris AI, a Lockheed wholly owned subsidiary.

Lockheed emphasized that its plan relies on using established 5G network spectrum, rather than changing existing cellular deployments. “As drones become more affordable and accessible, airspace awareness technology needs to evolve rapidly to outpace potential threats,” said Sarah Hiza, senior vice president for technology and strategic innovation at Lockheed. She said the company built NetSense to be deployable “at the time of need” by collaborating with the commercial technology industry within the established 5G environment.

Verizon executives said the key idea is applying AI to the RF announcement measurement data already moving across its infrastructure. “By applying AI to the RF announcement measurement data already traversing our 5G infrastructure, we can effectively turn our network into an intelligent sensing platform,” said Anil Guntupali, Verizon’s senior vice president of technology and product development. NVIDIA, meanwhile, said its components are meant to bring AI and accelerated computing toward the network edge so RF announcement disturbances can be analyzed quickly for public safety use cases.

Keysight highlighted what it called an engineering workflow spanning RF modeling, simulation, digital twin, emulation and test, saying the collaboration is meant to help customers prove feasibility, reduce complexity and bring mission-critical technologies to market faster. ODC’s chairman of the board of directors, Shaygan Kheradpir, described the effort as a milestone in counter-UAS and layered homeland air defense for low-flying objects, saying ODC’s AI-Native, Zero Trust AI-RAN platform enables integrated sensing and communications through a standards-compliant commercial architecture.

The partners did not disclose detailed performance metrics from the Miami test, such as detection range, false-alarm rates, classification accuracy, or latency. The release also did not specify the number or type of drones used in the demonstration, nor did it describe how the alerting system integrates with customer security operations beyond the subscription and “real-time alerts” framing. As pilots approach, those omissions may become important for evaluating how the solution performs across different environments, RF conditions, and regulatory constraints.

For Lockheed and its partners, the next step is piloting the subscription model with select customers and testing how well the approach scales from a controlled demonstration to ongoing operations at high-risk sites. Watch for additional detail on pilot customer selection, integration requirements with existing security systems, and any published benchmarks tied to detection, tracking, and alert responsiveness. If the concept delivers on its promise of sensing without changes to cellular deployments, it could also accelerate interest in layered counter-drone architectures built around commercial telecommunications infrastructure.