Purpose

This tool helps you see the physical and environmental scale of modern AI data centres by drawing an illustrative footprint on any map location you choose — your neighbourhood, workplace, or a searched place — so you can compare site size, hall area, and water use in familiar surroundings.

Suggestion - Try it on your own neighbourhood. It will very likely seem huge. But also try it outside built-up areas. "As big as a village" is one way to look at it. "As big as a golf course" is another. "As big as an airport" is yet another. They all trigger different mental associations.

Of course, this doesn't help visualise energy or water use -- for that you'll need a different tool. Nor does it help visualise the utility of a datacentre -- it could be making slop pictures, it could be animating the next blockbuster, or it could be curing cancer. It's probably working on a mixture of these.

Map key

The overlay is a simplified representation of what a real data centre might look like - a fenced boundary, buildings containing the data halls, and a substation and grid plant providing power. Of course a real data centre may be laid out differently.

• Site perimeter — fenced campus boundary (orange outline and tint).
• Data halls — main IT / server building footprints (dark grey).
• Substation & grid plant — electrical intake and switchyard area (striped orange). Backup generators and other plant are not drawn separately on this simplified overlay.

Positioning & accuracy

The overlay is centred on the middle of the map view, not on a real facility’s coordinates. Tier footprints are illustrative representations loosely based on published examples (Kao Data Harlow, VIRTUS Stockley Park, Google Campus 2 Iowa, proposed Blackdog AI Campus). Layout blocks inside each tier are approximate — useful for scale comparison, not planning or engineering.

Metric readouts (hectares, m², m³/day) come from researched public sources where available; sources and assumptions are documented in the project repository.

How to use

• Find a place or My location — move the map to where you want to compare scale.
• Facility tier — switch between regional hub, hyperscale, mega-campus, and proposed mega-campus classes. The map stays put; only the overlay and metrics change.
• Overlay rotation — rotate the footprint on the map without changing map north (bearing).
• Pan and zoom — the overlay recentres on the viewport when you stop moving the map.
• Right-click and drag — rotate the map itself (MapLibre bearing).

How to read them

The figures shown for each tier — site area, data hall area, and water use — are order-of-magnitude estimates meant to convey scale, not precise specifications for any one facility. Treat them as "roughly this big," and use them to compare tiers and to picture the footprint against places you know. The drawn grey hall blocks are sized to match the stated data-hall area in every tier, so the shapes and the numbers tell the same story.

How we reached them

• Site area & data halls — taken from published campus sizes and technical floor areas for each tier's real-world example (Kao Data Harlow, VIRTUS Stockley Park, Google Campus 2 Iowa, proposed Blackdog AI Campus), converting any imperial source figures to metric.
• Water — Tier 1 — derived directly from the operator's reported annual potable water use (Kao Data ESG report: ~18.2 million litres/year ≈ 50 m³/day).
• Water — Tiers 2–4 — illustrative estimates for the cooling load at that power level. The population comparison divides the daily water use by the UK average mains use of ~137 litres per person per day, so "about N people" means that many residents would use a similar amount of tap water each day.

Caveats

Real facilities vary widely with cooling design, climate, and utilisation — a free-air or adiabatic site can use a small fraction of the water of an evaporative-cooled one at the same power. Figures and layouts are loosely based on public examples and are not engineering or planning data. Full sources and the arithmetic behind each number are documented in the project repository.