Autonomous unmanned surface vessels (USVs) aren’t a lab experiment anymore. Ukraine has used them to push the Russian Black Sea Fleet back from its own coastline. The Philippines is running them today. The capability is real and proven. The question isn’t whether the U.S. Navy needs these vessels, it’s whether the Pentagon can actually get them into the fleet at the speed the moment demands. 

We’ve watched this movie before. 

Rewind 25 years and swap “USV” for “UAV.” In the late 1990s, drones were a curiosity, useful for reconnaissance, mistrusted for anything more, and stuck outside the services’ real planning and budgeting processes. It took the shock of 9/11 and two ground wars to force the issue. Only then did the Predator go from a niche ISR tool to an armed, mission-critical platform, and only then did the doctrine, training pipelines, and acquisition pathways catch up to what the technology could already do. That process took the better part of two decades. Nobody would design it that way on purpose. 

USVs are sitting roughly where UAVs sat in the mid-1990s: the hardware works, early adopters are already fielding it in combat, and the institutional plumbing hasn’t caught up. Task Force 59, the Navy’s first unmanned and AI task force, is the current version of those early Predator experiments: proof that autonomous systems aren’t just a nice-to-have supplement, but a genuine force multiplier, if the Navy lets them be. 

The harder problem was never the technology. It’s the system built around it. Defense acquisition still runs on a model designed for platforms that took a decade to develop and years to field, a pace that made sense when the alternatives to hard tradeoffs like unmanned surface vessels didn’t exist. It doesn’t work when adversaries are iterating on autonomy in months, not years, and every month of delay is a month of operational disadvantage. The FY2026 NDAA made real progress loosening some of those legal constraints on USVs―a good start, but just a start. What comes next has to go further: breaking the service-specific stovepipes that slow everything down and building acquisition authorities designed for how fast this technology actually moves. 

Look no further than the Strait of Hormuz for the stakes. Putting a manned ship within range of active threats risks lives and billion-dollar hulls in the same breath. A USV can sit in that same water doing ISR, holding sensor-to-shooter links, and absorbing the risk instead of a sailor facing it. Ukraine didn’t wait for a doctrine manual to figure that out, and neither will our adversaries. 

There’s a deeper shift buried in this that’s easy to miss. Past leaps in military hardware, faster jets, longer-range munitions, and stealth were exquisite-technology races. Whoever could out-engineer and out-spend the other side won, and that favored the side with the deeper budget. That was the U.S. Navy’s game to lose for eighty years. 

Autonomy inverts it. A cheap, attritable drone or unmanned boat lets a much poorer adversary impose real costs on a much wealthier one. Ukraine’s Black Sea campaign is the proof of concept, built substantially on low-cost unmanned platforms punching well above their price tag. That’s precisely why institutional resistance to USVs and UAVs alike has been so stubborn: the entire acquisition system was built to reward “faster, farther, more expensive.” Autonomy rewards “cheap, numerous, attritable” instead. It’s not just a new platform category, it’s a different muscle memory, and that’s a harder thing for a bureaucracy to relearn than any single technology ever was. 

None of this is really about hardware, either―it’s about trust. Just like it took a generation of pilots and commanders to trust an unmanned aircraft with a real mission, it’ll take training, repetition, and demonstrated performance before operators trust a USV the same way. That’s not a reason to slow-walk it. It’s the actual roadmap: build the training pipelines and the doctrine alongside the platforms, not after them. 

The Navy has adapted to transformational technology before: carrier aviation, nuclear propulsion, precision strike. But those were all still exquisite-platform transitions, played on the field the U.S. was built to win. Autonomous systems are the first shift in a generation that plays on the other field, and we already know how the story goes if we wait too long to adjust. We watched it happen with drones. The industrial base is ready to build the hybrid fleet now, manned and unmanned, networked together. The only open question is whether the institutions can learn the new game fast enough, or take another 20 years to relearn the lesson the hard way.