Cyclops Labs

Datacenter costs, pricing and site planning

See what each site costs and earns, and where to build next.

The platform records the cost and income of every datacenter site hour by hour, flags unusual changes, tests decisions before they are made and rates new locations from public data.

  • Hourly cost and income Energy at spot prices, cooling, water, leases, staff and depreciation per site, against GPU, compute and storage income.
  • Weather, power and fuel prices Two years of real weather, day-ahead electricity prices and fuel prices feed the cost model.
  • Anomaly detection Unusual costs and income are flagged with the most likely cause.
  • Racks, nodes and capacity Load per rack and node, and the hardware to add where capacity runs short.
  • What-if scenarios Change prices, contracts, staff, weather or demand, compare with the current plan and export a PDF.
  • Site intelligence Rate any location from 0 to 10 on power, climate, connectivity, policy, talent and public acceptance.
  • Contract register PDF and Word contracts in any language are read, and their parties, dates and amounts registered.
  • Data sources and access Connectors with credentials kept outside the database, and viewer, editor and administrator roles.

This demonstration combines simulated site data with real public weather, price and location data.

Cyclops Labs Cyclops Labs

Welcome back

Sign in with your Cyclops Labs account.

Trouble signing in? Your administrator can reset your password or unlock your account.

Cyclops Labs Cyclops Labs Cyclops Labs
Sites What-if Site intelligence Data sources Contracts Notes Usage & cost
· — —

Portfolio

Datacenter input costs

What every site costs to run this hour, from every party involved (grid energy, water, cooling, leases, maintenance, integrators, employees, consultants and depreciation), and what it earns alongside.

Simulated data

Hourly history

Cost, income and profit for all sites

Range
Chart
Cost by
Income by

Click any point or bar to see that hour (or day) broken down into every component. Striped means a loss. Stacked charts split cost and income into their parts.

Cost
Income
Profit / loss

Across all sites

Cost this hour

In EUR at reference rates. Striped bars are depreciation, a non-cash cost.

← All sites

Site

—

What-if scenario

Simulated live data

Location · 25 miles around the site

Grid, talent, competitors and people

Power lines by voltage, transmission tap-ins and substations, universities, other datacenters, energy supplier offices and population pockets, from OpenStreetMap and Wikidata. Hover over a pocket for its density. Zoom one level out or three in; the map stays within the 25 miles.

Loading the area map…

Hourly cost

Cost this hour

Striped bars are depreciation, a non-cash cost. Energy is split by what uses it, so the cooling cost includes the electricity used for cooling.

Hourly history

Cost, income and profit

Range
Chart
Cost by
Income by

Click any point or bar to see that hour (or day) broken down into every component. Striped means a loss. Stacked charts split cost and income into their parts.

Cost
Income
Profit / loss

Meter readings, same period

Grid prices, power, cooling and water

Grid spot price
Facility power
Cooling load
PUE
Outside temperature
Water use

Real open data · last two years and ahead

Weather, power market and fuel

What drives this site's cooling and energy cost, from public sources. These real hours also feed the meter readings above, the cost history and the what-if projections.

Loading the weather, market and fuel data…

Floor view

Racks and nodes

Colour by
—

    Layout as CSV, in the import format: racks.csv · devices.csv

    Hardware

    Hardware and what it is sold as

    Hardware Units Sold as Capacity In use now List price Max power

    Utilisation by product, same period

    Parties

    Suppliers, utilities and integrators

    Party Category Pricing basis Term and indexation This hour Notes and actions

    People

    Employees and consultants

    Role FTE Pay basis Term Per month This hour Actions

    Assets

    Depreciation

    Asset Cost Method and age This year Book value Per hour Notes and actions

    Site intelligence

    How suitable is a location for a new datacenter?

    Enter a country, city and postcode. The tool maps 50 miles around the place and gathers open data on power, prices, climate, connectivity, politics, talent, suppliers, tax and hazards, then rates it against datacenter site-selection markers on a scale from 0 to 10.

    Compare

    Compare two locations side by side

    Drag two analyses from the history into the boxes, choose them from the list, or enter a new location to assess. Compare becomes available when both analyses have finished.

    Standard floor plan

    Typical 1.2 MW air-cooled data hall

    Hot and cold aisles on a raised floor, with perimeter cooling and 2N power. Drag to rotate, scroll to zoom, right-drag to pan; click any rack for its details.

    Scenario analysis

    What-if scenarios

    Change contract terms, energy prices or loads for one site and compare the next months with the current plan. Nothing here is saved. Your contracts, prices and meter data do not change.

    Period

    Changes

    Market and site

    Weather uses the forecast, then the chosen weather year from two years of real data; prices use the published day-ahead prices, then a projection from the last two years. Diesel and outage hours apply to sites with standby generators.

    Products: demand and price
    ProductDemand change % Price nowScenario price

    People
    OnRoleFTE Salary or day rateDays/month

    Contract terms
    OnPartyPricing Rate · quantityIndex %/y

    Result

    Current plan versus scenario

      Profit or loss per day
      Cumulative profit
      Income per day
      Cost per day

      Change in cost by category

      Change in income by product

      Contract register

      Contracts 0

      AI cost this month
      —
      All time
      —

      Drop contract files here

      PDF or Word (.docx), up to 24 MB each, in any language. Employment, consulting, supplier and vendor contracts are parsed automatically by AI.

      Choose files
      Counterparty Type Term Value Status

      Admin overview

      Notes 0
      Authors 0
      Known users 0

      New note

      You have read-only access. Ask an administrator for the app-editor role to create or change notes.

      Notes 0

      Data sources

      Where the data comes from

      Every feed behind the tracked history, and whether it is in use. Each fact has one source. Where a contract already provides a figure, the matching feed is greyed out and not ingested, so nothing is counted twice.

      Checking…

        Overview

        Every feed at every site

        By site

        Feeds

        Shared

        Feeds used by every site

        Usage & cost

        Cost of reading contracts

        Every AI call is recorded, including reparses and failed attempts. Costs are list-price estimates from token counts, so discounts and tax are not included.

        By model

        Model Calls Input tokens Output tokens Cost

        Recent calls

        When (UTC) Document Outcome Cost Tokens in / out Cache read / written Time Model
        Cyclops Labs GmbH checking…
        Loading data… Large calculations and PDF reports can take a minute or more. Please be patient: the page updates by itself.