Sustainable Cloud Planner Calculator

Input planner

Workload and scenario settings

Responsive form grid: 3 columns large, 2 medium, 1 mobile.

View example table

Workload (monthly)

Compute vCPU-hours

Total vCPU-hours consumed by the workload.

Storage (GB-month)

Average stored data across the month.

Data transfer (GB)

Total network egress/transfer for the month.

Energy factors (editable assumptions)

kWh per vCPU-hour

Use provider measurements when available.

Storage kWh per GB-month

Reflects storage media and redundancy choices.

Network kWh per GB

Higher for long-haul egress and heavy processing.

Cost assumptions

Currency code

Example: USD, EUR, GBP, PKR.

Cost per vCPU-hour

Use blended rate for your instance mix.

Storage cost per GB-month

Account for tiered storage if needed.

Transfer cost per GB

Often the biggest bill driver for egress-heavy apps.

Scenario A

Scenario name

Region label

Grid intensity (gCO₂e/kWh)

Lower is better for location-based emissions.

PUE

Facility overhead multiplier.

Renewable coverage (%)

Used for market-based estimate.

Average utilization (%)

Lower utilization implies more provisioned capacity.

Scenario B

Scenario name

Region label

Grid intensity (gCO₂e/kWh)

PUE

Renewable coverage (%)

Average utilization (%)

Reset

Tip: If you have provider-specific energy or emissions data, replace the default factors and grid intensity values for tighter estimates.

Example data table

Sample inputs and outputs

Field	Example value	Notes
Compute vCPU-hours	5,000	Monthly workload demand.
Storage (GB-month)	2,000	Average stored data across the month.
Transfer (GB)	3,000	Monthly network transfer.
kWh per vCPU-hour	0.050	Editable factor for compute energy.
Scenario A: PUE / Grid / Renewables / Utilization	1.40 / 475 / 20% / 45%	Representative “current” conditions.
Scenario B: PUE / Grid / Renewables / Utilization	1.20 / 250 / 80% / 65%	Representative “planned” improvements.
Typical outcome	Lower emissions and cost	Improvement depends on your inputs.

Run the calculator to generate exact kWh, emissions, cost, and deltas.

Formula used

Computation model

Provisioned capacity

provisioned_vcpu_hours = compute_vcpu_hours ÷ (utilization% / 100)

This approximates how lower utilization requires more provisioned capacity for the same workload.

IT energy

it_compute_kwh = provisioned_vcpu_hours × kWh_per_vcpu_hour
it_storage_kwh = storage_GB_month × storage_kWh_per_GB_month
it_network_kwh = transfer_GB × network_kWh_per_GB

Total IT energy is the sum of compute, storage, and network energy.

Facility energy

facility_kwh = it_total_kwh × PUE

PUE captures non‑IT overhead like cooling and power distribution.

Emissions

location_kgCO2 = facility_kwh × (grid_gCO2_per_kWh ÷ 1000)
market_kgCO2 = location_kgCO2 × (1 − renewables% / 100)

Market-based emissions are reduced by renewable coverage.

Cost estimate

compute_cost = provisioned_vcpu_hours × cost_per_vcpu_hour
storage_cost = storage_GB_month × cost_per_GB_month
network_cost = transfer_GB × cost_per_GB_transfer
total_cost = compute_cost + storage_cost + network_cost

Note: This is a planning model. For reporting-grade numbers, replace factors with your provider’s published values and your organization’s accounting rules.

How to use

Steps

Enter your monthly compute, storage, and transfer volumes.
Adjust the energy factors if you have measured or vendor data.
Fill Scenario A with your current region and operations metrics.
Fill Scenario B with your planned region or optimization targets.
Press Submit to see emissions, energy, cost, and deltas above.
Download CSV or PDF to share plans with stakeholders.

Workload Baseline and Demand

Start by estimating monthly vCPU hours, storage GB-month, and transfer GB. These three drivers represent compute, data retention, and egress. The planner converts workload demand into provisioned capacity using the utilization input. When utilization is 40%, provisioned vCPU hours become 2.5× the workload. That multiplier highlights waste from idle instances, overprovisioned clusters, and always-on test environments.

Energy Model and PUE Impact

IT energy is calculated for compute, storage, and network using editable kWh factors, then scaled by PUE to include cooling and power distribution. PUE is a facility overhead multiplier, so improvements reduce all IT energy categories. For example, reducing PUE from 1.40 to 1.20 lowers facility energy by about 14.3%. Combine lower PUE with autoscaling and efficient instance families to compound savings.

Emissions Accounting Choices

Location-based emissions follow the grid intensity in gCO2e per kWh, reflecting where electricity is generated. Market-based emissions apply renewable coverage to represent contractual instruments or matched clean energy. If grid intensity drops from 475 to 250 gCO2e/kWh, location emissions nearly halve at the same energy level. Increasing renewables from 20% to 80% further reduces market emissions for planning scenarios.

Cost and Optimization Levers

Monthly cost is estimated from provisioned vCPU hours, storage GB-month, and transfer GB, using blended unit rates you provide. Compare scenario deltas to prioritize initiatives with the best combined value. If network dominates energy, reduce egress with compression, caching, and CDNs. If compute dominates, right-size, schedule batch jobs, and turn off idle workloads. Storage-heavy systems benefit from lifecycle policies and tiering.

Governance and Reporting Cadence

Use the sustainability score to communicate progress across teams, combining grid intensity, PUE, renewable coverage, and utilization into one signal. Set targets per service, review results quarterly, and update factors when vendors publish better measurements. Export CSV for spreadsheets and PDF for leadership briefings. Keep assumptions documented, especially grid factors and renewable matching, to support consistent decision-making. For mature programs, segment results by environment, business unit, and region to reveal hotspots. Pair carbon KPIs with reliability and latency SLOs so efficiency does not erode user experience. When a change is approved, re-run the planner to confirm expected savings before implementation and after.

FAQs

What does the planner estimate?

It estimates facility energy, location-based emissions, market-based emissions, and monthly cost for two scenarios, using your workload demand, utilization, PUE, grid intensity, renewables coverage, and unit rates.

How should I choose the energy factors?

Start with vendor measurements when available. Otherwise use conservative defaults and update as you learn. Keep factors consistent across comparisons so scenario deltas reflect decisions, not changing assumptions.

Why are there two emissions numbers?

Location-based follows the regional grid intensity and reflects where electricity is produced. Market-based applies renewable coverage to reflect contractual clean energy claims. Many reporting frameworks ask for both views.

How do I reduce emissions fastest?

First reduce wasted capacity by improving utilization and autoscaling. Next pick lower-carbon regions or providers and target lower PUE facilities. Finally increase renewable coverage for market-based reductions, especially for workloads that must stay in one region.

What inputs matter most for cost?

Provisioned vCPU hours and data transfer often dominate. Low utilization increases compute spend, while high egress inflates network charges. Use blended rates from your bills and compare scenario deltas to prioritize the biggest drivers.

Can I use this for reporting-grade disclosures?

Use it for planning and prioritization. For disclosures, replace factors with audited provider data, align boundaries with your inventory method, and document grid sources and renewable instruments. Keep evidence for each assumption and refresh values regularly.