Investing in data centers is a capital-intensive infrastructure undertaking characterized by long development timelines and complex cost structures. Unlike many conventional real estate or infrastructure projects, the financial performance of a data center depends not only on its initial construction costs but also on available power capacity, IT load utilization, electricity tariffs, operating expenses, customer mix, and the ability to scale capacity over time.
Accordingly, a data center investment should be evaluated across its entire lifecycle. This article examines four key financial metrics that investors need to understand when assessing a project: Capital Expenditure (CAPEX), Operating Expenditure (OPEX), Return on Investment (ROI), and Internal Rate of Return (IRR).
CAPEX - DETERMINING THE INITIAL CAPITAL INVESTMENT
Capital Expenditure (CAPEX) represents the total investment required to develop, construct, equip, and commission a data center for commercial operations.
For a data center project, CAPEX typically includes:
Land acquisition or land-use costs and site preparation;
Civil construction and building works;
Electrical infrastructure and power distribution systems;
Generators, uninterruptible power supply (UPS) systems, and energy storage systems;
Cooling systems;
Fire detection and suppression systems;
Physical security and access control systems;
Network infrastructure and telecommunications connectivity;
Information technology (IT) equipment;
Design, engineering consultancy, project management, and permitting costs; and
Contingency allowances and financing costs incurred during construction.
Electrical infrastructure and cooling systems are particularly significant cost components that require careful assessment, especially for high-density data centers supporting artificial intelligence (AI) and high-performance computing (HPC) workloads.
Therefore, when developing a financial model, investors should not focus solely on the total capital investment. CAPEX should also be broken down by development phase and cost component to establish an accurate assessment of funding requirements and the timing of capital deployment.
Basic CAPEX formula:
CAPEX = Land Costs + Construction Costs + Electrical Infrastructure + Cooling Systems + IT and Network Equipment + Other Equipment + Consultancy and Permitting Costs + Contingency
A project with relatively low CAPEX does not necessarily deliver superior financial returns, particularly if its utilization rate is low or its cumulative OPEX over the project lifecycle is high.
OPEX - DATA CENTER OPERATING AND MAINTENANCE COSTS
Once construction has been completed and the facility has been commissioned for commercial operations, a data center incurs recurring operating expenses, commonly referred to as Operating Expenditure (OPEX).
OPEX generally varies with the scale of operations and the facility's occupancy or capacity utilization, although certain fixed costs remain payable regardless of the actual IT load.
Key OPEX components may include:
Electricity costs;
Water consumption and cooling system operations;
Personnel and staffing costs;
Equipment maintenance and servicing;
Physical security and facility operations;
Network and connectivity costs;
Insurance premiums;
Management and administrative expenses; and
Equipment replacement and upgrades.
Energy is often one of the most significant components of data center OPEX, potentially accounting for more than 50% of total operating expenses. Electricity tariffs, energy efficiency, and the effectiveness of cooling system management can therefore have a material impact on project cash flow.
Human resources are another critical consideration. Maintaining a skilled workforce capable of operating a data center reliably, efficiently, and in accordance with stringent safety and security requirements to protect customers' data requires a long-term workforce development strategy.
This often begins with selecting a site that offers convenient transport and logistics access and proximity to technical education and training institutions. Depending on the location and labor market, investment in accommodation for technical specialists may also be necessary.
Balancing CAPEX and OPEX through strategic investment
At the individual project level, certain CAPEX and OPEX components involve trade-offs. An investment that increases upfront capital expenditure may reduce recurring operating costs over the facility's lifecycle.
For example, implementing liquid cooling typically requires a higher initial investment in specialized equipment and infrastructure than conventional air cooling. However, for large-scale facilities serving AI workloads, liquid cooling may improve cooling efficiency, reduce electricity and water consumption, and support the reliable operation of high-performance computing equipment over the long term.
Workforce planning presents another example. Selecting a site with access to a qualified labor pool may involve higher upfront costs, such as land acquisition or investment in training partnerships with educational institutions. These costs may, however, be offset by lower recurring expenses for salaries, allowances, transportation, and accommodation for employees recruited from distant locations.
More sophisticated solutions may also help optimize these trade-offs. Examples include collaborating with industrial park developers to co-invest in workforce accommodation that can subsequently be leased to employees, and exploring cooling solutions that recover cold energy from liquefied natural gas (LNG) regasification processes to improve energy efficiency.
The key consideration for foreign investors entering an emerging data center market is to establish mutually beneficial partnerships with existing infrastructure owners, technology providers, and complementary service operators. Such partnerships can help unlock operational efficiencies and improve project economics that may otherwise be difficult to achieve through standalone investment.
ROI - RETURN ON INVESTMENT
Return on Investment (ROI) is a widely used metric for assessing the return generated relative to the capital invested.
For a data center, financial returns are derived from revenue generated through services such as colocation, data hosting, and the provision of computing capacity, among others. These returns should be evaluated against the investment and operating costs associated with the project.
General formula:
ROI = Net Return on Investment / Investment Cost × 100%
ROI provides investors with an initial indication of capital efficiency. It can also help assess the competitiveness of proposed rental rates and service pricing in the enterprise market.
One useful approach is to evaluate ROI from the customer's perspective. For example, a large technology company seeking to expand its data infrastructure to support AI research and applications may compare leasing data center capacity with developing and operating a dedicated facility.
The following example illustrates a simplified comparison between these two alternatives:
Based on these assumptions, leasing results in approximately USD 2.95 million in cost savings over the 10-year period, as measured by the difference between the present values of the two alternatives.
The resulting ROI of the leasing option is approximately 34%, calculated as USD 2.95 million divided by USD 8.54 million, from the perspective of the customer in this example.
Furthermore, allocating the self-build costs over the relevant period, excluding relocation and other ancillary consumables, results in an estimated break-even level of approximately USD 230/kW/month.
This provides one potential reference point for data center investors seeking to establish and negotiate service pricing for facilities intended for lease. Actual pricing decisions should also account for utilization, service specifications, contract duration, financing structure, and market conditions.
Nevertheless, given the long operating lifecycles and multi-year cash flow profiles characteristic of data center projects, ROI alone may not fully capture the time value of money or the overall economic viability of an investment.
IRR — INTERNAL RATE OF RETURN
Internal Rate of Return (IRR) represents the annualized rate of return implied by a project's or investment's projected cash flows.
Conceptually, IRR is the discount rate at which the Net Present Value (NPV) of an investment equals zero. It is therefore useful for evaluating the relationship between the timing of an investment's cash outflows and its expected future cash inflows.
Calculating NPV
NPV can be calculated using the following formula:
[
NPV=-C_0+\sum_{t=1}^{n}\frac{C_t}{(1+r)^t}
]
Where:
C₀: Initial investment at Year 0.
Cₜ: Net cash flow received in year t.
t: The relevant year, from 1 to n.
n: The total number of periods in the investment analysis.
r: The discount rate, representing the minimum required rate of return. Depending on the cash flows and valuation framework, this may be based on the weighted average cost of capital (WACC) or another appropriate required rate of return.
The denominator, ((1+r)^t), discounts future cash flows to their present value. This reflects the time value of money: receiving USD 100 five years from now is not economically equivalent to receiving USD 100 today because of factors such as inflation, interest rates, and opportunity costs.
The higher the discount rate, the lower the present value of a given future cash flow.
IRR is the discount rate that solves the equation NPV = 0.
Interpreting IRR for investment decisions
IRR can be considered alongside other financial indicators when deciding whether to proceed with a data center investment.
A common preliminary decision framework compares IRR with the project's required rate of return or applicable hurdle rate:
IRR < Required rate of return: The project generally does not meet the investor's minimum return requirement and may be rejected or require restructuring.
IRR = Required rate of return: The project meets the minimum return threshold, but the investment decision should consider additional financial and strategic factors.
IRR > Required rate of return: The project may be financially attractive, subject to the reliability of its assumptions and the assessment of other material risks.
In the current environment, where AI infrastructure investment is influenced by expectations of future growth, debt financing, and cross-investments among technology companies, ROI and IRR should be assessed alongside other indicators, including NPV and the Payback Period, which will be examined in greater detail in subsequent articles.
For foreign investors entering Vietnam's data center market, financial feasibility must be assessed in conjunction with legal and regulatory requirements, land availability, power supply, construction, telecommunications, environmental compliance, and operational considerations.A financial model is only as reliable as its underlying assumptions. Investors must establish whether a proposed site is suitable, how and when the project can access sufficient power, what regulatory approvals are required and how long they may take, what development costs are likely to arise, and whether the facility can attract customers once it becomes operational.
As a project advisory partner supporting data center development in Vietnam, DCompass can assist investors in assessing investment conditions, site suitability, legal and regulatory matters, infrastructure readiness, implementation partners, and other factors affecting project development. Integrating legal and regulatory, site selection, power, construction, technology, financial, and operational considerations from the outset enables investors to develop financial models that more closely reflect actual implementation conditions and provide a stronger basis for investment decisions.
REFERENCES
[1] Cushman & Wakefield. (2026, September 3). Cushman & Wakefield Releases 2026 Data Center Development Cost Guide, Citing 21% Rise in Per-MW Construction Costs. Press release, Business Wire.
https://finviz.com/news/388803/cushman-wakefield-releases-2026-data-center-development-cost-guide-citing-21-rise-in-per-mw-construction-costs
[2] CRE Daily. (2026). Data Center Construction Costs Jump 21% Since 2024.
https://www.credaily.com/briefs/data-center-construction-costs-jump-21-since-2024/
[3] Donnellan, D., Lawrence, A., Bizo, D., Rogers, O., Judge, P., O’Brien, J., Davis, J., Smolaks, M., & Weinschenk, R. (2026, July). Uptime Institute Global Data Center Survey 2026 (UII Keynote Report 209). Uptime Institute.
https://intelligence.uptimeinstitute.com/resource/uptime-institute-global-data-center-survey-2026
[4] Sermsuk, M., Sukjai, Y., Wiboonrat, M., & Kiatkittipong, K. (2021). Utilising cold energy from liquefied natural gas (LNG) to reduce the electricity cost of data centres. Energies, 14(19), 6269.
https://doi.org/10.3390/en14196269
[5] Everitt, N. (2019, October 25). Data Centres Could Harness LNG Cool Energy. Cooling Post.
https://www.coolingpost.com/features/data-centres-could-harness-lng-cool-energy/
[6] Shahi, P., Mathew, A., Akash, Saini, S., Bansode, P., Kasukurthy, R., & Agonafer, D. (2022). Assessment of reliability enhancement in high-power CPUs and GPUs using dynamic direct-to-chip liquid cooling. Journal of Enhanced Heat Transfer, 29(8), 1–13.
https://par.nsf.gov/servlets/purl/10454830
[7] Green, A., Tai, H., Noffsinger, J., & Sachdeva, P. (2024, September 17). How Data Centers and the Energy Sector Can Sate AI’s Hunger for Power. McKinsey & Company.
https://www.mckinsey.com/industries/private-capital/our-insights/how-data-centers-and-the-energy-sector-can-sate-ais-hunger-for-power
[8] Pacheco, M. (2023, October 24; updated December 11, 2025). Data Center Build vs. Buy: Decision Guide. TierPoint.
https://www.tierpoint.com/blog/data-center-build-vs-buy
[9] Brealey, R. A., Myers, S. C., & Allen, F. (2020). Principles of Corporate Finance (13th ed.). McGraw-Hill Education. ISBN 978-1-260-56555-3.
[10] Financial Times. (2026). Nvidia’s Chips May Be Novel, but Its “Circular Financing” Isn’t. Republished by Female Invest.
https://www.femaleinvest.com/inspiration-articles/ft---nvidias-chips-may-be-novel-but-its-circular-financing-isnt
[11] Tilleke & Gibbins. (2025, September 26). Vietnam’s Data Center Market: A Strategic Guide for Investors.
https://www.tilleke.com/insights/vietnams-data-center-market-a-strategic-guide-for-investors/3/
[12] Vietnam News. (2026, September 12). Data Centre Boom Courted, but Power Must Keep Pace.
https://vietnamnews.vn/economy/1799380/data-centre-boom-courted-but-power-must-keep-pace.html
