Cost · Carbon

Cloud Carbon & Green IT Sustainability Estimator

Cloud computers use electricity. Electricity has a carbon cost that depends on the grid. This page turns vCPUs, a region band, storage, and data transfer into a monthly footprint and a green score you can explain in a meeting.

Enter vCPUs, a region band, storage TB, transfer GB, and a renewable commitment. An in-browser AI agent shows tonnes of CO₂, a green score, and a 90-day reduction sequence. No cloud login. No API key.

Four inputs, said in plain words

vCPUs. How much compute stays on. Each one is priced at 12 watts for 730 hours in the month.
Region band. A rounded grid intensity, not a live reading from a cloud account. Europe is the lower band. Asia-Pacific is the higher band.
Storage and transfer. Terabytes sitting still, and gigabytes moving. On a typical estate these are smaller than compute.
Renewable commitment. The share you can defend with a contract or a certificate. It lowers the market-based number only. The grid number stays the same.

A worked example, before you touch a slider

200 vCPUs, United States — mixed grid band at 380 gCO₂/kWh, 50 TB stored, 5,000 GB transferred, 20% renewable.

Energy: 1,752 kWh compute + 75 kWh storage + 150 kWh transfer = 1,977 kWh. Compute is about 89% of that.
Grid footprint (location-based): 0.751 t CO₂e, which is 751 kg. That is energy times the region band.
After the renewable commitment (market-based): 0.601 t CO₂e, which is 601 kg. The commitment removes 20% of the grid number. It does not change the grid.
Green score 57, band Mixed. The score measures how clean the electricity is, against a 700 g ceiling. The tonnes measure how big the estate is.
Largest lever in this example: move the same energy to the Europe band and save about 285 kg. Raising renewable by 20 points saves about 150 kg. Turning off 20% of vCPUs saves about 107 kg.

Your estate, in round numbers

No cloud login. Type the totals you already know.

Compute (vCPUs on for the month) 200 is the example. Each vCPU uses 12 watts.

200

Region band Rounded grams of CO₂ per kilowatt-hour. Not a live grid feed. Storage (TB) 1.5 kWh per TB in the month. 50 TB is the example. Data transfer (GB this month) 0.03 kWh per GB. 5,000 GB is the example. Renewable commitment (%) Share you can defend. 20% is the example. 100% does not erase the grid number.

20%

Footprint, score, and the biggest cut

AI agent is ready. Run it, or move a slider.

Green score — How clean the electricity is

Market-based this month — After the renewable commitment

Location-based this month — The grid number, before certificates

Energy this month — Compute, storage, transfer

In plain words — Same math, said as a sentence

Top three cuts

The AI agent ranks these after it runs.

90-day sequence

The AI agent writes this after it runs.

Directional planning aid. Rounded grid bands. Not a cloud-account carbon report and not an assurance opinion.

How the model works

Compute energy = vCPUs × 12 watts × 730 hours ÷ 1,000.
Storage energy = TB × 1.5 kWh. Transfer energy = GB × 0.03 kWh.
Location-based kg = total kWh × the region grams ÷ 1,000. One tonne is 1,000 kg.
Market-based kg = location-based kg × (1 − renewable %). The grid number does not move.
Green score = 100 − (effective grams ÷ 700 × 100), held between 0 and 100. Effective grams are the region band after the renewable %. 75–100 is Cleaner. 45–74 is Mixed. Under 45 is Heavy.
Region shift reprices the same energy on the Europe band (200 g).
Renewable step adds 20 points, and stops at 100%.
Rightsize assumes 20% of vCPUs can be turned off. Colder storage assumes 40% of TB moves to a tier that uses 50% of the storage energy.

The AI agent fills this line after it runs.

Unlock a printable summary

The numbers above stay visible with no email. A work email unlocks Print / Save as PDF. Same subscribe flow as the newsletter.

FAQ

Is this my AWS, Azure, or Google carbon report?

No. This page does not log into a cloud account. You type vCPUs, a region band, storage, and transfer. The grams per kilowatt-hour are rounded planning bands, not a live grid feed and not a vendor carbon export.

How does the math work?

Compute energy is vCPUs times 12 watts times 730 hours, divided by 1,000, to get kilowatt-hours. Storage adds 1.5 kWh per TB. Transfer adds 0.03 kWh per GB. Location-based carbon multiplies that energy by the region band. Market-based carbon then keeps only the share that is not covered by the renewable commitment you typed. One tonne is 1,000 kg.

What does the green score measure?

How clean the electricity is, not how big the estate is. Effective grams are the region band after the renewable percentage. The score is 100 minus that number divided by a 700 g heavy-grid ceiling. 75 to 100 is Cleaner, 45 to 74 is Mixed, and under 45 is Heavy. The tonnes are the size. A large clean estate can score well and still show a large footprint.

Do I need an API key?

No API key is required. No OpenAI, Anthropic, or Google key. Results are directional heuristics with an AI agent experience: a badge, a short thinking step, and a stamp. The formulas run in your browser. Numbers stay on screen before any email ask. Email is only for the printable summary.

Is the AI agent calling a cloud model?

No. The AI agent on this page is the in-browser analysis experience. It uses rule-based heuristics. No vCPU count, region, or storage figure is sent to a third-party model.

Can I file this number in an ESG report?

No. The result is a directional planning aid. It is not an assurance opinion, a location-based inventory from a meter, or a filing. Use it to see whether region, renewable commitment, idle vCPUs, or colder storage moves the number, then replace it with your provider’s carbon export before any external report.