The U.S. Army put roughly 1,900 drones into a major exercise in Germany while Ukrainian troops worked inside American formations to pass along battlefield lessons, a sign that four years of drone warfare are beginning to change how U.S. ground units train for large-scale combat.

About a dozen Ukrainian personnel joined Saber Junction 26 at the Joint Multinational Readiness Center in Hohenfels, advising the 173rd Airborne Brigade on drones, electronic warfare and tactics developed against Russia. Brig. Gen. Terry Tillis, commander of the 7th Army Training Command, said the aircraft represented the latest lessons from Ukraine, according to an AP report published Saturday.

The number is striking, but the more consequential result was how quickly drones shaped contact between forces. Army Chief Technology Officer Anthony Miller said about half of the brigade’s engagements with the opposing force began with a drone sighting. That finding points to a battlefield where concealment, electronic protection and rapid decisions may matter as much as the aircraft themselves.

A brigade-level laboratory

Saber Junction is an annual readiness exercise rather than a technology demonstration. The Army’s official exercise account says approximately 5,480 personnel participated from Aug. 15 through Sept. 13, including 3,360 members of the 173rd and forces from nine allied and partner nations. The scenario tested rapid deployment, an airborne entry, offensive and defensive operations, and the regeneration of combat power.

That setting matters because a drone that works at a test range can fail when it must share radio spectrum, intelligence and airspace with an entire brigade. At Hohenfels, units had to integrate reconnaissance aircraft, one-way attack systems, electronic warfare and conventional fires while an opposing force tried to disrupt them. The Ukrainians were not there to supply a finished doctrine; they were helping American units expose weaknesses under pressure.

The exchange also ran in both directions. Ukraine has adapted commercially derived aircraft rapidly because combat provides immediate feedback. A recent field report described the constant cycle of changing frequencies, components and tactics as each side develops new defenses. The U.S. Army has larger budgets and different missions, but its acquisition and training systems generally move more slowly than a wartime workshop.

The exercise therefore tested an institutional question: whether lessons learned from Ukraine can spread across a U.S. formation fast enough to survive contact with an adaptive enemy. Sending American soldiers to Ukraine and embedding Ukrainian advisers in Germany shortens that learning loop, but a few exchanges cannot replace a durable system for fielding, repairing and updating equipment.

The Army still lacks a settled structure

Army leaders agree that small drones must become common at lower echelons. The service is putting one-way attack aircraft down to company level, according to the AP, but officials have not settled whether specialized formations or distributed teams should carry most of the expertise. Those options solve different problems.

A dedicated unit can concentrate pilots, engineers and electronic-warfare specialists, accelerate experimentation and maintain scarce equipment. Distributing operators throughout infantry and armor formations can make drones immediately available to commanders and keep the technology tied to ordinary missions. The risk is that every unit gets aircraft but not enough trained personnel, maintenance capacity or protected communications to use them well.

That debate sharpened when the Army ended the specialized drone focus of a battalion within the 173rd after nearly a year of experimentation. The service said the change was planned and that learning should spread across the force; critics treated it as a possible retreat. Separate reporting documented the shutdown, and senators later sought an explanation through congressional questions.

Miller argued that centralizing drone knowledge would make it harder to scale lessons across the Army. Saber Junction offers evidence for that position: drone-enabled contact was not confined to a niche test unit. But the exercise does not by itself prove that a fully distributed model will preserve technical expertise once units return to routine training, personnel rotate and equipment changes.

Drones require defenses as well as operators

The 1,900-aircraft figure can obscure the other half of the contest. Every additional friendly drone adds demand for spectrum management, identification and airspace coordination. Every adversary drone creates a detection and interception problem that cannot be solved economically by firing a costly missile at a cheap aircraft.

During Saber Junction, U.S. air defenders used SGT STOUT short-range systems to protect artillery positions against simulated unmanned threats, according to an official Army account. That pairing of offensive drones with mobile counter-drone defense is more representative of current war than treating unmanned systems as a stand-alone capability.

Electronic warfare further complicates the picture. Jamming can break an enemy control link, but it can also interfere with friendly aircraft, navigation and communications. Autonomous navigation may reduce dependence on a radio link, while fiber-optic controls and changing frequencies can bypass some defenses. Effective formations will need operators who understand not only flight but also electromagnetic signatures, network security and how to keep conventional fires from attacking friendly systems.

Ukraine’s expanding use of unmanned ground vehicles illustrates the same combined-arms logic. A Reuters report found that Ukrainian units were using ground robots for logistics, evacuation, surveillance and attacks, while commanders acknowledged limits in terrain and production. The lesson is not that robots replace soldiers outright. It is that air and ground systems can shift dangerous tasks away from people when communications, tactics and supply are aligned.

Scale depends on an industrial base

An exercise can draw from a temporary pool of 1,900 aircraft. Sustaining that density across multiple brigades would require a much larger flow of batteries, motors, cameras, controllers and replacement airframes. Training losses and rapid obsolescence mean small drones behave more like consumable munitions than traditional aircraft.

The Pentagon’s Drone Dominance Program is designed to address that scale problem. Its supply framework describes a $1.1 billion advance market commitment intended to procure more than 300,000 small one-way attack systems over four phases. It also sets a path away from components sourced from countries considered security risks, reflecting concern about Chinese control of important parts of the commercial drone supply chain.

Volume alone will not create readiness. A large inventory can become obsolete if software, radio modules and payload interfaces are locked to one vendor. The more useful procurement standard is a system that soldiers can repair, reconfigure and replace without waiting years for a new program. Open interfaces and realistic field testing are therefore operational requirements, not merely contracting preferences.

Cost comparisons also need discipline. A small drone may be inexpensive beside a missile or crewed aircraft, but the full system includes operators, sensors, relays, electronic protection, training and replacement stocks. Cheap hardware can still produce an expensive and fragile capability if the network around it is poorly designed.

What the Army must prove next

Saber Junction demonstrated that drones can saturate a large American exercise and that Ukrainian experience can materially influence how a brigade trains. It did not answer how many operators each unit needs, which capabilities belong at company, battalion or brigade level, or how the Army will measure performance when both sides jam, spoof and hunt one another’s control systems.

The next tests should publish outcomes that go beyond aircraft counts: mission success rates, losses to electronic warfare, repair times, the speed from detection to fire, and the burden placed on communications and command posts. They should also distinguish reconnaissance, attack, logistics and counter-drone missions, because success in one does not establish readiness in the others.

The Army’s choice between specialized drone units and broadly distributed expertise may ultimately be false. A sustainable model could require both: centers that preserve engineering and training depth, and organic teams that make drones part of routine maneuver. Ukraine’s contribution at Hohenfels is valuable precisely because it treats adaptation as continuous rather than complete.

The most important result from 1,900 drones is not that the Army now possesses a new fleet. It is that half of one brigade’s simulated engagements reportedly began with unmanned contact. If that pattern holds across harder exercises, drones are no longer an attachment to ground warfare. They are becoming part of the environment in which every unit must move, hide, communicate and fight.