Why America’s drone swarms are still just crowds in the sky

In a recent critique published by War on the Rocks, experts in distributed systems and military modernisation dismantle the myth that drone swarming is already here. Despite flashy military demos and awe-inspiring drone light shows, they argue the U.S. Department of Defense has not achieved true swarming capability — not even close. Instead, the article warns that what’s being labelled as “swarming” is little more than synchronised, robotic mass — visually impressive, but strategically limited.

At its core, the authors contend that real swarming isn’t about numbers or speed; it’s about architecture. Swarms aren’t just many drones acting at once — they are cohesive, resilient, adaptive systems that operate without a central controller. Drawing parallels with biological collectives like ant colonies or bee hives, the article defines true swarming as “collaborative, autonomous, machine-speed adaptation” where the group thinks and responds as one. This is a far cry from current U.S. capabilities, which remain mired in leader-follower dynamics or one-to-many operator models — all of which are vulnerable to single points of failure.

The authors outline why swarming represents not just an incremental change but a paradigm shift in military operations. In warfare’s evolutionary arc — from melee to mass to manoeuvre — swarming is the next developmental leap. It promises more than just more efficient operations; it holds the potential to overwhelm current strategies through coordination, speed, and adaptability. However, the Pentagon has so far focused on hardware and AI improvements while overlooking the necessary backbone: distributed systems infrastructure.

AI, they argue, is not enough. Intelligence alone can’t make machines coordinate. Without a common operating picture — one created and maintained in a decentralised, cloud-independent manner — autonomous systems remain isolated. Swarming requires drones to share information in real time, even when communications are jammed or individual nodes fail. This level of decentralisation and synchronisation exists in theory and in civilian data centres — but not yet in deployed military systems.

A particularly troubling insight is how industry misuse of the term “swarming” has diluted its meaning. Defence contractors, eager to sell the next big thing, are branding simple drone groupings as swarms. Policymakers, imagining sci-fi-style hive minds, don’t realise they’re being shown glorified remote-controlled formations. This mismatch between expectation and reality muddies acquisition priorities and stifles innovation. The result? U.S. forces are optimising for the wrong fight.

To fix this, the authors issue a clarion call: acquisition leaders must clearly specify the need for resilient, distributed systems that are cloud-independent and locally self-sufficient. This shift in focus would not only make true swarming possible — it would future-proof American capabilities against adversaries who may be closer to achieving this breakthrough.

By Vugar Khalilov