Autonomous swarm defense

Named industrial players include Saab, Boeing Company, BAE Systems, Northrop Grumman and Swarm Aero. Saab is associated with distributed coordination in the organization data, while Swarm Aero reported command-and-control software exercises at U.S. Navy Fleet Experimentation in June 2026. These examples span platform providers, defense contractors and software developers; the available data does not establish that every named company sells a complete swarm system.

National University of Defense Technology, Nanjing University of Aeronautics and Astronautics, and Northwestern Polytechnical University are active research groups, with 25, 16 and 13 documents, respectively, in the recorded window. Their work includes spectrum optimization, task allocation, reinforcement learning and swarm air-combat methods. Beihang University also published recent work on role-aware representations and uncertainty-driven policy regularization.

The recorded funding includes $4,069,731 across 10 grants since 2020 and $252,890,199 across 2 funding rounds. The round entries name University of Exeter and Aalto University, so they may not represent defense-specific investment. Treat these totals as recorded funding signals, not a verified measure of capital directed solely to autonomous swarm defense.

H2 Clipper, Inc. is the named leading patent holder, with 1 patent. The data records 3 patents in total and 1 patent in the 24-month window, but it does not provide a patent-growth rate. This limited patent count contrasts with the much larger publication activity, so the available figures do not show a clear upward patent trajectory.

The field remains research-heavy: FACTS records 190 publications in the last year, up 18%, but only 1 patent in the 24-month window. Operational demonstrations provide a sign of progress; Auterion reported a live multi-manufacturer combat-drone swarm demonstration in December 2025. However, published activity and demonstrations do not by themselves establish routine deployment or proven performance in combat.

Recent developments include Beihang University’s role-aware reinforcement-learning method for UAV swarm air combat, published in October 2026, and the National University of Defense Technology’s spectrum-optimization work, published in September 2026. A September 2026 study from Nanjing University of Aeronautics and Astronautics proposed software-defined resilient networking for topology reconstruction. Swarm Aero also reported command-and-control exercises with crewed aircraft in June 2026.

The United States leads the recorded country data with 1,395 documents and a 54.3% share, followed by China with 814 documents and a 31.7% share. The United Kingdom is also prominent, with 478 documents and an 18.6% share. Country shares may overlap, so they should not be read as mutually exclusive portions of a single total.

Recorded applications include counter-drone defense with 139 documents, intelligence and surveillance with 84, coordinated strike with 63, maritime operations with 55, and force protection with 27. Recent activity rose in several areas: coordinated strike increased 115% and intelligence and surveillance 47.1% in the comparison period. These categories describe the application taxonomy and do not establish operational effectiveness.

The five approaches are centralized orchestration, distributed coordination, learning-based coordination, rule-based autonomy and resilient swarm networking. Centralized systems simplify oversight but depend on a central controller; distributed systems reduce that dependence but complicate global coordination. Learning-based methods adapt to conditions but can be hard to validate, while rule-based systems are predictable but less flexible. Resilient networking focuses on keeping coordination links usable.