The question is not whether unmanned aviation will operate from ships by 2035. It already will. The sharper question is what form “unmanned carrier” will take in a decade and a half and whether any navy will field a large deck, fully autonomous aircraft carrier in that time frame. My answer, grounded in current programs and measured against engineering and operational realities, is that we will see widespread adoption of unmanned air wings and specialized unmanned motherships by 2035 but a fully autonomous, nuclear-powered supercarrier operating without human supervision is unlikely within that horizon.

What is already funded and getting to sea matters. The U.S. Navy has formalized a strategy to scale unmanned systems across the fleet and to pursue manned-unmanned teaming as an operational norm. That institutional commitment creates both the budgetary momentum and the doctrinal imperative to integrate carrier-capable unmanned aircraft into carrier strike groups and to experiment with distributed, unmanned platforms as force multipliers.

On the air side, the most concrete program is the MQ-25 Stingray. The MQ-25 is intended as a carrier-based tanker and an initial proof point for carrier-launched unmanned aviation. It is being integrated into the Carrier Air Wing alongside human-piloted fighters and is explicitly the Navy’s pathfinder for future carrier-based unmanned systems. Getting a proven, repeatable launch, recovery, deck handling, and command chain in place for MQ-25-class aircraft will lower technical and procedural risk for larger unmanned air wings.

At the platform level navies are already experimenting with two distinct architectural approaches that will both be important through 2035.

  • Repurposed or light carriers and amphibious ships acting as drone carriers. Turkey’s adaptation of the TCG Anadolu into a short-deck drone carrier and similar experiments show an operationally pragmatic path: take a hull with a large aviation deck, add datalinks, automated launch/landing aids, and shipboard handling to support persistent, medium endurance unmanned aircraft. This is a lower barrier to entry and a fast way to field meaningful capability.

  • Unmanned motherships and large unmanned surface vessels operating as distributed nodes. China’s early public demonstrations of an unmanned mothership able to carry and coordinate dozens of aerial, surface, and undersea drones foreshadow another route. The unmanned mothership concept is attractive because it separates the most expensive, survivable capital ship from expendable or attritable unmanned assets that can be mass-produced and risked forward. Expect more experimentation and fielding of these concepts through 2035.

Those two trends are already visible in procurement lines. The U.S. Navy’s Large Unmanned Surface Vessel and Medium Unmanned Surface Vessel efforts, and the experiments tied to the Overlord/Ghost Fleet work, are designed to prove autonomy, signature management, reliability, and a command and control architecture that can be scaled. If those programs succeed in producing durable, remotely supervised USVs and robust C2, they will be the maritime backbone for unmanned carriers of sensors and weapons. But the schedule is deliberate because shipboard reliability, power management, and modular payload integration are hard engineering problems.

Hard technical limits slow the leap from proof of concept to fully autonomous supercarriers. Two concrete engineering choke points illustrate why.

1) Launch and recovery systems. Modern catapult and arresting technologies were designed around predictable, human-piloted aircraft and tightly integrated shipboard power and maintenance chains. New systems such as the electromagnetic aircraft launch system are powerful enablers for a broader range of aircraft weights and arresting gear improvements. But they have shown reliability and maintenance challenges in service-test regimes, which in turn affect sortie generation and operational availability. Those problems matter more for an unmanned-centric carrier because an unmanned air wing depends on high sortie throughput and minimal human intervention in maintenance loops. Fixing those issues at scale will take iterative engineering, spares pipelines, and mature maintenance doctrine.

2) Deck handling, ordnance supply, and sustainment automation. A carrier’s combat effectiveness is not just launches and recoveries. It is how quickly aircraft are refueled, rearmed, repaired, and reconfigured between missions. Current carrier designs rely on complex weapons elevators, specialized shops, and a large skilled complement. Automation of those processes is possible and will be pursued vigorously, but it requires proven robotics, ship integration, and cybersecurity-hardened control systems before commanders confidently remove humans from the loop on a large deck ship.

Operational and doctrinal friction will also shape adoption. Fully unmanned high-value vessels present difficult command, legal, and escalation questions. Navies will therefore prefer hybrid solutions in which human crews remain responsible for high-level decisions and for damage control, while unmanned platforms undertake surveillance, strike in permissive environments, attritable swarm attacks, and logistics runs. The ‘‘mothership plus swarm’’ construct reduces single-point risk and leverages bodies of cheaper uncoupled platforms rather than a single enormous remotely piloted capital ship. Evidence for early adoption of that pattern exists today in multiple navies.

Putting the pieces together yields a realistic 2035 landscape.

  • Common by 2035: Carrier-capable unmanned aircraft routinely integrated into carrier strike group operations for roles such as refueling, ISR, and limited strike. Early enablers like MQ-25 will be operational and doctrinally embedded.

  • Common by 2035: Drone motherships and LUSV/MUSV deployments in distributed formations conducting persistent surveillance and layered strike with human supervision. These will appear in exercises and forward deployments as risk-tolerant complements to manned ships.

  • Less likely by 2035: A true nuclear supercarrier sized hull that routinely sails and fights entirely without a human crew. The engineering and operational hurdles around survivability, damage control, legal and command arrangements, and the political appetite for delegating lethal decisions to an unmanned capital ship make that an unlikely near-term outcome. The industrial and budgetary inertia behind manned carriers will slow any wholesale replacement. Instead expect manned carriers to evolve into networked hubs that control and sustain large numbers of unmanned systems operating at range.

A short scenario analysis clarifies risk tradeoffs. A navy seeking to field an unmanned carrier force quickly can choose a low-cost route: refit an amphibious ship or commercial hull to operate medium endurance UCAVs plus robust satcom links and specialized deck handling. That will field an operationally useful capability within a few years and is the path Turkey and some others have pursued. The higher-cost, higher-capability route of building a new, survivable, large-deck unmanned capital ship that can operate in high-threat environments requires breakthroughs in autonomous damage control, redundant power systems, hardened C2 links, and legal frameworks for remote lethal action. Those breakthroughs are plausible but will take longer than a single decade.

Policy and procurement implications for 2024 and beyond are straightforward. Navies should accelerate investments in the components that unlock unmanned carrier operations: robust shipboard autonomy standards, automated deck and ordnance handling, secure and resilient beyond-line-of-sight datalinks, and logistics models for rapid unmanned aircraft turnover. They should treat EMALS, arresting gear, weapons elevators, and large-ship power distribution as dual-use investments that benefit both manned and unmanned air wings. At the same time they must fund realistic experimentation in attritable swarms and mothership concepts to avoid concentrating risk in a single expensive hull.

Final assessment in one line. By 2035 we will have carrier aviation architectures dominated by mixed manned-unmanned air wings and by a diversity of mothership concepts operating together. Full replacement of manned supercarriers with autonomous equivalents by that date is unlikely. The transition will be evolutionary and modular rather than revolutionary and monolithic. The sensible planning assumption for policymakers and program offices is to fund both incremental integration of unmanned carrier assets and a portfolio of distributed, attritable platforms that can be produced and risked at scale.