L3Harris and Shield AI Fused Autonomous Drones with Electronic Warfare — No Human in the Loop. The Spectrum Just Got a Mind of Its Own.
On March 11, 2026, L3Harris Technologies and Shield AI announced they had completed a first-of-its-kind demonstration: fully autonomous electronic warfare with no human in the decision loop.[1] The test fused L3Harris' Distributed Spectrum Collaboration and Operations (DiSCO™) electromagnetic battle management system with Shield AI's Hivemind mission-autonomy software — creating a system where unmanned aircraft detected, analyzed, and responded to electromagnetic threats entirely on their own.
This wasn't a single drone dodging a jammer. DiSCO collected and processed threat intelligence from multiple unmanned aerial systems simultaneously, building a shared real-time picture of the electromagnetic battlespace. Hivemind then used that fused picture to autonomously identify safe operating zones and execute tactical maneuvers — all without a human operator touching the controls.[1][2]
Electronic warfare moves at machine speed, and operational advantage depends on autonomy.
The integration pairs two systems designed for very different problems — and the combination creates something neither could achieve alone.
DiSCO™ (Distributed Spectrum Collaboration and Operations) is L3Harris' electromagnetic battle management ecosystem. Think of it as the nervous system: it ingests signals data from every sensor in the network — airborne, ground-based, sea-based — and fuses it into a single, coherent picture of who is transmitting what, where, and on which frequencies. It doesn't just detect threats; it builds a common operating picture of the entire electromagnetic spectrum in real time.[1]
Hivemind is Shield AI's mission-autonomy software — the brain. Originally designed to fly aircraft without human pilots, Hivemind doesn't follow preplanned routes. It senses, decides, and acts. It can reroute around threats, avoid obstacles, engage targets, and complete missions without human intervention. It's platform-agnostic, meaning it can run on drones, ground vehicles, or naval systems.[4]
When you plug the brain into the nervous system, you get something new: an autonomous entity that perceives the electromagnetic battlespace and makes tactical decisions within it at machine speed. DiSCO tells Hivemind what the spectrum looks like. Hivemind decides what to do about it. No human analyst sits between sensing and action.
The demonstration also integrated L3Harris' Green Wolf — an electronic warfare vehicle from the company's "Wolf Pack" family of launched effects, introduced in July 2025. Green Wolf carries electronic attack and Detect, Identify, Locate and Report (DILR) payloads. In the exercise, it served as both a jamming platform and a threat-detection node feeding data into DiSCO.[5] A software-defined radio payload provided electronic support measures, while a Shield AI UAS served as a communications relay — completing a multi-node autonomous EW mesh.[2]
Electronic warfare has always been a human-speed domain disguised as a machine-speed problem. Analysts monitor waveforms. Operators decide which threat to jam. Commanders approve electronic attack plans. The signals move at light speed, but the decisions move at the speed of a human reading a screen.
The adversary doesn't wait. Russia demonstrated in Ukraine that electronic warfare environments shift in seconds — frequencies hop, jammers reposition, new emitters appear. China's integrated air defense networks coordinate across dozens of radar types simultaneously. The human operator monitoring a scope isn't slow because they're bad at their job. They're slow because they're human.
The DiSCO-Hivemind integration eliminates that bottleneck. When a new emitter appears in contested airspace, the system doesn't alert an operator and wait for instructions. It fuses the detection into the common operating picture, recalculates safe maneuvering zones, and repositions platforms — all before a human could have processed the initial alert.
This is the same logic that drove Anduril's Pulsar system to use machine learning instead of threat libraries. But Pulsar is a jammer that learns autonomously. DiSCO-Hivemind is a multi-platform autonomous force that maneuvers autonomously within the electromagnetic environment. One teaches itself how to jam. The other teaches itself where to fly, what to attack, and how to survive — all based on what the spectrum is telling it in real time.
The distinction matters. Pulsar is a weapon system that fights autonomously in the spectrum. DiSCO-Hivemind is a command-and-control architecture that operates autonomously through the spectrum. The weapon gets smarter. The force gets autonomous.
Shield AI was founded in 2015 with a deceptively simple mission: fly drones into buildings so soldiers don't have to. Its first product, Nova, was a quadcopter that could navigate GPS-denied indoor environments — dark rooms, tunnels, underground bunkers — using onboard AI and no external communications.[4]
That capability — autonomous operation without GPS, without comms, without a pilot — turned out to be the seed of something much larger. Hivemind evolved from a building-clearing AI into a platform-agnostic autonomy stack that now flies fixed-wing aircraft, manages drone swarms, and coordinates multi-vehicle operations.
The trajectory in 2025-2026 alone has been remarkable. In August 2025, Hivemind first flew the Navy's BQM-177A subsonic target drone, validating advanced vehicle control laws. In December 2025, Hivemind autonomously piloted two BQM-177As simultaneously in a Live-Virtual-Constructive environment alongside a virtual F-18 and simulated adversary aircraft — executing coordinated defensive maneuvers to protect Combat Air Patrol positions.[6]
Then, in March 2026, Hivemind fused with DiSCO for the autonomous EW demonstration. The pattern is unmistakable: Shield AI is methodically proving that its AI pilot can handle increasingly complex military functions — from navigation, to multi-aircraft coordination, to electromagnetic warfare — without human intervention at each step.
By mid-2026, Shield AI intends to make its safety assurance framework (MM-RTA) available as part of the Hivemind Autonomy SDK, enabling third parties to integrate certified safety assurance into their own autonomous systems.[7] Hivemind isn't just flying Shield AI's own aircraft anymore. It's becoming the operating system for autonomous military platforms.
The DiSCO-Hivemind demo represents something the legacy defense primes have been forced to confront: the integration layer is becoming the product.
L3Harris — a $18B defense prime formed from the 2019 Harris/L3 Technologies merger — is one of the world's largest electronic warfare companies. It builds the systems on the EA-18G Growler, the ALQ-211 SIRFC, and the Navy's Next Generation Jammer. Yet for this demonstration, it partnered with a 10-year-old startup to provide the autonomy layer. L3Harris brought the sensors and the EW payloads. Shield AI brought the brain.
This is happening across the defense sector simultaneously. CACI International just won an $85 million contract for the Navy's Spectral electronic warfare system — a program valued at approximately $1.2 billion designed to give naval vessels electromagnetic awareness across radar, communications, and the full spectrum.[8] The Pentagon's $13.4 billion FY2026 budget request for autonomous weapons and systems includes significant electronic warfare integration requirements.[9]
The Wolf Pack family — Red Wolf for kinetic strikes, Green Wolf for electronic warfare — represents L3Harris' own evolution toward modular, autonomous launched effects. Introduced in July 2025, these small, attritable platforms are designed to be launched from ground vehicles, aircraft, or naval vessels in coordinated packs. The USMC has already conducted live-fire demonstrations with Red Wolf.[5]
The message from industry is unanimous: the future of electronic warfare is autonomous, distributed, and software-defined. The question isn't whether it's coming — it's which integration architecture wins.
There is a problem that no one in the March 11 press releases mentioned: what happens when autonomous EW systems on both sides make decisions at machine speed?
If DiSCO-Hivemind detects a threat emitter and autonomously repositions drones to jam it, and the adversary's autonomous EW system detects the jamming and autonomously escalates to a different countermeasure, and DiSCO-Hivemind autonomously responds to that escalation — you have an electromagnetic engagement happening entirely between machines, at speeds no human can monitor, let alone control.
This is the logical endpoint of removing the human from the EW loop. The electromagnetic spectrum becomes an arena where AI systems engage each other in real time, with humans learning about the outcome after the fact. Each side's autonomous EW agent probes, adapts, and counters at machine speed in a continuous escalation spiral that may resolve — or crash — before a human decision-maker even knows it started.
China's Atlas drone swarm system, demonstrated on March 25, 2026, already integrates reconnaissance, electronic warfare, and strike drones in a single autonomous kill chain — one operator managing 96 drones.[10] Russia's electronic warfare has forced Ukrainian drones toward full autonomy as the only way to survive in a contested spectrum.[11] The convergence is global, and it's accelerating.
The Pentagon's nonkinetic effects cell — created after Operation Absolute Resolve — was designed to integrate cyber, EW, and directed energy into operational planning with humans at the center. DiSCO-Hivemind suggests the technology is already moving faster than the doctrine. The humans may get their planning cell. The machines may not wait for it.
The DiSCO-Hivemind demonstration is the moment autonomous systems and electronic warfare formally merged. Not in a PowerPoint slide or a concept paper — in a hardware-in-the-loop test where multiple unmanned platforms detected, analyzed, and responded to electromagnetic threats without human intervention.[1]
Ghost Spectrum covered how Anduril's Pulsar gave individual EW systems the ability to learn autonomously. The Hivemind takes the next step: autonomous forces that maneuver through the electromagnetic battlespace as a coordinated unit, with AI making the tactical decisions that humans used to make — at speeds humans cannot match.
The strategic trajectory is clear. Legacy electronic warfare — human analysts monitoring scopes, operators deciding which threat to jam, library-based systems fighting the last war — is being replaced by distributed, AI-driven networks that sense, decide, and act across the full electromagnetic spectrum in real time. L3Harris, Shield AI, Anduril, CACI, and the entire defense industrial base are converging on this architecture simultaneously.
The live flight tests later this year will prove whether the concept works in real RF environments, not just simulations. But the direction is set. The electromagnetic spectrum is no longer a domain where humans fight with electronic tools. It's becoming a domain where machines fight each other — and humans read the after-action report.
By integrating autonomous decision-making with advanced battle management technology, we're answering the Pentagon's urgent call for coordinated command and control of multiple unmanned systems.