Fort Irwin exercise exposes gaps in the U.S. Army’s AI-warfare push as drone failures and network problems complicate plans for autonomous weapons

A 10-day U.S. Army exercise at Fort Irwin, California, exposed technical and institutional hurdles in the military’s effort to bring AI-enabled systems and autonomous weapons into the field, The Wall Street Journal reported. About 9,700 U.S. and allied personnel took part in the July exercise.

Newly acquired drones, including models from Anduril, faced overheating, battery failures, signal loss and lengthy launch times. Soldiers limited flights to conserve battery power, while experimental communications systems and augmented-reality headsets also encountered problems.

The exercise highlighted the challenge of adapting to rapidly changing battlefield technology, particularly in comparison with Ukraine. Ukrainian drone operators defeated a U.S. armored brigade during a military exercise in Germany earlier this year, the report said.

Starlink connectivity and AI-assisted planning systems improved during the Fort Irwin exercise, but officials said significant technical and institutional challenges remain before autonomous weapons can be deployed reliably at scale.

The exercise, Project Convergence Capstone 6, ran from July 20 to 29 at the Army’s National Training Center. The Army described it as a test of its Next Generation Command and Control system at division scale, involving U.S. forces and allied partners.