Understand CAPEX and OPEX in PPPs and concessions: lifecycle costs, reinvestments, maintenance, asset management, availability, handback, risks, and cost assumptions.
Check it out!
CAPEX and OPEX in PPPs and concessions are central assumptions for long-term contracts because they represent, respectively, the investments required to implement, expand, renew, or adapt infrastructure and the recurring costs needed to operate it and keep it available. In a concession, however, this separation cannot be treated as an isolated accounting exercise: CAPEX, OPEX, performance, maintenance, demand, reinvestments, and the final condition of assets form a single technical-economic lifecycle.
The most common mistake is to optimize the initial investment and treat future costs as a secondary consequence. In contracts spanning decades, a decision that appears economical during implementation may increase energy consumption, maintenance, unavailability, replacements, obsolescence, or upgrade needs. Engineering’s objective is to build cost assumptions consistent with the expected behavior of assets throughout the contract term.
What are CAPEX and OPEX in PPPs and concessions?
CAPEX represents expenditures associated with the creation, expansion, replacement, or material improvement of assets. OPEX represents costs required for day-to-day operations, maintenance, support, consumption, and service management. This distinction is useful, but it does not by itself resolve concession structuring.
In long-term contracts, certain expenditures may change classification according to the project’s accounting and financial policies. Engineering should not determine tax or accounting treatment; its role is to technically characterize what will be done, why it will be necessary, when it will occur, which asset will be affected, and what impact it will have on capacity, availability, and service life.
The Brazilian Law No. 8,987/1995 links concessions to investments, service quality, tariffs, oversight, and reversible assets. In PPPs, Brazilian Law No. 11,079/2004 incorporates risk allocation, performance-linked remuneration, and financial sustainability. These obligations require a lifecycle view, not merely an implementation budget.
Why do CAPEX and OPEX need to be analyzed together?
A technical alternative changes investment and operations simultaneously. More efficient equipment may require higher CAPEX and reduce energy use. Redundant architecture may increase investment and maintenance but reduce unavailability penalties. Automation may reduce labor while increasing software, telecommunications, and support costs.
Integrated analysis prevents false savings: reducing investment by removing redundancy, protection, instrumentation, maintenance accessibility, or expansion capability may transfer cost and risk into decades of operations.
The article on TCO and Lifecycle Cost in Engineering explores this logic in greater depth. In PPPs and concessions, it becomes even more important because the same contractual structure must support implementation, performance, and asset preservation over periods far longer than the lifecycle of a conventional project.
CAPEX: initial investment is not the entire investment
Concession CAPEX needs to originate from quantities, design maturity, interfaces, schedule, and traceable contingencies. The investment should reflect the asset that will actually be implemented and renewed.
A concession’s CAPEX does not end when operations begin. It is necessary to distinguish initial investment, expansions, regulatory upgrades, end-of-life replacements, technology modernization, and investments required to meet asset handback conditions.
Initial CAPEX typically concentrates design, civil works, electromechanical systems, electrical infrastructure, automation, telecommunications, equipment, auxiliary facilities, permitting, integration, testing, and commissioning. Each component has its own maturity, disbursement curve, and uncertainty.
When the economic model uses a single aggregated value, visibility is lost over which components are sufficiently defined and which remain subject to revision. Cost engineering should break the investment into traceable packages compatible with scope and schedule.
The level of project definition determines CAPEX reliability
An estimate produced during the conceptual phase does not have the same accuracy as a budget based on detailed design and consolidated quantities. This does not mean early estimates are useless; it means their uncertainty needs to be made explicit.
Structuring should record the source of quantities, price base date, inclusions, exclusions, productivity assumptions, logistics, taxes considered in the technical budget, contingencies, and items not yet defined. The value should be accompanied by a technical calculation record capable of explaining how it was formed.
The problem is not working with uncertainty. The problem is presenting an immature estimate as a deterministic number and allowing it to support tariffs, availability payments, or financing needs without a margin compatible with the project’s stage.
The cost structure should align with the WBS and schedule
CAPEX decomposition should be compatible with the project’s physical structure. Engineering, procurement, construction, integration, testing, and commissioning packages need to be associated with periods and milestones.
This linkage turns a static budget into a disbursement curve. Long-lead equipment may require advance payments. Civil works may consume resources in phases different from systems. Commissioning occurs close to the start of operations but may require mobilization and specialized resources contracted in advance.
When the schedule and budget use different structures, the financial model must distribute costs using artificial rules. This weakens the relationship between physical progress and disbursement.
Expansion CAPEX should be treated through triggers
Not every future investment occurs on a fixed date. Demand growth, coverage expansion, or performance deterioration may require expansion before or after the originally expected date.
The concession needs to define objective triggers. Utilized capacity, volume served, demand backlog, geographic expansion, or regulatory requirements may trigger additional investments. The model should represent not only the value of the expansion but also its engineering, permitting, procurement, implementation, and commissioning lead time.
A poorly structured trigger can produce two extremes: premature investment and idle capacity, or late expansion with service deterioration.
Reinvestment CAPEX needs to be separated from routine maintenance
Replacing a consumable component or carrying out routine maintenance is not equivalent to renewing an entire system at end of life. For modeling, this distinction is essential because reinvestments create material cash-flow peaks.
Electrical and electronic assets, servers, communication systems, automation, UPS systems, batteries, embedded software, and other technology components may go through several replacement cycles during a long concession. Civil works and structures typically follow different cycles.
The strategy should map asset families, expected service life, criticality, obsolescence, and replacement cost. In this way, reinvestment stops being a generic percentage and begins to reflect the actual asset portfolio.
OPEX: operations need to be modeled using cost drivers
OPEX is more robust when built from observable cost drivers. Number of teams, operating hours, energy consumption, asset count, area served, processed volume, software licenses, support contracts, and inspection frequency are examples.
Applying only inflation to an initial OPEX assumes that operations will remain structurally unchanged for decades. This is rarely true. Demand changes, assets age, technologies are replaced, service levels may be revised, and regulatory requirements evolve.
The model should separate fixed, variable, and semi-variable costs. This structure makes it possible to simulate demand and expansion scenarios without distorting all costs in the same proportion.
OPEX, maintenance, and asset management need to form a single logic
OPEX, maintenance, and reinvestments need to share the same asset base. Asset register, criticality, condition, and lifecycle make operating costs technically verifiable.
The maintenance plan is not a document parallel to the economic model. It is the source of a significant share of OPEX and reinvestments. Preventive, predictive, corrective, and condition-based strategies have different costs, resources, and risks.
Asset management connects asset register, criticality, condition, service life, failures, interventions, and cost. In concessions, this integration also supports evidence that infrastructure is being preserved and prepared for handback.
When OPEX is reduced without revising the maintenance strategy, the savings may appear only in the spreadsheet. In actual operations, less maintenance may increase failures, unavailability, penalties, and the need for early replacement.
Energy and utilities can dominate certain OPEX categories
In data centers, water and wastewater systems, transportation, lighting, HVAC, industrial facilities, and other energy-intensive assets, consumption may represent a material share of operating cost. The assumption needs to originate from loads, operating regime, efficiency, utilization factors, and applicable tariffs.
It is not enough to multiply installed power by 8,760 hours. Systems operate under variable profiles and have losses, redundancy, standby modes, seasonality, and different loading levels.
Engineering should provide consumption curves or scenarios consistent with operations. This makes it possible to assess efficiency, automation, and control alternatives with verifiable impacts on OPEX.
Staffing and outsourcing need to reflect the operating philosophy
The number of operators, maintenance personnel, supervisors, security staff, customer service staff, and support personnel should not be estimated solely by benchmark. It should be compatible with architecture, automation, criticality, geographic distribution, shifts, response time, and contractual obligations.
Outsourcing can convert part of the fixed structure into service contracts, but it does not eliminate the need for governance, oversight, and technical management. The model should avoid double counting and also avoid assuming that outsourcing fully transfers operational risk.
A clear operating philosophy improves comparability among bids and reduces the risk of undersized service levels being used to support an artificially low price.
Software, licenses, and digital services need to be included in the lifecycle
Modern infrastructure incorporates VMS, SCADA, BMS, EMS, CMMS, data platforms, control systems, database licenses, and cloud services. Costs may be annual, per device, per user, by data volume, or by capacity.
These expenses should not appear only as an IT line item. They are part of the asset’s functional availability. End of support, licensing changes, or version incompatibility may require simultaneous hardware and software upgrades.
A long-term concession needs to provide for a technology-upgrade strategy and avoid dependence on commercial assumptions that cannot be sustained for decades.
Does OPEX increase as assets age?
Not always linearly, but the assumption of constant real OPEX needs to be justified. As assets age, failure frequency may increase, parts may become scarce, and interventions may require greater effort.
On the other hand, reinvestments may reduce maintenance and energy costs. The correct behavior depends on the asset family and renewal strategy.
The model should therefore connect OPEX and reinvestments. A renewal CAPEX peak may be followed by a temporary reduction in maintenance. Ignoring this interaction creates double counting or underestimation.
Availability, quality, and OPEX cannot be separated
Performance-based contracts procure outcomes, not merely the presence of assets. To maintain high availability, the concessionaire may need redundancy, spare-parts inventory, standby teams, support contracts, and preventive maintenance.
Reducing these resources lowers OPEX but may increase the risk of deductions from availability payments or loss of revenue. The model needs to represent this relationship consistently with the performance measurement system.
Engineering contributes by defining architecture, failure modes, repair time, redundancy criteria, and maintenance requirements. These elements make it possible to assess whether the proposed OPEX supports the contracted service level.
CAPEX vs. OPEX: there is no universal ratio
Market percentages may help with screening, but they do not replace analysis. A system with high CAPEX and low OPEX may coexist with another with lower initial investment and intensive operations. Sector, technology, scale, environment, and service level completely change the relationship.
Comparison among alternatives should use equivalent bases and a common horizon. It should also consider reinvestments, transition costs, demobilization, and residual asset condition.
A decision based on a generic ratio may favor a solution unsuitable for the specific contract context.
Inflation and indexation need to reflect the nature of each cost
Engineering does not define monetary policy or contractual adjustment mechanisms, but it should help break costs down according to their drivers. Imported equipment, labor, energy, civil materials, and software may respond to different indices and markets.
When all lines are escalated by the same index, the model may hide material exposure. Financial and legal structuring will determine how to treat adjustments; engineering should provide a sufficiently granular cost breakdown for that decision.
It is also necessary to maintain the distinction between nominal and real prices to avoid mixing inflation with physical increases in quantity, aging, or expansion.
Uncertainty, contingency, and risks need to be treated explicitly
Contingency only supports governance when it arises from characterized risks and coherent contractual allocation. Generic percentages do not replace technical risk analysis.
Contingency should not be a percentage used to correct any uncertainty in the project. It should be compatible with identified risks and with the level of estimate definition.
Geotechnical risk, permitting, interferences, land acquisition, critical supplies, productivity, foreign-exchange exposure on imported equipment, obsolescence, and availability are examples of events that may affect CAPEX or OPEX through different mechanisms.
The risk matrix should indicate who manages and bears each event. Cost estimating and modeling should be consistent with this allocation to avoid charging for the same risk twice or transferring it without real management capability.
Contingency reserve does not replace engineering maturity
The less defined the scope, the wider the uncertainty range. However, increasing contingency without developing surveys, design, and interfaces does not solve the information problem.
Sustainable uncertainty reduction comes from engineering: investigation, requirements definition, coordination, quantities, execution planning, and clarification of responsibilities.
Contingency should decrease or be redistributed as risks are mitigated and the project matures. A reserve that remains unchanged regardless of stage is a sign of weak governance.
Sensitivity analysis helps identify which costs really matter
Not every assumption deserves the same refinement effort. Sensitivity analysis helps identify which CAPEX and OPEX lines most affect tariffs, availability payments, return, or debt coverage.
O conteúdo sobre Sensitivity and Scenario Analysis in Engineering Projects shows how to work with switching values and scenarios. In concessions, this approach helps prioritize where to invest in due diligence and engineering maturation.
If a small variation in energy consumption, demand, or schedule changes the decision, those assumptions require stronger evidence than low-materiality lines.
CAPEX and OPEX need to be consistent with the risk matrix
Transferring construction risk to the private partner tends to influence price, contingency, and guarantees. Retaining land-acquisition risk with the granting authority requires the model to represent consistent responsibilities and schedules. Shared risk may require triggers and rebalancing mechanisms.
Cost modeling cannot assume a base case incompatible with the contract. If a given risk is fully transferred, its pricing needs to be analyzed; if it is retained, it should not appear as a generic concessionaire margin without justification.
Consistency between cost and risk is part of Value for Money and bankability.
The reinvestment schedule needs to align with financing
Major replacements in the middle of the concession may occur while the initial debt is still being amortized. The project needs cash, reserves, or a financing structure capable of supporting these events.
Engineering does not define leverage, but it needs to deliver a technically defensible reinvestment curve. If all critical systems are assumed to last forever, the financing structure will be built on an asset that does not exist.
Periodic reviews of the asset plan make it possible to update dates and costs without losing the long-term view.
Handback creates CAPEX at the end of the contract
Handback conditions may require minimum residual life, absence of maintenance backlog, demonstrated functionality, documentation, and compliance. These requirements may generate inspections, major maintenance, and replacements in the final years.
If the model does not reserve resources for handback, an incentive arises to reduce maintenance near the end of the concession. The risk returns to the granting authority in the form of deteriorated assets.
Therefore, handback criteria need to be defined early and connected to the maintenance plan, asset management, and reinvestments.
How to structure a lifecycle CAPEX-OPEX matrix
A useful matrix organizes asset families, initial investment, maintenance frequency, annual cost, service life, probable replacement date, renewal cost, criticality, and handback requirements.
| Cost family | Initial CAPEX | Recurring OPEX | Reinvestment | Primary driver |
| Civil works | high | inspection/maintenance | occasional renewal | physical condition |
| Electromechanical | high | energy/maintenance | medium/long term | hours and load |
| Automation/control | medium | support/licenses | shorter cycles | obsolescence |
| IT/telecom | medium | connectivity/support | recurring | technology/capacity |
| Security and monitoring | medium | operations/maintenance | recurring | performance and coverage |
The table does not replace the detailed model, but it helps verify whether any asset group disappears after implementation.
How to audit cost assumptions before procurement
The review needs to verify whether two competent bidders receive enough information to develop comparable prices. Quantities, condition of existing assets, availability requirements, demand, interfaces, and responsibilities need to be clear.
A technical audit may verify:
- price basis and date;
- maturity level of quantities;
- consistency between scope and CAPEX;
- disbursement curve versus schedule;
- operating philosophy and OPEX composition;
- maintenance and reinvestment plan;
- energy and utility consumption;
- licenses and recurring contracts;
- expansion triggered by demand;
- contingencies and risks;
- handback condition and residual life.
The lowest CAPEX is rarely the lowest contractual cost
The concession should be assessed by its ability to deliver performance throughout the lifecycle. A solution that is cheap to implement may generate higher operating expenditure, unavailability, and replacements. A more robust solution may reduce OPEX but also create uncompensated excess capacity.
The appropriate decision is the one that balances requirements, risks, and total cost within the contractual logic. This requires engineering of alternatives, not merely comparison of unit prices.
Final considerations
CAPEX and OPEX in PPPs and concessions need to be treated as parts of the same lifecycle architecture. Initial investment, operations, maintenance, reinvestments, and handback are consecutive stages of the same asset and should share consistent technical assumptions.
Engineering gives traceability to these assumptions, shows where cost originates, identifies uncertainties, and connects expenditure to performance. The stronger this basis, the lower the risk of structuring an apparently economical contract that becomes unviable, unavailable, or costly during operations.
Technical references
[1] BRAZIL. Law No. 8,987 of February 13, 1995. Establishes the concession and permission regime for public services. Available at: [https://www.planalto.gov.br/ccivil_03/leis/l8987compilada.htm](https://www.planalto.gov.br/ccivil_03/leis/l8987compilada.htm).
[2] BRAZIL. Law No. 11,079 of December 30, 2004. Establishes general rules for procurement and contracting of public-private partnerships. Available at: [https://www.planalto.gov.br/ccivil_03/_ato2004-2006/2004/lei/l11079compilado.htm](https://www.planalto.gov.br/ccivil_03/_ato2004-2006/2004/lei/l11079compilado.htm).
[3] PINHEIRO, Armando Castelar et al. Structuring PPP and concession projects in Brazil: diagnosis of the Brazilian model and proposals for improvement. São Paulo: IFC, 2015. Available at: [https://web.bndes.gov.br/bib/jspui/handle/1408/7211](https://web.bndes.gov.br/bib/jspui/handle/1408/7211).
[4] PROJECT MANAGEMENT INSTITUTE. A Guide to the Project Management Body of Knowledge (PMBOK Guide). 8th ed. 2025.
Frequently asked questions
CAPEX represents investments in implementation, expansion, replacement, or material improvement of assets; OPEX represents recurring costs of operations, maintenance, support, and consumption.
Technically, they are investments in asset renewal or replacement. The specific accounting treatment should be defined by the competent function, but engineering needs to characterize the value, timing, and reason for the reinvestment.
Percentages may serve for initial screening, but long-term contracts require technical drivers such as staffing, energy, maintenance, licenses, demand, criticality, and asset age.
Because design choices simultaneously affect investment, operations, maintenance, availability, and renewal needs.
Yes. Handback conditions may require major maintenance, replacements, inspections, and minimum residual life of assets.
Risks affect quantities, schedule, productivity, maintenance, consumption, reinvestments, and contingencies. The estimate should be consistent with the contractual risk-allocation matrix.
Additional technical resources
Related services
- Cost Engineering for Engineering Works and Services
- Engineering Asset Management
- Engineering Risk Management
- Technical and Economic Feasibility Study
- Engineering Technical Due Diligence
