REPMUS has evolved from a Portuguese-led experimentation exercise into the de facto NATO proving ground for sea robotics and manned-unmanned integration. After REPMUS 2024 demonstrated a string of practical innovations from floating autonomous “islands” to integrated anti-submarine warfare concepts, expectations for REPMUS 2025 have tilted firmly toward autonomy as an operational multiplier and force enabler rather than a pure technology showcase.

Three concrete trends will shape the 2025 agenda. First, persistent layered sensing with unmanned surface vessels, unmanned underwater vehicles, and aerial drones will move from ad hoc demos to scripted operational experiments. NATO and national commands have been explicitly pursuing Task Force X concepts to mass maritime autonomous systems for persistent surveillance and response. That organizational intent, paired with live USV demonstrations conducted by Allied Maritime Command in February 2025, means REPMUS 2025 will be evaluated against mission-end states such as continuous detection, classification, and handoff to manned platforms.

Second, undersea critical infrastructure protection will be a central use case. Political and military attention on undersea cables and pipelines has translated into operational initiatives like Baltic Sentry and supporting activity to harden maritime vigilance in the Baltic. REPMUS 2025 will therefore emphasize integrated concepts for protecting undersea lines of communication, combining UUVs for detailed inspection, USVs for persistent surface watch, and airborne assets for cueing and long-range detection. Expect experiments that measure detection timelines, cueing fidelity, and the end-to-end chain of custody for forensic evidence collection.

Third, anti-submarine warfare experimentation will continue to be reframed around distributed autonomous assets. The ASW barrier experiments first prototyped at REPMUS 2022 and refined in 2023 and 2024 moved beyond single-platform demonstrations toward networked sensing with sonobuoy fields, large AUVs carrying towed arrays, and distributed processing nodes. REPMUS 2025 will be judged on the degree to which multiple vendor systems can establish a shared acoustic picture and support persistent tracking across shallow and deep water gradients.

From a technical vantage point there are four immediate metrics NATO should prioritize when assessing autonomous demonstrations at REPMUS 2025.

1) Interoperability at the data layer. Demonstrations must show not only that a USV can hand a contact to a frigate, but that the acoustic and metadata formats, timestamps, and location provenance survive system handoffs with measurable end-to-end latency and error bounds. Past REPMUS events have already used COTS assets and academic nodes to prototype data exchange across domains. REPMUS 2025 needs a repeatable interoperability test harness and common message profiles to produce quantitative results.

2) Trustworthy autonomy, defined by predictable failure modes and recovery behaviors. Fielded autonomy is not useful if it degrades unpredictably under communications contention, GPS denial, or moderate sea states. Experimental scenarios should include contested comms, GNSS denial, and degraded sensor suites to stress resilience. The aim is to articulate a set of autonomy assurance metrics that map directly to operational risk tolerances.

3) Mission-integrated human-machine teaming. REPMUS must move beyond “operator-in-the-loop” demos where humans simply monitor. Instead the exercise should evaluate delegated authorities, escalation pathways, and human oversight in time-constrained kinetic and non-kinetic tasks. That requires telemetry of operator workload, decision latency, and the fidelity of human intent propagation into autonomous plans.

4) Sustainment and logistics for long-duration maritime autonomy. Systems operating at sea must be evaluated on refuel or recharge cycles, data offload mechanisms, and at-sea maintenance concepts such as the autonomous “island” mothership prototypes trialed in 2024. REPMUS 2025 provides a realistic environment to measure true operational availability rather than lab-run endurance numbers.

Operationally, NATO’s ongoing move to institutionalize Task Force X creates both an opportunity and a constraint for REPMUS 2025. Task Force X aims to codify how Allies pool maritime autonomous capabilities to provide persistent surveillance and deterrence. REPMUS is the deconfliction ground in which national architecture meets Alliance-level concepts of employment. If REPMUS 2025 can demonstrate a federated mission with national contributors delivering plug-and-play observability, it will accelerate Doctrine to Tactics to Acquisition cycles. Conversely, if the exercise exposes brittle interfaces, proprietary chokepoints, or opaque AI behaviors, it will underline how much investment remains in standards and operational engineering.

There are persistent risks that REPMUS 2025 must address head on. First, dependence on commercial sensors and vendor-specific datalinks introduces supply chain and sovereignty concerns. Second, an overemphasis on capability demonstrations risks producing optimism bias in commanders who must later accept system limitations in combat. Third, legal and escalation dynamics associated with autonomous platforms operating near critical infrastructure require clear rules of engagement and forensic chains that survive contested conditions. These are not technical problems alone. They are policy, legal, and industrial challenges that must be exercised in parallel with technical trials.

Recommendations for a meaningful REPMUS 2025 outcome

  • Require instrumented baseline scenarios. Each experiment should include objective metrics for detection, track continuity, latency, and false alarm rates. Avoid purely qualitative “success” narratives.

  • Publish stakeholder interface specifications. NATO should publish minimal data profile specifications for ASW cueing, CUI detection reporting, and USV-to-ship handoff that industry can implement and test in advance.

  • Stress contested environments. Include GNSS denial, comms jamming, and cyber intrusion red teams in at least one major experiment to measure realistic robustness.

  • Exercise legal and forensic chains. Validate evidence collection for undersea infrastructure incidents from sensor through to prosecution or attribution workflows. Baltic Sentry and associated efforts make clear this is an operationally salient requirement.

REPMUS 2025 will not solve the autonomous navy problem in a single season. However, if the exercise prioritizes measurable interoperability, resilience under contention, and mission-relevant human-machine teaming, it can move autonomous maritime systems from impressive demos to credible operational tools. The Alliance needs hard numbers on detection timelines, track survivability, and system availability captured under repeatable test conditions. Those numbers will determine whether autonomy delivers a durable operational advantage or simply a faster path to new integration headaches.