Lattice24 Savings Assessment

Enter your organisation's published electricity figures and see the energy, CO₂e and cost Lattice24 can avoid by compressing compute before it runs.

Every number on this page is an estimate. It is built from your published annual figures plus default assumptions. It is not a forecast for your site and not a quote. Two inputs are not in any annual report: how much of your load is AI/ML, and how much of that AI uses feature-heavy models. Those are sliders. Real results come from a pilot on your own hardware.

Your figures

Electricity used per year (MWh)

From your ESG or sustainability report (energy or electricity consumption).

Grid carbon intensity (kg CO₂e per kWh)

Location-based Scope 2 tonnes ÷ electricity MWh from the same report. Default 0.40 if unknown.

Power price ($ per kWh)

Carbon price ($ per tonne CO₂e)

Default: EU ETS allowance, 9 Oct 2026, at EUR/USD 1.1206. A compliance-market reference only.

We run data centres (the figure includes cooling)

AI / ML share of load:

Feature-heavy share of that AI:

Estimated result per year

Avoided energy = IT load × AI share × feature-heavy share × 93.7% (the inference-energy cut measured in Lattice24's published compression test). Value = avoided kWh × power price + avoided tonnes × carbon price. Lattice24 quotes 80% of that value, so the measured result comes in above the promise.

How Lattice24 is paid

There is no upfront fee.
Lattice24 is only paid from savings measured on your own equipment.
You keep 75% of the measured savings and Lattice24 receives 25%.
If the waste does not go down, you pay nothing.

Next step: measure it on your own data

Run the free Lattice24 assessment in your own Databricks workspace. It reads your job runs and billing, nothing leaves your account, and it takes about 10 minutes.

Download the notebook below.
In Databricks, import Lattice24_Free_Assessment.ipynb.
Type your company name in the Customer Name box and click Run all.
Wait for the green READY, then download your report from Catalog > workspace > lattice24 > reports.

Download the assessment notebook

Source code: github.com/JJardine919/lattice24-assess. Running Slurm instead of Databricks? Email us for the offline version.