What is the difference between Scope 2 location-based and market-based?

Location-based Scope 2 uses the national average grid emission factor regardless of energy procurement. Market-based Scope 2 reduces the factor in proportion to renewable energy certificates (RECs) or PPA coverage: effectiveGrid = gridIntensity × (1 − renewablePct/100). The calculator applies the market-based method when renewable coverage > 0%.

What material emission factors does the LCA use?

Concrete C30 = 150 kgCO2e/t, C40 = 250 kgCO2e/t, low-carbon GGBS blend = 100 kgCO2e/t. Structural steel: world-average (30% recycled) = 1850 kgCO2e/t, EAF (90%+ recycled) = 500 kgCO2e/t, BOF virgin = 2200 kgCO2e/t. Copper (primary) = 4100 kgCO2e/t. Aluminium (primary) = 10,000 kgCO2e/t. PVC cable = 2800 kgCO2e/t. Sources: ICE Database v4.1 and RICS/LETI DC embodied-carbon studies.

How does the tool apply regulatory compliance (SEC, EU CBAM)?

The compliance tab maps each emission scope to the relevant regulatory regime: US SEC Climate Disclosure Rule (final 2024) requires Scope 1+2 disclosure for large accelerated filers; EU CBAM (Carbon Border Adjustment Mechanism, 2026 full entry) prices embedded emissions at the EU ETS rate (~€85/tCO2e in 2025); Singapore National Carbon Tax (S$45/t from 2026) applies to Scope 1 direct emitters. The tool computes annual financial exposure under each regime using the total tCO2e from the relevant scopes.

Home / DC Solutions / Technical Manuals / Carbon Footprint Analyzer

Technical Manual · Carbon & Sustainability

Carbon Footprint Analyzer — Methodology & Formulas

Every input, equation, constant, output, and reference behind the Carbon Footprint Analyzer. All math covers GHG Protocol Scope 1/2/3 annual emissions and a full life-cycle assessment (EN 15978 stages A1–D), powered by the deterministic engine RZEngine.models.carbon anchored to locked data in DATA.carbon — no random values, no back-solved constants.

Engine RZEngine v2.5.x Basis GHG Protocol Corporate Standard LCA Basis ISO 14044 / EN 15978 Countries 32 Tabs 3 (Operational · LCA · Carbon Budget)

▶ Open the live Carbon Footprint Analyzer

01 Purpose & engineering basis

The Carbon Footprint Analyzer quantifies greenhouse-gas emissions from a data center across three boundary layers defined by the GHG Protocol Corporate Standard: direct emissions you own (Scope 1), electricity you purchase (Scope 2), and supply-chain and construction emissions you induce (Scope 3). It then extends those annual figures across the full cradle-to-grave facility lifetime via a life-cycle assessment aligned to EN 15978 (Sustainability of construction works, calculation method for environmental performance of buildings, stages A1–D).

A third tab allocates the resulting annual emissions against the host country's nationally-determined contribution (NDC) under the Paris Agreement, mapping the facility's share of the national carbon budget across sectors (AI training, inference, cloud, enterprise, government) and computing the gap to net-zero by the user's target year. This is a screening-grade planning model, not a certified corporate carbon disclosure.

02 Inputs

Inputs span three operational groups (power, location, Scope 1 sources, renewables) plus a lifecycle group for the LCA tab and a budget-allocation group for the Carbon Budget tab.

Tab 1 — Operational

Field / ID	Symbol	Unit	Range / default	Meaning
IT Load `itLoad`	`P_IT`	kW	50–100 000 · default 1 000	Rack-level IT power draw (servers + storage + network). Does not include cooling or power-distribution overhead.
PUE `pueValue`	`PUE`	ratio	1.1–2.5 · default 1.58	Total facility power ÷ IT power. Multiplies IT load to give facility power. Industry average 1.58 (Uptime 2026).
Country `countrySelect`	—	ISO-2 enum	32 countries	Selects grid emission factor and electricity rate from `COUNTRIES[]` (patched at runtime from `RZEngine.data.countries`).
Grid intensity `gridIntensity`	`EF_grid`	kgCO₂e/kWh	read-only · auto-filled	National average emission factor for purchased electricity. Source: IEA 2026 + Ember. Range 0.02 (SE) to 0.90 (ZA).
Generator fuel `genType`	—	enum	diesel / hvo	Standard diesel 2.68 kgCO₂/L; HVO (hydrotreated vegetable oil) 0.54 kgCO₂/L (~80% lower).
Generator test hours `genHours`	`H_gen`	h/yr	50–500 · default 200	Annual hours of generator operation including monthly testing, load-bank tests, and unplanned outages.
Cooling type `coolingType`	—	enum	air_crac / inrow / rdhx / dlc	Selects refrigerant GWP and charge rate. R-410A (GWP 2088), R-134a (GWP 1430), R-1234ze (GWP 7).
Annual leak rate `leakRate`	`L_r`	%	1–15 · default 5	Fraction of total refrigerant charge that escapes annually. Industry average 5–10% for DX; 2–3% for chilled-water plants.
Renewable strategy `renewableType`	—	enum	none / solar / solar_bess / ppa	Procurement mechanism. Determines co-investment note only; the emission reduction is applied via `renewablePct`.
Renewable coverage `renewablePct`	`R`	%	0–100 · default 0	Percentage of facility electricity matched by renewable energy or certificates. Applied to derive market-based Scope 2.

Tab 2 — Life-Cycle Assessment

Field / ID	Unit	Default	Meaning
Facility size `facilityMW`	MW IT	linked to itLoad	Total IT capacity; scales material quantities, floor area, and server counts per MW.
Building type `buildingType`	enum	purpose	Multiplier on material quantities: purpose-built = 1.0, converted warehouse = 0.70, modular = 0.85.
Concrete grade `concreteGrade`	enum	c30	C30 = 150 kgCO₂e/t; C40 = 250 kgCO₂e/t; low-carbon GGBS blend = 100 kgCO₂e/t. ICE Database v4.1.
Steel sourcing `steelSource`	enum	world_avg	World-average (30% recycled) = 1850 kgCO₂e/t; EAF (90%+) = 500 kgCO₂e/t; BOF virgin = 2200 kgCO₂e/t.
Seismic zone `seismicZone`	0–4	0	Structural mass multiplier: Zone 0 = 1.0×, Zone 1 = 1.05×, Zone 2 = 1.12×, Zone 3 = 1.25×, Zone 4 = 1.40×.
Workload type `workloadType`	enum	traditional	Sets server embodied carbon (kgCO₂e) and density (servers/MW): traditional 1200/1250, cloud 1200/1000, H100 4000/140, B200 3500/200.
Server refresh cycle `serverRefresh`	yr	4	Replacement period; drives number of IT refresh cycles over project lifetime.
Network refresh `networkRefresh`	yr	5	Network-switch replacement cycle.
Project lifetime `projectLifetime`	yr	20	Total facility operational life for LCA amortisation and lifetime operational total.
Battery chemistry `batteryType`	enum	vrla	UPS battery: VRLA = 68 kgCO₂e/kWh (5yr life); LFP = 62 kgCO₂e/kWh (10yr); NMC = 74 kgCO₂e/kWh (8yr).
Battery runtime `batteryRuntime`	min	10	UPS hold-time at full IT load; sets total battery kWh = facilityMW × 1000 × (runtime/60) × redundancyMult.
Redundancy `redundancyLCA`	enum	n1	N = 1.0×, N+1 = 1.25×, 2N = 2.0×, 2N+1 = 2.25× on battery and MEP quantities.
Equipment origin `equipOrigin`	enum	regional	Transport multiplier on A4+A5: local (<500 km) = 0.8×, regional = 1.0×, international = 1.5×.
Construction duration `constructDuration`	months	18	Scales A4+A5 construction-activity carbon linearly vs the 18-month reference basis.
Decommission plan `decommPlan`	enum	demolition	End-of-life multiplier on C1–C4: demolition = 1.0×, refurbishment = 0.5×, deconstruction/reuse = 0.2×.
Steel recycle % `steelRecycle`	%	92	Module-D recycling credit on structural steel (92% industry average).
Copper recycle % `copperRecycle`	%	85	Module-D recycling credit on copper cabling and busbars.

03 Calculation methodology

All emission calculations follow GHG Protocol Corporate Standard (2015 edition) for Scope 1/2/3 boundaries, and EN 15978:2011 / ISO 14044:2006 for LCA stages. Function names map one-to-one to RZEngine.models.carbon in rz-engine.js.

Scope 2 — Purchased grid electricity

facilityKWh = P_IT × PUE × 8 760 h/yr effectiveEF = EF_grid × (1 − R / 100) [market-based] Scope 2 = (facilityKWh / 1 000) × effectiveEF [tCO₂e/yr] In RZEngine.models.carbon.annualTonnes(mw, pue, region). Market-based method: renewable coverage R% directly reduces the effective emission factor. Location-based = R set to 0. GHG Protocol · IEA 2026

Scope 1a — Generator diesel / HVO combustion

fuelConsumed = P_IT [kW] × 0.3 L/kW·h × H_gen [h/yr] [L/yr] genEmissions = fuelConsumed × EF_fuel / 1 000 [tCO₂e/yr] EF_fuel: diesel 2.68 kgCO₂/L (EPA 2024); HVO 0.54 kgCO₂/L (~80% life-cycle reduction). Diesel consumption rate 0.3 L/kW·h is the engine constant DIESEL_CONSUMPTION_RATE. EPA Emission Factors 2024 · GHG Protocol

Scope 1b — Fugitive refrigerant leakage

totalCharge = P_IT [kW] × chargeKgPerKw [kg refrigerant] refrigEmissions = totalCharge × (L_r / 100) × GWP / 1 000 [tCO₂e/yr] chargeKgPerKw and GWP are per-cooling-type constants: air_crac 0.30 kg/kW / GWP 2088; inrow 0.20 / 2088; rdhx 0.15 / 1430; dlc 0.05 / 7. GWP₁₀₀ values: IPCC AR4 (sitewide-consistent). GHG Protocol · IPCC AR4 · EPA GreenChill

Scope 3 simplified (operational tab)

Scope 3_ops = P_IT [kW] × 0.15 tCO₂e/kW [tCO₂e/yr upstream] A screening upstream-supply-chain factor used on the operational tab only (0.15 tCO₂e/kW IT = ~15 kgCO₂e per MWh IT). The LCA tab provides a full cradle-to-gate Scope 3 breakdown with material-level emission factors. GHG Protocol Scope 3 Standard · RICS/LETI

LCA — Stage A1–A3: Material production (embodied carbon)

concreteCarbon = facilityMW × 1 000 t/MW × EF_concrete × buildMult × seismicMult / 1 000 steelCarbon = facilityMW × 200 t/MW × EF_steel × buildMult × seismicMult / 1 000 copperCarbon = facilityMW × 20 t/MW × 4 100 kgCO₂e/t / 1 000 aluminumCarbon = facilityMW × 10 t/MW × 10 000 kgCO₂e/t / 1 000 pvcCarbon = facilityMW × 7 t/MW × 2 800 kgCO₂e/t / 1 000 materialsTotal = concreteCarbon + steelCarbon + copperCarbon + aluminumCarbon + pvcCarbon [tCO₂e] All results in tCO₂e. EF_concrete: C30 150, C40 250, low-carbon 100 kgCO₂e/t. EF_steel: world-avg 1850, EAF 500, BOF 2200 kgCO₂e/t. Quantities per MW IT: concrete 1 000 t, steel 200 t, copper 20 t, aluminium 10 t, PVC 7 t — from RICS/LETI DC embodied-carbon studies. buildMult: purpose 1.0, converted 0.7, modular 0.85. seismicMult: zone 0–4 = [1.0, 1.05, 1.12, 1.25, 1.40]. ICE Database v4.1 · RICS/LETI · EN 15978

LCA — IT equipment embodied carbon (Scope 3 upstream)

totalServers = facilityMW × serversPerMW serverCycles = ceil(projectLife / serverRefresh) networkCycles = ceil(projectLife / networkRefresh) itEmbodied = ( totalServers × EF_server × serverCycles + totalServers × 0.05 × 450 kgCO₂e × networkCycles + totalServers × 0.02 × 2 000 kgCO₂e × serverCycles ) / 1 000 [tCO₂e] EF_server per workload: traditional 1 200 kgCO₂e (1 250 servers/MW), cloud 1 200 (1 000/MW), H100 GPU 4 000 (140/MW), B200+ GPU 3 500 (200/MW). Network switches at 5% of server count × 450 kgCO₂e; storage arrays at 2% × 2 000 kgCO₂e. Dell/HPE/NVIDIA published LCA · GHG Protocol Scope 3

LCA — UPS battery embodied carbon

batteryKWh = facilityMW × 1 000 × (batteryRuntime / 60) × redundancyMult batteryCarbon = batteryKWh × EF_battery / 1 000 [tCO₂e/cycle] batteryReplacements = ceil(projectLife / battChemLife) totalBatteryCarbon = batteryCarbon × batteryReplacements [tCO₂e] EF_battery: VRLA 68 kgCO₂e/kWh (5yr life); LFP 62 (10yr); NMC 74 (8yr). redundancyMult: N=1.0, N+1=1.25, 2N=2.0, 2N+1=2.25. PEFCR Battery (2020) · IVL Swedish Environment Institute

LCA — MEP equipment, construction activity, end-of-life (A4–C4)

mepCarbon = facilityMW × 150 tCO₂e/MW (UPS + cooling + gensets + switchgear) floorArea = facilityMW × 500 m²/MW constructCarbon = floorArea × 40 kgCO₂e/m² × transportMult × (duration / 18) / 1 000 endOfLifeCarbon = floorArea × 15 kgCO₂e/m² × decommMult / 1 000 transportMult: local 0.80, regional 1.00, international 1.50. decommMult: demolition 1.0, refurbishment 0.5, deconstruction 0.2. Construction activity factor 40 kgCO₂e/m² covers plant diesel, formwork, site waste (A5). RICS Embodied Carbon Primer 2020 · EN 15978

LCA — Module D: End-of-life recycling credits

steelCredit = facilityMW × 200 t × (steelRecycle% / 100) × 1 500 kgCO₂e/t / 1 000 copperCredit = facilityMW × 20 t × (copperRecycle% / 100) × 3 000 kgCO₂e/t / 1 000 alCredit = facilityMW × 10 t × 0.85 × 9 000 kgCO₂e/t / 1 000 moduleDCredit = −(steelCredit + copperCredit + alCredit) [tCO₂e, negative] Recycling credit = avoided primary production. Factors: steel 1 500, copper 3 000, aluminium 9 000 kgCO₂e/t saved. Aluminium recycled fraction fixed at 85% (industry average). Module D is reported separately and does not reduce the A1–C4 total. ICE Database v4.1 · EN 15978 Module D

Carbon budget allocation (Tab 3)

dcSectorEmissions = nationalTotal × (dcShare% / 100) [MtCO₂e/yr] targetDcEmissions = dcSectorEmissions × (1 − ndcReduction) cumulativeBudget = (dcSectorEmissions + targetDcEmissions) / 2 × budgetYears [linear] annualBudgetPerMW = cumulativeBudget / budgetYears / nationalMW × 1 000 [tCO₂/MW/yr] NDC reduction = |ndcTarget%| / 100. nationalMW estimated as dcSectorEmissions / (0.475 kgCO₂/kWh × 8.76 × 1.58). Three budget methodologies: linear trajectory, front-loaded (×0.85), back-loaded (×1.15). UNFCCC NDC Registry · Paris Agreement Art. 4

Carbon offset and compliance cost

offsetCost = totalEmissions [tCO₂e] × $35 /tCO₂e [voluntary market] taxExposure = totalEmissions × $65 /tCO₂e [EU ETS 2025 proxy] In RZEngine.models.carbon.offsetCost(tonnes) and the inline EU ETS exposure. DATA.carbon.offsetPrice = $35 (voluntary market 2026 blend; reconciles DCMOC $45 vs legacy $18). EU ETS proxy $65 = ~€60/tCO₂ (2025 forward). Singapore NCCS S$45/t from 2026 also surfaced. World Bank Carbon Pricing Dashboard · OECD ECR 2025 · DATA.sources['carbon']

04 Constants & data sources

All grid factors and carbon prices are locked in DATA.carbon with provenance tags in DATA.sources['carbon']. Material emission factors live in the inline LCA constant in carbon-footprint.html, crossreferencing ICE Database v4.1 and RICS/LETI studies.

Grid emission factors (kgCO₂e/kWh) — `DATA.carbon.gridFactor`

Region	EF (kgCO₂e/kWh)	Notes
US	`0.37`	US EIA national average (coal+gas+nuclear+renewables mix)
EU	`0.23`	EU27 average (high nuclear FR/SE, coal-heavy PL pulled)
APAC	`0.55`	Blended Asia-Pacific (China/India dominate)
LATAM	`0.20`	High hydro share (BR, CO, CL)
ID (Indonesia)	`0.68`	Coal-heavy PLN grid; per-country COUNTRIES[] = 0.70
SG (Singapore)	`0.41`	CCGT gas-dominant grid; per-country = 0.40
JP (Japan)	`0.47`	Post-Fukushima gas/coal, nuclear restart partial
IN (India)	`0.63`	Coal-heavy; per-country = 0.72
MY (Malaysia)	`0.55`	Gas+coal Peninsular; per-country = 0.60
FR (France)	`0.06`	~70% nuclear (per-country table)
SE (Sweden)	`0.02`	Hydro + wind dominant
ZA (South Africa)	`0.90`	Coal-dominant Eskom grid

Source: DATA.sources['carbon'] = 'IEA 2026 + Ember grid intensity + ICAP carbon prices', asOf 2026.

At runtime, carbon-footprint.html patches per-country values from RZEngine.data.countries[id].environment.gridCarbonIntensity (the single source of truth generated from dcmoc/src/constants/countries.ts). The 32-country per-country table shown in the tool supersedes the 9-region DATA.carbon.gridFactor blends for the calculator.

Carbon prices ($/tCO₂e) — `DATA.carbon.carbonPrice`

Region	Price ($/tCO₂e)	Regime
US	`40`	Blended state-level compliance + RGGI + voluntary
EU	`85`	EU ETS 2025 (compliance, ~€80/t converted)
APAC	`25`	Blended SEA/APAC compliance schemes
LATAM	`15`	Colombia/Chile voluntary + compliance blend
ID	`12`	Indonesia ETS pilot + voluntary market
SG	`18`	NCCS S$25/t 2024 (S$45/t from 2026)
JP	`30`	Japan GX-ETS + J-Credit scheme
IN	`10`	India PAT + voluntary carbon market
MY	`14`	Bursa Carbon Exchange + voluntary blend

Source: DATA.sources['envCosts'] = 'World Bank Carbon Pricing Dashboard + OECD ECR 2025 + NCCS Singapore', asOf 2026-07.

Material emission factors — `LCA` constant

Material	kgCO₂e/t	Quantity/MW IT	Source
Concrete C30	`150`	1 000 t	ICE Database v4.1, RICS/LETI DC
Concrete C40	`250`	1 000 t	ICE Database v4.1
Concrete low-carbon (GGBS 50%)	`100`	1 000 t	RICS/LETI
Steel — world average (30% recycled)	`1 850`	200 t	ICE Database v4.1
Steel — EAF (90%+ recycled)	`500`	200 t	ICE Database v4.1
Steel — BOF virgin	`2 200`	200 t	ICE Database v4.1
Copper (primary)	`4 100`	20 t	ICE Database v4.1
Aluminium (primary)	`10 000`	10 t	ICE Database v4.1
PVC cable	`2 800`	7 t	ICE Database / Ecoinvent 3.9

Engine model — `DATA.carbon` anchors

Constant	Value	Unit	Notes
`DATA.carbon.embodiedPerMw`	3 200	tCO₂e/MW	Bulk embodied carbon for engine's `models.carbon.embodiedTonnes(mw)`. Covers concrete+steel+MEP screening estimate. RICS/LETI DC studies.
`DATA.carbon.dieselKgCo2PerL`	2.68	kgCO₂/L	EPA Emission Factors 2024.
`DATA.carbon.genTestHoursPerYear`	52	h/yr	Engine default (1 h/week monthly test). Calculator uses user-provided 50–500 h.
`DATA.carbon.refrigerantLeakTco2ePerMwYr`	8	tCO₂e/MW·yr	Engine Scope 1 screening; GHG Protocol + EPA GreenChill typical ranges.
`DATA.carbon.offsetPrice`	35	$/tCO₂e	Voluntary market 2026 blend (nature-based + engineered); reconciles DCMOC $45 vs legacy $18. `DATA.sources['carbon.offsetPrice']`.

05 Outputs

Output	Formula / source	Unit	Interpretation
Annual total emissions	`scope1 + scope2 + scope3`	tCO₂e/yr	GHG Protocol total boundary. Primary KPI on the operational tab.
Scope 1	`genEmissions + refrigEmissions`	tCO₂e/yr	Direct on-site emissions: generator fuel combustion + fugitive refrigerant leaks.
Scope 2 (market-based)	`annualMWh × effectiveEF`	tCO₂e/yr	Purchased electricity adjusted for renewable certificates/PPAs.
Scope 3 (simplified)	`P_IT × 0.15`	tCO₂e/yr	Upstream IT supply chain screening factor (operational tab only).
Carbon intensity	`totalEmissions × 1000 / (P_IT × 8760)`	kgCO₂e/kWh IT	Efficiency grade: A+ <0.1, A <0.2, B <0.4, C <0.6, D <0.8, E <1.0, F ≥1.0.
Offset cost	`totalEmissions × $35`	$/yr	Voluntary market neutralisation cost at 2026 blend price.
Tax exposure	`totalEmissions × $65`	$/yr	Indicative EU ETS-proxy compliance cost (not the applicable rate outside EU).
LCA embodied (A1–A3)	`materialsTotal`	tCO₂e	One-time cradle-to-gate: concrete + steel + copper + aluminium + PVC.
IT embodied (Scope 3)	`itEmbodied`	tCO₂e	Lifetime server+network+storage manufacturing emissions across all refresh cycles.
Battery carbon (total)	`batteryCarbon × batteryReplacements`	tCO₂e	UPS battery manufacturing over project life (all replacement cycles).
Module D credit	`−(steelCredit + copperCredit + alCredit)`	tCO₂e	End-of-life recycling benefit — negative value, reported separately from A1–C4 total per EN 15978.
Lifetime operational	`totalEmissions × projectLife`	tCO₂e	Cumulative operational footprint over the facility lifetime.
Lifetime total (LCA + ops)	`totalLCA + lifetimeOperational`	tCO₂e	Cradle-to-grave total: all construction, IT, and operational emissions over project life.
Annual budget per MW	NDC derivation (see § 03)	tCO₂/MW/yr	National DC-sector carbon budget allocated per MW of installed capacity.

06 Worked example

Default inputs: 1 000 kW IT, PUE 1.58, United States (grid 0.40 kgCO₂e/kWh), diesel generator 200 h/yr, air_crac cooling (R-410A, GWP 2088, 0.30 kg/kW, 5% leak), no renewables, 1 MW facility, purpose-built, concrete C30, steel world-average, traditional workload, VRLA battery 10 min N+1, 20-year life. All numbers verified by node -e against the engine logic.

Facility energy: 1000 kW × 1.58 PUE × 8760 h = 13 840.8 MWh/yr
Scope 2: 13 840.8 MWh × 0.40 kgCO₂/kWh = 5 536 tCO₂e/yr
Scope 1a — generator diesel: 1000 kW × 0.30 L/kW·h × 200 h × 2.68 kgCO₂/L / 1000 = 160.8 tCO₂e/yr
Scope 1b — refrigerant: total charge 1000 × 0.30 = 300 kg; leak 300 × 0.05 × 2088 / 1000 = 31.3 tCO₂e/yr; Scope 1 total 192 tCO₂e/yr
Scope 3 (simplified): 1000 kW × 0.15 = 150 tCO₂e/yr
Annual total: 192 + 5536 + 150 = 5 878 tCO₂e/yr → Carbon intensity 5878 × 1000 / (1000 × 8760) = 0.671 kgCO₂e/kWh IT → Grade D (below average)
Offset cost: 5878 × $35 = $205 730/yr
LCA A1–A3 materials (1 MW, C30, world-avg steel): concrete 1000 t × 150 = 150 t, steel 200 t × 1850 = 370 t, copper 20 t × 4100 = 82 t, aluminium 10 t × 10 000 = 100 t, PVC 7 t × 2800 = 20 t → 722 tCO₂e embodied (A1–A3)
IT embodied (traditional, 20yr / 4yr refresh = 5 cycles): 1 250 servers × 1 200 kgCO₂e × 5 cycles + 62.5 switches × 450 × 4 cycles + 25 storage × 2 000 × 5 cycles = 7 863 tCO₂e over life
Battery carbon (VRLA, 10 min, N+1, 5yr replace → 4 cycles): 1000 kW × (10/60) min × 1.25 × 68 kgCO₂e/kWh / 1000 × 4 cycles = 56.7 tCO₂e
MEP + construction + EOL: 150 + 20.0 + 7.5 = 177.5 tCO₂e
Module D recycling credits (steel 92%, copper 85%, Al 85%): steel 200 × 0.92 × 1500 = 276 t, copper 20 × 0.85 × 3000 = 51 t, Al 10 × 0.85 × 9000 = 76.5 t → −404 tCO₂e (benefit)
Total LCA (excl. ops): 722 + 7863 + 56.7 + 177.5 − 404 = 8 415 tCO₂e
Lifetime operational (20yr): 5878 × 20 = 117 560 tCO₂e
Cradle-to-grave total: 8415 + 117 560 = 125 975 tCO₂e over 20 years

Reading: for this typical 1 MW colo in the US, Scope 2 electricity dominates at ~94% of annual emissions. Switching to 100% renewable PPA reduces Scope 2 to near zero, cutting the annual total to ~342 tCO₂e and flipping the grade from D to A+. On a lifetime basis, IT equipment manufacturing (7 863 t) exceeds the entire building structure (722 t) by ~11×, reflecting the embodied-carbon intensity of AI/GPU hardware. Low-carbon steel (EAF) would reduce the steel line from 370 t to 100 t — a 73% saving on structural steel alone.

07 References & standards

GHG Protocol (2015) — Corporate Accounting and Reporting Standard (Revised): Scope 1/2/3 boundary definitions, market-based vs location-based Scope 2, REC treatment. ghgprotocol.org
ISO 14064-1:2018 — Specification with guidance at the organisation level for quantification and reporting of GHG emissions and removals.
ISO 14044:2006 + AMD 1:2017 — Environmental management — Life cycle assessment — Requirements and guidelines.
EN 15978:2011 — Sustainability of construction works: calculation method for the environmental performance of buildings; defines stages A1–A5 (product + construction), B (operational), C (end-of-life), D (beyond system boundary).
IEA World Energy Statistics 2026 — National electricity emission intensity factors (kgCO₂e/kWh). Source for DATA.carbon.gridFactor and COUNTRIES[].grid.
Ember Climate Grid Intensity 2026 — Real-time country grid emission factors used to cross-check IEA values. DATA.sources['carbon'].
ICAP Emissions Trading Worldwide 2026 — Carbon prices by jurisdiction. Source for DATA.carbon.carbonPrice.
ICE Database v4.1 (RICS/Circular Ecology) — Embodied carbon factors for construction materials: concrete, steel, copper, aluminium, PVC.
RICS Professional Statement: Whole Life Carbon Assessment (2023) — DC embodied carbon benchmarks; quantities per MW IT.
LETI Embodied Carbon Primer (2020) — Low-carbon material specification (GGBS concrete, EAF steel targets).
EPA Emission Factors for Greenhouse Gas Inventories (2024) — Stationary combustion: diesel 2.68 kgCO₂/L. Source for DATA.carbon.dieselKgCo2PerL.
IPCC AR4 (2007) — GWP₁₀₀ values: R-410A 2088, R-134a 1430, R-1234ze 7. Sitewide-consistent (matches published values in DATA.refrigerants).
IVL Swedish Environment Institute (2023) — Battery LCA emission factors: VRLA 68, LFP 62, NMC 74 kgCO₂e/kWh. Basis for LCA.BATTERY_* constants.
UNFCCC NDC Registry (2024) — Nationally Determined Contributions: national emission totals, reduction targets, and target years. Source for NDC_DATA (32 countries).
Paris Agreement Art. 4 (2015) — Net-zero trajectory and budget methodology basis for the Carbon Budget tab.
Engine source — rz-engine.js DATA.carbon + RZEngine.models.carbon (gridFactor / annualTonnes / embodiedTonnes / annualCost / offsetCost / scopes).

08 Assumptions & limitations

Scope boundaries: Scope 2 is calculated on the market-based method only (location-based = renewable coverage set to 0%). Scope 3 upstream on the operational tab is a screening factor (0.15 tCO₂e/kW IT); the LCA tab provides a more granular Scope 3 upstream via per-material emission factors and IT-equipment LCAs, but does not include purchased goods and services beyond IT hardware, business travel, waste, or water.

Grid factors: National average emission factors from IEA/Ember do not represent the marginal emission rate of an additional data-center load on any specific grid or hour. Time-of-use or hourly matching (24/7 CFE methodology) is not modelled. For high-renewable grids (France, Sweden), the national average already reflects near-zero grid intensity.

Material quantities per MW: Concrete 1 000 t/MW, steel 200 t/MW, copper 20 t/MW, aluminium 10 t/MW, PVC 7 t/MW are industry planning estimates from RICS/LETI DC embodied-carbon studies. They are screening-grade benchmarks; actual quantities depend on structural design, site conditions, and equipment specification.

IT embodied carbon: Server-embodied values (1 200–4 000 kgCO₂e/unit) are from published manufacturer LCA declarations or peer-reviewed studies. AI GPU server values (H100 4 000, B200 3 500 kgCO₂e) are estimates based on the Dell/NVIDIA hardware and will improve as manufacturers publish more detailed LCA reports.

Regulatory compliance: Carbon tax exposure is computed using the EU ETS proxy ($65/tCO₂e) as an indicative figure only. Actual compliance obligations depend on the jurisdiction, emission category, facility classification, and applicable scheme thresholds. SEC Climate Disclosure, EU CBAM, and Singapore NCCS exposures cited in the tool are informational only and do not constitute legal or compliance advice.

Carbon budget allocation: The NDC-derived per-MW budget is a macro screening estimate derived from national total emission budgets and a static DC-sector share percentage. It does not account for electricity generation emissions already included in Scope 2, temporal emission pathways within each country's sector, or sub-national policy variations.

This calculator is an engineering education and pre-design planning aid. Carbon footprints for corporate GHG reporting under GHG Protocol, ISO 14064, SEC Rule, or EU taxonomy must be prepared or validated by a qualified sustainability professional with access to site-specific metered data and certified grid factors.