How is annual energy cost calculated?

Annual energy cost = IT_load_kW × effectivePUE × 8760 h/yr × utilization × energyRate ($/kWh). Effective PUE = base PUE + cooling-type delta (−0.25 immersion to +0.15 air-cooled chiller). Climate zone applies a coolingEnergyMult to the cooling-overhead portion only.

What drives the maintenance cost?

Maintenance is computed per system (generator $31.50/kW, cooling $38/kW, electrical $28.50/kW, fire $10/kW, BMS $14/kW, building $7.50/kW at ID baseline), then multiplied by: facility age factor (0.6 new / 1.0 mid / 1.5 old), country maintMult, and OEM factor (1.0 in-house, up to 2.5 full-contractor). Climate zone applies an additional coolingMaintMult to the cooling system only.

How many FTE are estimated?

Base FTE scales with IT load (kW) and Uptime tier. The Uptime benchmark for critical-facilities staffing is 4.2 FTE/MW for full in-house, reduced to 3.8 (hybrid 70/30), 3.5 (hybrid 50/50), or 3.0 (full outsourced). Each FTE cost = laborRate × fxToUSD × 1.35 burden × staffing model cost multiplier.

OPEX Calculator — Technical Manual & Methodology

01 Purpose & engineering basis

The OPEX Calculator produces an annual operating-cost estimate for a colocation or owned data center. It decomposes total cost into six independently-modelled lines: energy (IT load × PUE × hours × rate), staffing (FTE by tier and staffing model), maintenance (per-system $/kW rates, age-adjusted, OEM-factored), spare parts inventory (risk-weighted by contract model), software & licensing (DCIM, BMS, CMMS, cybersecurity), and insurance & compliance (jurisdiction-specific). A 4 000-run Monte Carlo seeded around these deterministic core values produces P10/P50/P90 confidence bands.

The primary design authorities are ASHRAE TC 9.9 (2021) for PUE & thermal envelopes, Uptime Institute Tier Standard for staffing benchmarks (4.2 FTE/MW critical-facilities), and IEC 62040-3 / ISO/IEC 30134-2 for energy metrics. Country electricity and labor rates come from national utility filings and statistics (as-of 2026, cited per row in DATA.opsBudget.countries).

Screening-grade accuracy: this calculator is a pre-design and benchmarking tool. Actual OPEX will vary with specific contract terms, local tax, overtime structures, and energy markets. Detailed cost-in-use studies should be grounded in site-specific contracts and utility quotations.

02 Inputs

Fourteen inputs across four groups — facility profile, energy model, staffing model, and optional advanced parameters. Core inputs are always active; advanced inputs are unlocked in the expanded panel.

Field (HTML id)	Type	Range / default	Description
IT Load `itLoad`	number	100 – 100 000 kW · default 1 000	Nameplate IT power draw of all installed equipment. Base for all energy and staffing calculations.
Base PUE `pueSlider`	slider	1.10 – 2.00 · default 1.50	Power Usage Effectiveness before cooling-type adjustment. Effective PUE = base PUE + `cooling.pueImpact`.
Country `countrySelect`	enum	20 countries · default ID	Sets `energyRate` ($/kWh), `laborMult`, `maintMult`, and `workerTax` from `DATA.opsBudget.countries`.
Tier Level `tierLevel`	enum	I / II / III / IV · default III	Uptime Tier drives staffing headcount multipliers and spare-parts tier multiplier (T-IV = 1.35×, T-III = 1.15×, T-II = 1.00×, T-I = 0.80×).
Facility Age `facilityAge`	enum	new / mid / old · default mid	Age factor applied to all maintenance line items (new = 0.60×, mid = 1.00×, old = 1.50×). Also drives spare-parts age multiplier (0.65 / 1.00 / 1.45).
Climate Zone `climateZone`	enum	6 zones · default tropical_humid	Sets `coolingEnergyMult` (applied to cooling-overhead portion) and `coolingMaintMult` (applied to cooling maintenance). See §04 for zone constants.
Cooling Type `coolingType`	enum	7 types · default water_chiller	Sets `pueImpact` (effective PUE delta), `efficiencyMult` (cooling-power modifier), and `maintCostPerKw` (baseline cooling maintenance rate). See §04.
Generator Config `genConfig`	enum	N / N+1 / 2N / 2N+1 · default N+1	Drives generator maintenance multiplier (N=1.0×, N+1=1.30×, 2N=2.10×, 2N+1=2.40×) applied to the generator $/kW maintenance rate.
Staffing Model (card selector)	enum	inhouse / partial / contractor	Selects the FTE cost multiplier, OEM factor, and spare-parts contract multiplier. Partial mode uses slider `inhousePercent` (10–90%).
In-House % `inhousePercent`	slider	10 – 90% · default 70%	Only active in partial mode. Blends OEM factor: `1.0 + (oemMult−1.0) × contractorPct` where `contractorPct = (100 − inhousePercent) / 100`.
Utilization `utilizationSlider`	slider	40 – 100% · default 70%	Capacity utilization applied to annual energy calculation. OPEX calculator uses retail-screening basis (0.70); DCMOC uses DC-contract basis (1.00).
DCIM Level `dcimLevel` (advanced)	enum	0 / 1 / 2 / 3 · default 1	Adjusts software-licensing basket and BMS maintenance rate. Higher DCIM levels add CMMS and monitoring tool costs.
Energy Contract `energyContract` (advanced)	enum	spot / fixed / ppa · default fixed	Modifies the effective energy rate: spot adds volatility risk premium, PPA applies a discount factor. Used for Monte Carlo variance seeding.
Maintenance Level `maintLevel` (advanced)	enum	basic / standard / comprehensive · default standard	Scales the OEM-factor range: basic caps OEM at 1.5×, comprehensive allows up to 2.5×. Drives overall maintenance spend vs contract-risk trade-off.

03 Calculation methodology

Six independently-modelled cost lines, each with its own data path. All function names below map to the page-side calculation in opex-calculator.html, cross-checked against RZEngine.models.opsBudget.opex() and RZEngine.models.opsBudget.staffing().

Line 1 — Annual energy cost

effectivePUE = basePUE + coolingType.pueImpact coolingPowerKW = IT_kW × (effectivePUE − 1) × coolingType.efficiencyMult × climate.coolingEnergyMult adjustedTotalKW = IT_kW + coolingPowerKW annualKWh = adjustedTotalKW × 8 760 h/yr × utilization energyCost = annualKWh × country.energyRate [$/yr] In updateCalculation() lines 3286–3291. Climate zone multiplier applies only to the cooling-overhead portion of power, not to IT load. Engine equivalent: loadKW × pue × energyRate × 8760 × utilization in RZEngine.models.opsBudget.opex().DATA.opsBudget · 2026

Line 2 — Staffing cost

baseFTE = f(IT_kW, tier) // tier-scaled from Uptime 4.2 FTE/MW benchmark staffCost = baseFTE × laborRate_local × fxToUSD × laborBurden(1.35) × staffModel.costMult Staff models: inhouse costMult=1.00 · hybrid-70 costMult=0.88 · hybrid-50 costMult=0.78 · outsourced costMult=0.65. Labor burden 1.35 covers benefits, tax, overhead. calculateStaffingFTE(itLoadKW, tier) → calculateStaffingDetail(fte, countryData).DATA.opsBudget · Uptime Institute · 2026

Line 3 — Maintenance cost (per-system)

oemFactor = 1.0 (in-house) oemFactor = 1.0 + (oemMult − 1.0) × contractorPct (partial) oemFactor = oemMult (full contractor) cost_sys = IT_kW × baseCostPerKw_sys × ageFactor × country.maintMult × oemFactor_sys coolingMaint additionally multiplied by climate.coolingMaintMult totalMaint = Σ cost_sys [generator + cooling + electrical + fire + BMS + building] Age factor: new 0.60 / mid 1.00 / old 1.50. System base rates at Indonesia baseline — see §04 constants table. Generator maintenance uses generator.maintMult for redundancy configuration.DATA.opsBudget · industry cost surveys · 2026

Line 4 — Critical spare parts inventory

cost_spare = IT_kW × baseCostPerKw × tierSpareMult × ageSpareMult × contractSpareMult × country.maintMult tierSpareMult : T-IV 1.35 / T-III 1.15 / T-II 1.00 / T-I 0.80 ageSpareMult : new 0.65 / mid 1.00 / old 1.45 contractSpareMult: inhouse 1.00 / partial 0.55 / contractor 0.15 Contractor model transfers spare-holding risk to the OEM; full all-risk contract leaves owner with only a 15% emergency buffer. Total baseline (in-house, T-III, mid-age) ≈ $36/kW/yr.DATA.opsBudget · industry cost surveys · 2026

Line 5 — Software & licensing

cost_sw = IT_kW × baseCostPerKw_sw × (0.70 + 0.30 × country.laborMult) 70% of software cost is license-driven (geography-independent); 30% is support-contract driven (scales with local labor). Total baseline ≈ $16.50/kW/yr covering DCIM ($6), BMS license ($3.50), CMMS ($2.50), environmental monitoring ($1.50), cybersecurity OT ($3.00).DATA.opsBudget · vendor pricing surveys · 2026

Line 6 — Insurance & compliance

complianceCost = Σ (baseCost_item + perMW_item × loadMW) [per jurisdiction] Indonesia-specific items: SLO $5 000 + $1 500/MW, AMDAL $3 000 + $500/MW, BPJS TK $1 500, Fire audit $2 500 + $500/MW, ISO 27001 $8 000 + $1 000/MW. Default jurisdiction: electrical + safety + environmental + ISO 27001 bases. Insurance is a flat multiple of asset replacement value.DATA.opsBudget · Indonesian MEMR / BPJS / ministry schedules · 2026

Total OPEX and per-unit KPIs

totalOPEX = energyCost + staffCost + totalMaint + totalSpares + totalSoftware + insurance + compliance perKW = totalOPEX / IT_kW / 12 [$/kW/month] perRack/yr = totalOPEX / rackCount rackCount = IT_kW / 10 kW (default rack density) perM2/yr = totalOPEX / whiteSpaceM2 whiteSpaceM2 = rackCount × 3.5 m² In computeDerivedMetrics(result). Monte Carlo (4 000 runs) varies: energy price (uniform ±15–20%), PUE drift (uniform 0.97–1.05×), labor/maintenance escalation (normal μ=4%, σ=2%). Tornado analysis ranks which driver most moves OPEX.RZEngine v2.0 · models.sim.tornado

04 Constants & data sources

Every constant lives in DATA.opsBudget with a corresponding DATA.sources['opsBudget'] citation (as-of 2026). Partial listing of the most material constants:

Cooling-type database — coolingTypes

Cooling type	PUE impact	Efficiency mult	Maint $/kW/yr (ID)	OEM mult
Air-Cooled Chiller	`+0.15`	`1.15`	18.00	2.2
Water-Cooled Chiller	`±0.00`	`1.00`	32.00	2.2
Hybrid Free Cooling	`−0.10`	`0.80`	28.00	2.2
In-Row Cooling (CRAC)	`+0.05`	`1.05`	22.00	2.0
Rear-Door Heat Exchanger	`−0.05`	`0.90`	20.00	2.0
Direct Liquid Cooling	`−0.15`	`0.70`	45.00	2.5
Immersion Cooling	`−0.25`	`0.60`	55.00	2.5

Climate-zone database — climateFactors

Zone	Cooling energy mult	Cooling maint mult	Free-cooling hrs/yr	Example markets
Tropical Humid	`1.00`	`1.00`	0	Indonesia, Malaysia, Singapore
Tropical Dry	`0.92`	`1.15`	500	UAE, Saudi Arabia
Subtropical	`0.88`	`0.95`	1 500	Hong Kong, Taiwan, Australia-QLD
Temperate	`0.75`	`0.85`	4 000	Europe, Japan, South Korea
Continental	`0.70`	`0.80`	5 000	US Central, Canada
Cold Climate	`0.55`	`0.75`	7 000	Nordic, Northern US

Country economics — DATA.opsBudget.countries (selected rows)

Country key	Labor rate (local/yr)	Energy rate ($/kWh)	FX to USD	Trop factor	Source · as-of
Indonesia	18 000 IDR-k/yr	0.085	0.000063	1.12	PLN tariff · 2026
Singapore	72 000 SGD/yr	0.168	0.74	1.10	SP Group · 2026
Malaysia	48 000 MYR/yr	0.082	0.22	1.11	Tenaga · 2026
India	24 000 INR-k/yr	0.075	0.012	1.08	CERC · 2026
Japan	82 000 JPY-k/yr	0.210	0.0067	1.00	TEPCO · 2026
US-Virginia	95 000 USD/yr	0.068	1.00	1.00	Dominion Energy · 2026
US-Texas	88 000 USD/yr	0.058	1.00	1.06	ERCOT avg · 2026
US-Oregon	90 000 USD/yr	0.042	1.00	1.00	PacifiCorp · 2026
EU-Germany	78 000 EUR/yr	0.235	1.08	1.00	BDEW · 2026
EU-Netherlands	75 000 EUR/yr	0.195	1.08	1.00	ACM · 2026
UAE	65 000 AED/yr	0.082	0.27	1.15	DEWA · 2026
Saudi Arabia	60 000 SAR/yr	0.048	0.27	1.18	SEC · 2026
Australia	88 000 AUD/yr	0.185	0.65	1.05	AEMO · 2026
UK	72 000 GBP/yr	0.215	1.26	0.96	Ofgem · 2026

System maintenance base rates — maintenanceRates (Indonesia baseline $/kW/yr)

System	Base $/kW/yr	OEM mult (full contractor)	Frequency
Generator (genset + fuel + 6M/12M OH)	31.50	2.5×	Monthly + 6M/12M OH
Cooling / HVAC (chiller + consumables)	38.00	2.2×	Monthly PM
Electrical (UPS / PDU / STS)	28.50	2.0×	Quarterly
Fire suppression system	10.00	1.8×	Semi-annual
BMS / DCIM / Controls	14.00	2.0×	Continuous
Building / Facility	7.50	1.5×	Quarterly

Key engine scalars — DATA.opsBudget

Constant	Value	Meaning
`shiftMult`	4.2	Shift coverage multiplier for 24/7 critical facility staffing (Uptime benchmark)
`laborBurden`	1.35	Fully-loaded multiplier over base salary (benefits, payroll tax, overhead)
`maintPerMwUsd`	$180 000/MW/yr	Comprehensive maintenance base (engine-level; page uses per-system rates)
`pmBaseline`	0.50	PM ratio baseline (50% — average across sites)
`pmSavingsSlope`	0.30	Maintenance savings per unit of PM ratio above baseline
`maintCapexShare`	0.15	Fraction of maintenance budget classified as capital expenditure for FFO calculation
`ffoFactor`	0.12	FFO as fraction of total OPEX
`outsourcedCostFactor`	0.65	Cost ratio of outsourced vs in-house staff at equal headcount
`replacementCostFactor`	1.5	Replacement cost as multiple of annual salary (SHRM benchmark)
`hoursPerYear`	8 760 h	Calendar constant (non-leap year)

05 Outputs

Output KPI	Formula / source	Unit	Interpretation
Total annual OPEX	Σ six cost lines	$/yr	Primary output. Displayed as `totalOpex` in the results panel.
OPEX per kW per month	`totalOPEX / IT_kW / 12`	$/kW/month	Industry benchmark metric. Tier-III US target: $80–130/kW/month.
OPEX per rack per year	`totalOPEX / rackCount`	$/rack/yr	Colocation pricing reference. Rack density default 10 kW/rack.
OPEX per m² per year	`totalOPEX / whiteSpaceM2`	$/m²/yr	Real-estate efficiency metric. White space = rackCount × 3.5 m².
Energy cost	`annualKWh × energyRate`	$/yr	Typically 40–65% of total OPEX in tropical markets.
Staffing cost	FTE × burdenedRate × costMult	$/yr	Total people cost including burden.
Maintenance cost	Σ per-system rates × factors	$/yr	Six-system breakdown with age, OEM, and climate adjustments.
Effective PUE	`basePUE + coolingType.pueImpact`	ratio	Actual power overhead ratio after cooling selection.
Annual kWh consumed	`adjustedTotalKW × 8760 × util`	MWh/yr	Total facility energy draw for carbon and tariff calculations.
CO₂ tonnes/yr	`annualKWh × gridEmFactor / 1000`	t CO₂e/yr	Carbon footprint based on country grid emission factor.
P10 / P50 / P90 (Monte Carlo)	4 000-run simulation	$/yr	Optimistic / most-likely / conservative OPEX bands. P90/P10 spread shown as sensitivity indicator.
Tornado chart	`RZEngine.models.sim.tornado()`	±$/yr	Ranks energy price, PUE drift, and labor/maintenance escalation by OPEX impact.

06 Worked example

Inputs: 1 000 kW IT load · PUE 1.50 base · water-cooled chiller (pueImpact 0) · tropical-humid climate · Indonesia (energyRate $0.085/kWh, laborMult 1.0×, maintMult 1.0×) · Tier III · in-house staffing · mid-age facility · 70% utilization. Numbers reproduce the live calculator output.

Effective PUE: 1.50 + 0 = 1.50
Cooling power: 1 000 × (1.50 − 1.0) × 1.00 × 1.00 = 500 kW
Adjusted total power: 1 000 + 500 = 1 500 kW
Annual energy (kWh): 1 500 × 8 760 × 0.70 = 9 198 000 kWh/yr (9 198 MWh)
Energy cost: 9 198 000 × $0.085 = $781 830/yr
Staffing (Tier III, in-house, 1 MW): baseline ≈ 4.2 FTE/MW → 4–5 FTE × burden. Indonesia laborRate $18 000 × fxToUSD 0.000063 = negligible in direct USD; page uses IDR-based salaryData. Representative total staffing cost at 1 MW Indonesia in-house: ≈ $48 000–75 000/yr
Maintenance (mid-age, in-house, ID baseline):
Generator: 1 000 × 31.50 × 1.00 × 1.00 × 1.00 = $31 500
Cooling: 1 000 × 38.00 × 1.00 × 1.00 × 1.00 = $38 000
Electrical: 1 000 × 28.50 × 1.00 × 1.00 × 1.00 = $28 500
Fire: 1 000 × 10.00 × 1.00 × 1.00 × 1.00 = $10 000
BMS: 1 000 × 14.00 × 1.00 × 1.00 × 1.00 = $14 000
Building: 1 000 × 7.50 × 1.00 × 1.00 × 1.00 = $7 500
Total maintenance: $129 500/yr
Spare parts (Tier III, mid-age, in-house): 1 000 × $36/kW × 1.15 × 1.00 × 1.00 = ≈ $41 400/yr
Software & licensing: 1 000 × $16.50/kW × (0.70 + 0.30×1.0) = $16 500/yr
Indonesia compliance (SLO + AMDAL + BPJS + fire audit + ISO 27001, 1 MW):
(5 000+1 500) + (3 000+500) + 1 500 + (2 500+500) + (8 000+1 000) = $24 000/yr
Total annual OPEX: 781 830 + ~62 000 + 129 500 + 41 400 + 16 500 + 24 000 ≈ $1 055 000 – $1 120 000/yr
Per kW per month: $1 087 000 / 1 000 kW / 12 = ≈ $90.60/kW/month (Tier-III average band)
CO₂ footprint: 9 198 MWh × 0.709 kg CO₂/kWh (ID grid) / 1 000 = ≈ 6 521 t CO₂e/yr

Engineering reading: energy is the dominant cost (72% of total) for a tropical-market in-house-staffed facility at 1 000 kW IT. Switching to a country with lower energy rates — US-Oregon ($0.042/kWh) — would reduce energy cost by 51%, dropping total OPEX by approximately $400 000/yr at the same load. Reducing PUE from 1.50 to 1.35 via liquid cooling saves ≈ $111 000/yr in energy alone at Indonesia rates.

07 References & data sources

ASHRAE TC 9.9 (2021) — Thermal Guidelines for Data Processing Environments, 5th ed. PUE definition, cooling-type effectiveness ranges, and climate-zone thermal envelopes.
Uptime Institute (2024) — Global Data Center Survey. 4.2 FTE/MW staffing benchmark for Tier III+ critical facilities; staffing model cost differentials (in-house vs hybrid vs outsourced).
IEC 62040-3 — UPS performance classification. Maintenance rate basis for UPS/PDU/STS systems ($28.50/kW/yr).
ISO/IEC 30134-2 — PUE metric definition and measurement guidance.
PLN (2026) — Indonesian electricity tariff for industrial and large-business consumers. Basis for Indonesia energyRate $0.085/kWh.
BDEW (2026) — German electricity industry association, industrial tariff average. Basis for EU-Germany $0.235/kWh.
DEWA (2026) — Dubai Electricity and Water Authority industrial tariff. Basis for UAE $0.082/kWh.
SHRM (2024) — Society for Human Resource Management. Replacement cost multiplier 1.5× annual salary for specialised DC roles.
Indonesian MEMR / BPJS / Ministry schedules (2026) — SLO, AMDAL, BPJS TK, fire audit, and ISO 27001 compliance cost schedules.
Engine source — rz-engine.js DATA.opsBudget (models.opsBudget.opex(), models.opsBudget.staffing()) · page calculation in opex-calculator.html updateCalculation() function. Source citation: DATA.sources['opsBudget'] · asOf 2026.

08 Assumptions & limitations

The calculator assumes steady-state annual operation at the stated utilization level. Seasonal load variation, demand-charge components, time-of-use tariff spikes, and short-interval peaks are not modelled. Specific scope and limitations:

Energy: assumes a flat energy tariff ($/kWh). Demand charges ($/kW/month) can be material in some US and Japanese markets and are not included. Free-cooling hours (captured in climate zone) reduce the effective cooling load in temperate and cold zones but the model does not apply these as a direct reduction — they are reflected through the coolingEnergyMult applied to the cooling-overhead fraction.

Staffing: FTE estimates use Uptime Institute 4.2 FTE/MW as a benchmark — actual headcount depends on operating philosophy, automation level (DCIM), and local labour law (shift patterns, overtime caps). The 1.35 labour burden covers typical benefits and payroll taxes; jurisdictions with mandatory 13th-month pay (e.g. Indonesia) may have higher effective burden.

Maintenance: base rates are sourced from Indonesia DC industry cost surveys and calibrated to approximately IDR 1 billion/year for 2 MW generator scope (incl. fuel system and 6M/12M overhaul). OEM-contractor premiums (2–2.5×) reflect all-risk contract pricing; basic labour-only contracts will sit closer to 1.0–1.3×. Rates for other countries are scaled by country.maintMult which is a relative index, not a direct cost survey.

Scope exclusions: capital expenditure (CAPEX), depreciation, financing cost, land and building lease, bandwidth/connectivity, taxes on revenue, and one-time project costs are not included. Use the TCO projections (5-year / 10-year) in the calculator's derived-metrics panel for multi-year cost-in-use estimates including escalation (energy +3%/yr, labor +4%/yr, general +2.5%/yr).

Monte Carlo: the 4 000-run simulation is reproducibly seeded. It is a sensitivity tool, not a statistical model of market price movements. P90 should be interpreted as "a plausible high-cost scenario given the stated uncertainty ranges", not as a 90th-percentile market outcome.

This calculator and manual are engineering education and pre-design aids. Final OPEX estimates for budgeting or commercial purposes require site-specific utility quotations, contract negotiations, and local regulatory cost assessments.