L3Harris DiSCO, Shield AI Hivemind, and the First Autonomous Electronic Warfare System That Doesn't Need a Human
On March 11, 2026, L3Harris Technologies and Shield AI announced what they called a first-of-its-kind demonstration: unmanned systems that autonomously detected, analyzed, and responded to electromagnetic threats in real time — with no human in the loop.[1]
The demonstration integrated L3Harris' Distributed Spectrum Collaboration and Operations platform (DiSCO) with Shield AI's Hivemind mission-autonomy software. Multiple unmanned aircraft systems collected electromagnetic signals, DiSCO fused them into a common operating picture of the spectrum, and Hivemind-powered drones used that picture to autonomously identify safe operating zones and execute tactical maneuvers.[1][2]
The test also incorporated L3Harris' Green Wolf electronic warfare ground vehicle — equipped with electronic attack and detection capabilities — alongside a software-defined radio payload for electronic support and a Shield AI UAS communications relay. All operating in a hardware-in-the-loop simulation designed to replicate real-world conditions.[2]
Electronic warfare moves at machine speed, and operational advantage depends on autonomy.
The DiSCO-Hivemind integration represents a fundamentally different approach to electronic warfare. Legacy EW puts a human at the center: analysts monitor the spectrum, operators make decisions, commanders authorize responses. This architecture worked when adversaries had a handful of radar systems operating on known frequencies. It breaks when the spectrum is saturated with hundreds of emitters — drones, jammers, decoys, spoofed signals — all evolving in real time.
DiSCO solves the sensing problem. L3Harris' platform collects electromagnetic data from multiple distributed platforms simultaneously and fuses it into a single coherent picture — what the companies call a "common operating picture of the electromagnetic spectrum." This is not a single sensor looking at one slice of spectrum. It's a distributed network of sensors building a real-time map of every signal in the battlespace.[1]
Hivemind solves the acting problem. Shield AI's autonomy software — the same platform that flies V-BAT drones without GPS and without a pilot — takes DiSCO's spectrum picture and makes tactical decisions: where to fly, what to avoid, what to jam, what to exploit. The drones don't wait for instructions. They see the spectrum, understand the threats, and maneuver accordingly.[2]
The combination is greater than its parts. DiSCO alone produces intelligence that still needs a human to act on it. Hivemind alone can fly autonomously but lacks spectrum awareness. Together, they create something new: an autonomous system that perceives the electromagnetic environment and acts within it at machine speed.
The Maven program of record — designated by Deputy SecDef Steve Feinberg on March 9 — gives AI the power to find and recommend targets.[3] The DiSCO-Hivemind integration gives AI something potentially more consequential: control of the electromagnetic spectrum itself.
Targeting is a step in the kill chain. Spectrum dominance is the substrate on which the entire kill chain runs. Every satellite link, every drone video feed, every GPS signal, every radar pulse, every communications channel — all of it flows through the electromagnetic spectrum. Whoever controls the spectrum controls what every other system can see, hear, and do.
The Pentagon's $13.4 billion FY2026 budget request for autonomous systems reflects this understanding.[4] But the money tells only part of the story. What the DiSCO-Hivemind demo proves is that the technology is real: AI systems can now perceive and act in the spectrum domain faster than humans can supervise them. The question is no longer whether autonomous EW is possible. It's whether any force that doesn't have it can survive.
Iran's Shahed drones cost $20,000-$50,000 each. A Patriot PAC-3 interceptor costs over $3 million.[5] That cost asymmetry drove the Pentagon's laser weapon push. But electronic warfare offers something lasers don't: you can disable an entire swarm by poisoning the spectrum it navigates through. No explosions, no ammunition, no cost-per-shot calculus. Just silence.
Six days after the DiSCO-Hivemind EW demonstration, a parallel development revealed the international reach of the same autonomy stack. On March 17, Mitsubishi Heavy Industries (MHI) and Shield AI announced the completion of autonomous flight tests in Japan, using Hivemind software integrated into MHI's ARMD prototype drone.[6]
The integration took less than two months. Two 20kg test vehicles — ARMD-01 and ARMD-02 — demonstrated learning behaviors and coordinated motions while tracking a virtual air vehicle. During the second flight, the drone performed more aggressive maneuvers based on learnings from the first. The AI didn't just execute — it improved between flights.[6]
MHI is the primary Japanese industrial partner in the tri-nation Global Combat Air Programme (GCAP) with BAE Systems and Leonardo. A sixth-generation GCAP fighter is designed to operate alongside lower-cost "attritable" collaborative combat aircraft — exactly the type of platform that Hivemind was built to control.
The implication is clear: Shield AI's autonomy software is becoming a de facto standard for allied autonomous operations. The same Hivemind that controls EW drones for L3Harris also flies MHI's next-generation combat aircraft in Japan. One software stack, spanning electronic warfare, air combat, and allied interoperability.
The DiSCO-Hivemind demonstration didn't happen in a vacuum. It landed in a week where every major defense storyline converged on the same theme: AI is taking over the battlefield faster than policy can keep up.
On March 9, Maven became a permanent program of record. On March 17, MIT Tech Review revealed the Pentagon's plan for AI companies to train on classified data. The same day, Senator Slotkin introduced the AI Guardrails Act — banning autonomous lethal strikes, AI surveillance of Americans, and AI in nuclear launch decisions.[3]
Meanwhile, India's DRDO unveiled a high-power microwave system at the Electronic Warfare Conference.[7] Australia is building autonomous laser turrets capable of destroying 200 drones on a single battery charge.[5] Turkey's Baykar flew five K2 kamikaze drones in autonomous swarm formation without GPS.
This is the context: every major military power is simultaneously pursuing autonomous EW, autonomous targeting, and autonomous swarm warfare. The nations that field these capabilities first don't just have better weapons — they have better weapons that get better with every engagement. A human-operated EW system fights the same way every time. An AI-operated one learns.
The DiSCO-Hivemind demonstration marks the moment electronic warfare crossed the autonomy threshold. For the first time, unmanned systems have demonstrated the full loop — sensing the electromagnetic spectrum, fusing multi-platform intelligence, and acting on it — without a human in the chain.
This matters more than it appears. Autonomous targeting (Maven) decides what to hit. Autonomous EW (DiSCO-Hivemind) decides what can see, hear, and communicate in the first place. One picks targets from a list. The other controls whether the list can be built at all. Spectrum dominance is upstream of everything.
The international dimension accelerates the timeline. Hivemind is not just an American capability — it's becoming the autonomy layer for allied combat aircraft through Japan's GCAP program. A single software stack now spans electronic warfare drones, combat air vehicles, and multi-national interoperability. This is how standards get set.
Live flight tests later this year will determine whether the simulation results hold in real-world contested spectrum. If they do, the Pentagon's vision of coordinated, autonomous, multi-domain electronic warfare moves from PowerPoint to tarmac. And the human EW operator — who has been the linchpin of spectrum warfare since World War II — begins the transition from decision-maker to supervisor.
The spectrum is learning to think for itself. The only question is whether human oversight can keep pace.
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.