About
The Cloud Kettle Index
The Cloud Kettle Index estimates Great Britain’s data-centre electricity load and expresses it in a familiar unit: kettle-boils per second. On bank holidays and quiet days, GB electricity demand falls by roughly 15–20%. Data-centre demand is generally much flatter than residential and commercial electricity use - so when national demand dips, data centres can represent a larger share of the grid. This site makes that contrast visible. All figures are modelled estimates based on published capacity data, not live meter readings. See the methodology page for the full formula, data sources, and what the site does not claim.
This site covers Great Britain only (England, Scotland, Wales). Northern Ireland operates a separate electricity system managed by SONI and is not included.
Prismatic Labs and vetch
The Cloud Kettle Index is built by Prismatic Labs, the team behind vetch - an open-source tool - planet-aware LLM observability with circuit breakers for runaway cost, energy, and carbon.
vetch on GitHub →Source code and licence
Open source under the Apache 2.0 licence.
github.com/prismatic-labs/cloud-kettle-index →In brief
Short
The Cloud Kettle Index is a public-interest website that estimates Great Britain’s data-centre electricity load and expresses it as kettle-boils per second. On bank holidays, GB electricity demand falls by around 15–20%. Data-centre load does not. The site estimates that data centres draw between 2,130 and 4,980 kettle-boils per second continuously - a modelled estimate based on published capacity data, not a live meter reading. Built by Prismatic Labs.
Longer
The Cloud Kettle Index is a public-interest website by Prismatic Labs that estimates Great Britain’s data-centre electricity load and translates it into a unit everyone can picture: kettle-boils per second, where one kettle-boil equals the energy a standard 3-kilowatt kettle uses to boil one litre of water from cold.
On bank holidays, Sundays, and overnight, GB electricity demand typically falls by 15–20%. Data-centre demand is generally far flatter than residential and commercial electricity use - it powers cloud services, streaming, banking, AI, and internet infrastructure around the clock. When national demand dips, data centres can represent a larger share of the grid. The Cloud Kettle Index makes this contrast visible.
The site estimates that GB colocation data centres draw between 2,130 and 4,980 kettle-boils per second on a continuous basis. This is a modelled estimate, not a live meter reading, derived from DSIT’s 2024 estimate of colocation data-centre IT capacity in Great Britain, adjusted for utilisation and Power Usage Effectiveness. Live GB electricity demand is sourced from Elexon BMRS.
Built by Prismatic Labs, which also develops vetch, an open-source tool for measuring LLM inference energy and carbon.
What this site does not claim
- It does not claim to show live metered data-centre electricity demand.
- It does not claim to measure AI-specific electricity use. Data centres power AI, cloud storage, streaming, banking, SaaS, and all internet infrastructure.
- It does not identify nearest data centres from a postcode.
- It does not claim DSIT’s 1.6 GW is a current 2026 live figure (it is an autumn 2024 estimate of colocation facilities).
- It does not include Northern Ireland (separate grid, operated by SONI).
Contact
Corrections, questions, or data enquiries: marco@prismaticlabs.ai