Autonomous Military AI: US Drone Fires AMRAAM Missile in Historic Test

What Happened?

The US Air Force (USAF) has confirmed that the autonomous combat drone YFQ-44A, developed by Anduril Industries, successfully fired a live AIM-120 AMRAAM missile against a digital target during a test over the Mojave Desert. According to TechRadar, the test included prior phases where the drone carried an inert version of the missile to validate flight stability and communication links.

General Ken Wilsbach, a senior USAF commander, stated that "it wasn't just an AMRAAM that was released, but it was tracking the target," highlighting the system's ability to integrate sensor and weapon in an autonomous environment. The test is part of the Collaborative Combat Aircraft (CCA) program, which aims to deploy combat drones as digital wingmen to support human pilots.

Why It Matters

This milestone represents a qualitative leap in military autonomy. Until now, armed drones like the MQ-9 Reaper required remote human control to fire. The YFQ-44A, in contrast, executed the complete sequence of detection, tracking, and launch autonomously, though the USAF insists that the decision to fire will always require human authorization. However, the test demonstrates that the underlying technology—sensors, targeting algorithms, and data links—is already mature for real combat operations.

The context is key: the USAF plans to acquire at least 1,000 CCAs over the next decade, with a cost target of one-third that of the F-35 (about $28 million vs. $83 million). The fiscal 2027 budget already requests $1.4 billion for development and $1 billion for procurement. This could transform the economics of air warfare, allowing low-cost drone swarms to saturate defenses while manned fighters remain at a safe distance.

Immediate and Future Consequences

The test accelerates the AI arms race in autonomy. China and Russia are already developing their own AI combat drones, and this milestone pressures their programs. It also raises ethical and strategic dilemmas: what level of autonomy is acceptable? The USAF states that "autonomous systems will not independently decide when to fire," but the line between tactical autonomy and lethal decision-making is blurry.

For industry, Anduril solidifies its position as a key player against giants like General Atomics (whose YFQ-42A Dark Merlin competes in the same phase). Integration challenges also emerge: CCAs will need to operate with F-35s and F-22s, sharing data in real time. The test validates that the sensor-effector loop can be closed without direct human intervention, reducing latency in high-intensity scenarios.

What Readers Should Know

This is not an autonomous killer drone straight out of movies. It is a system that automates certain tasks under human supervision, similar to advanced autopilot systems. However, the technological gap is narrowing: the ability to fire a live missile at a moving target implies that identification and threat prioritization algorithms already work in real conditions.

For tech companies, the CCA program represents a multi-billion-dollar market in autonomy software, sensors, and communications. Companies like Shield AI and Collins Aerospace are competing for the software contract, while Anduril and General Atomics vie for hardware. The final production decision is expected by June 2026.

"It wasn't just an AMRAAM that was released, but it was tracking the target." — General Ken Wilsbach, USAF

The future of air combat will be hybrid: humans and machines collaborating, but with machines taking on increasingly high-risk tasks. The question is not whether it will happen, but how it will be regulated.