Sustainment is the invisible tax that decides whether robotic systems are strategic enablers or orphaned curiosities. The economics of keeping robots operational over years and decades is not an addendum to acquisition. It is the dominant cost center of the enterprise and therefore must be the primary lens through which we design, buy, and field autonomous and remotely operated systems. Empirically, sustainment frequently accounts for the majority of life‑cycle cost for complex systems, which explains why sustainment policy and practice will determine the long run affordability of robotic fleets.

What do we mean by “economic model” for sustainment? At minimum it is an explicit decomposition of total life‑cycle cost into acquisition and operating and support elements, coupled to performance metrics that translate technical readiness into economic value. A practical model therefore has three linked layers: (1) financial accounting for costs over time, (2) operational metrics that matter to commanders, and (3) contract and supply‑chain structures that allocate risk and incentives between government and industry. Without all three layers present and reconciled, models produce misleading answers. The Department of Defense has recognized this need by institutionalizing condition‑based maintenance and performance‑oriented product support approaches as central sustainment policy.

Canonical cost decomposition and key drivers. The simplest useful formula is: LCC = Procurement + O&S, where O&S themselves break down into maintenance labor, depot and MRO (maintenance, repair, overhaul), spares and consumables, contractor logistics support and performance fees, infrastructure, training, software sustainment, cybersecurity, and disposal. Two pragmatic metrics grow out of this decomposition: cost per mission or cost per operational day, and fleet availability. Cost per mission focuses budget planners. Availability is what commanders pay for. The interplay between these metrics shapes economic tradeoffs. Investing to raise mean time between failure reduces cost per mission only if the up‑front investment and recurring modernization costs are less than the avoided O&S expense and operational risk.

Why robotic systems are different cost problems. Robots add three sustainment burdens not seen to the same degree in traditional platforms. First, software and data are central to capability and must be sustained with updates, patches, model retraining, and cybersecurity hardening for the life span of the platform. Second, rapid commercial obsolescence in electronics and sensors forces ongoing engineering refresh cycles rather than simple depot repair. Third, distributed autonomy generates large volumes of health and usage data that create both an opportunity and a cost: building the data infrastructure is expensive, but it enables predictive maintenance that can reduce O&S. The DoD roadmaps and sustainment guidance emphasize these peculiarities as reasons to address sustainment early in the life cycle.

Economic model primitives used in practice.

  • Life‑Cycle Costing (LCC): Standard present value models discounted over the program life. This is the baseline by which acquisition and sustainment tradeoffs are evaluated. LCC estimates are sensitive to assumptions about operating tempo, failure rates, and log‑chain responsiveness.

  • Availability‑based contracting and Performance Based Logistics (PBL): Instead of paying for parts and hours, PBL pays for outcomes such as mission capable rate or time to repair. PBL aligns incentives to reduce total O&S cost and is now a cornerstone of DoD product support strategy. It requires robust metrics, honest baselines, and attention to perverse incentives.

  • Condition‑Based and Predictive Maintenance (CBM+/PHM): Sensorization and prognostics shift costs forward into data systems and analytics but enable reductions in unscheduled maintenance and spare inventory by predicting failure windows. DoD policy codifies CBM+ as a primary sustainment strategy when feasible. The economics favor CBM+ when failure modes are detectable, diagnostics are reliable, and the data chain is trustworthy.

  • Attritable and Disposability Models: For certain classes of unmanned systems, force‑employment concepts assume loss is tolerable if unit cost is low enough. The attritable model substitutes procurement volume and lower unit price for expensive sustainment. The economics here hinge on not just unit price but logistics costs of resupply and the strategic value of expendability versus repairability. Recent analyses of conceptual attritable air vehicles demonstrate large variance in projected savings depending on assumptions about operating tempo and replacement cost.

  • Contractor Logistics Support and Public‑Private Hybrids: Outsourcing depot and sustainment to industry can produce scale economies but risks vendor lock and reduced government insight into costs and technical data. Historical cases in unmanned aviation show both the benefits of industry performed logistics and the dangers when data rights and commonality are not preserved.

Modeling recommendations and pitfalls. First, model at the fleet level not the unit level. Many sustainment costs scale nonlinearly. Spare parts inventories, depot facilities, and skilled maintainer pools benefit from aggregation; small fleets are disproportionately expensive to sustain per unit. Second, explicitly budget for software and cyber sustainment as a recurring O&S line item rather than burying it in ad hoc engineering. Third, avoid overconfidence in MTBF projections derived from laboratory tests; field failure distributions commonly differ and drive cost growth. Fourth, value data rights and modular interfaces early. Open architectures permit competition for spares and upgrades, which compresses long‑term costs.

Empirical anchors and a sober cost example. Operating and support costs for remotely piloted or unmanned air systems have long been a substantial portion of program budgets in historical studies. For example, detailed studies of persistent unmanned air systems show significant steady state annual O&S burdens when ground segments, manning, and sustainment support are included. Those examples anchor the intuition that acquisition is only the upfront capex; true affordability is determined by decades of O&S flows.

How to build a practical economic model for a robotic program. I recommend the following engineering and financial steps.

  1. Define operational demand scenarios. Specify realistic missions, sortie rates, and replacement rates across a plausible spectrum. Sustainment cost projections collapse into a single number only under a single OPTEMPO assumption.

  2. Itemize cost buckets across the life cycle. Separate mechanics labor, depot MRO, spares, transportation, contractor fees, software engineering, cybersecurity, training, and infrastructure. Include obsolescence and periodic technology refresh cycles.

  3. Build a failure and repair model. Use field data where available. If not, use conservative distributions that permit sensitivity analysis. Couple failures to spare consumption and depot throughput.

  4. Add contracting and incentive layers. Model scenarios with organic maintenance, CLS, PBL with availability targets, and mixed models. Simulate how each contracting approach changes firm incentives to reduce total cost.

  5. Run portfolio and fleet aggregation analyses. Evaluate how unit costs decline or rise with fleet scale and commonality assumptions.

  6. Value alternative architectures. Compare highly modular open designs against tightly integrated proprietary systems. Include the option value of future upgrades and third party competition for parts.

Policy implications and closing thoughts. Economically rational sustainment for robotic systems is not simply a matter of buying more sensors or better parts. It requires institutional alignment across acquisition, logistics, and operational communities. Policymakers must internalize that sustaining robotic capability is a decade long commitment that starts before first flight or first patrol. Invest in CBM+ and data infrastructure where the signal to noise ratio supports prognostics. Use PBL to tie contractor reward to readiness outcomes but preserve government access to technical data and to the ability to inject competition. Treat software and cyber as recurring O&S commitments on par with fuel and manpower.

Finally, there is a philosophical point. Robots are agents of leverage. Their promise is scaling. But leverage multiplies not only effect but also fragility if sustainment is treated as an afterthought. If we want robotic fleets to be a force multiplier and not an enduring fiscal lever that bends budgets, then sustainment economics must be designed into the system from the start. The good news is that modern data science, sensorization, and new contracting constructs give us tools to do so. The harder work is political and organizational. Without that, no model will save an otherwise unaffordable program.