Free kit for data-center builds

Know the last safe day to order every long-lead item on your build.

A month of delay can cost a typical 60 MW US build as much as $14.2M, about $470K a day1, and power transformers now average 128 weeks from order to delivery2. Give this free kit your equipment list and it works out the latest safe date to order each long-lead item and flags the ones that will arrive late.

One universal skill. Works with your AI assistant.

  • Claude
  • ChatGPT
  • Codex
  • Gemini
  • Grok
  • One ZIP: the skill, a cheat sheet and an Excel tracker that needs no AI
  • Your equipment list never comes to Zinit
Get the free kit

Every field is required. Your download starts here after you submit.

Choose the closest match.

The form records your answers, contact details, attribution and tracking choice. It does not collect your equipment list.

What lands in your hands

A report, a tracker and a one-page cheat sheet.

The report and the workbook are built from your own equipment list. These previews come from the kit’s fictional 40 MW example, so every date, price and line in them is made up.

report.html
The Long-Lead Check report for the fictional 40 MW example: counts of long-lead items that will miss, are at risk or are on track, the cost of a late week, and a list of what to do first.
The report: what to order first, what will miss and what a late week costs
Data-Center-Long-Lead-Cheat-Sheet-2026.pdf
The one-page Data Center Long-Lead Cheat Sheet 2026: lead-time bars for the five bottleneck items, what a late month costs and five questions to ask a supplier.
The cheat sheet: lead times, the cost of a late month and five supplier questions on one page
Data-Center-Long-Lead-Tracker-2026.xlsx
The Your build sheet of the Excel tracker with five example lines, each with its latest safe order date, arrival date, weeks to spare and a status of on track, at risk or will miss.
The tracker: type your lines and the formulas work out each latest safe order date and status, with no AI

What it does

Six things it works out from your equipment list

Made for the procurement, supply-chain, project and construction managers and owner’s reps who plan data-center builds.

Finds the long-lead lines

Power transformers, medium-voltage switchgear, generators, UPS and chillers, picked out of an equipment list or bill of materials in CSV or XLSX.

Works out each latest safe order date

It counts back from the date you need the item on site, using the supplier’s quoted lead time or the 2026 published range plus a buffer for receiving and testing.

Flags the lines that will be late

Every long-lead line gets a status of on track, at risk or will miss, with the weeks it has to spare or the weeks it will be late.

Prices a late week for your MW

It scales STL Partners’ figure for a 60 MW build to the MW you give it and calls that an assumption, so you can swap in your own cost of a day.

Ranks bids with the delay priced in

A cheap quote that arrives late is easy to spot once its priced delay sits beside its price. It is a screening view, and the award stays your call.

Drafts your supplier questions

Held production slots, phased delivery, standard ratings, factory test dates, freight and second sources, for every line at risk.

It writes an HTML report, an Excel workbook, JSON and a short summary, and it has no network, upload or telemetry code.

Get the free kit

How it works

From your equipment list to the orders that cannot wait

Step 1

Download the kit.

One ZIP holds the AI skill, the Excel tracker, the cheat sheet and a fictional 40 MW equipment list to try first.

Step 2

Give your AI assistant your equipment list.

Add the skill to Claude, ChatGPT, Codex, Gemini, Grok or another assistant that can run Python, attach your list as CSV or XLSX and tell it your build’s IT capacity in MW, or your own cost of a day of delay.

Step 3

Act on the late lines first.

The report puts the lines that will miss at the top, each with its latest safe order date and the questions to put to the supplier.

The lead times it checks against

How long the five bottleneck items take to arrive

Weeks from purchase order to delivery, as published in 2026. The check counts back from your need-on-site date with the slow end of each range, and a lead time your supplier has quoted replaces the range.

  • Published range or average
  • Longer figure reported

  1. Power transformers

    Step-up units average 144, and the check uses 144 to stay safe

    128 weeks2

  2. Medium-voltage switchgear

    52–80 weeks3

  3. Generators over 3 MW

    90–110 weeks3

  4. UPS systems

    30–60 weeks3

  5. Chillers and chilled-water plant

    Chillers: up to 85 weeks

    48–60 weeks4

Weeks from order to delivery

What’s inside the kit

Six parts in one ZIP

  • data-center-long-lead-check/

    The skill, for any AI assistant that can run Python: its instructions, the 2026 lead-time table with sources, and one Python script your assistant runs.

  • data-center-long-lead-check.zip

    The same skill packed as one archive, for assistants that add a skill by upload, such as Claude.ai.

  • Data-Center-Long-Lead-Tracker-2026.xlsx

    An Excel tracker with sheets for your build, the cost of delay, bid scoring, the lead times and their sources. It needs no AI.

  • Data-Center-Long-Lead-Cheat-Sheet-2026.pdf

    One page to print or pin up, with the lead times, what a late month costs and five questions to ask a supplier.

  • example-bom-40mw.csv

    A fictional 40 MW phase with made-up dates, prices and two competing transformer bids, so you can see a report before you use real data.

  • README.md

    Setup for Claude, ChatGPT, Codex, Gemini CLI, Grok CLI and any other agent that can run Python, and how to run the check with Python alone.

Get the free kit

Before you download

The questions people ask first

Do I need AI to use it?

The Excel tracker works on its own in Excel, Google Sheets, Numbers or LibreOffice. You type your long-lead lines and the formulas work out the latest safe order date, the weeks to spare and a status, and the cheat sheet is a single page to print.

Which AI assistant does the skill work with?

Any AI assistant that can read a skill and run Python. The README has setup steps for Claude, ChatGPT, Codex, Gemini and Grok, and for any other agent that can read files and run Python. We ran the fictional example and three harder made-up equipment lists in Claude Code, Codex, Gemini CLI and Grok CLI, and each one produced the same results as the script on its own. ChatGPT reads the same open skill format, according to OpenAI’s documentation, though we have not run the kit there ourselves.

What happens to my equipment list?

This page never sees it, since the form asks only for your contact details. The skill’s script reads your file and writes four files where your assistant runs (your computer for a desktop agent, the assistant’s own workspace on the web), with no network, upload or telemetry code. Your AI assistant’s own account and data settings apply to anything you share with it, so check them before you give it a confidential list.

What file does it need?

An equipment list, schedule or bill of materials as CSV or XLSX, one line per row. Only a description is required, and the answer gets sharper with quantities, need-on-site dates, order dates, quoted lead times, suppliers and prices. The kit explains which column names it recognises.

Where do the lead times come from?

From figures published in 2026 by Wood Mackenzie (via MGRID), Terrapin Construction Group and DC Atlas, used as each source states them and listed with links in the kit. Lead times move with order books and factory capacity, so a quote from your supplier replaces the range, and every date is worth confirming.

How does it price a delay?

It starts from STL Partners’ figure for a month of delay on a typical 60 MW US build and scales it in a straight line to the MW you give it. That is an assumption, so if you know your own cost of a day of delay, give it that and it uses your number instead.

Is the bid ranking a recommendation?

It is a screening view that adds the priced delay to each price, so quotes can be compared on the same footing. Scope, specification, commercial terms and the supplier’s record still decide the award.

Can I try it before using my own list?

Yes, the kit includes a fictional 40 MW phase with made-up dates, prices and two transformer bids, so you can see the whole report before you share anything real.

Get the free Data Center Long-Lead Check.

Start with the fictional 40 MW example, then run your own equipment list and share the report with your project team.

Get the free kit

Your download starts here after you submit the form. No phone number required.

Sources

Figures as each source published them. Lead times move with order books and factory capacity, so confirm every date with your supplier.

  1. [1] STL Partners (March 2025), Delays in data centre construction cost developers nearly USD15mn per monthdelays can cost as much as USD 14.2M per month on a typical 60 MW US data centre
  2. [2] MGRID (April 2026), Power Transformer Lead Times of 128 Weeks Are Driving the 2025 Solid-State Transformer Commercial OpportunityWood Mackenzie’s Q2 2025 survey: power transformers averaged 128 weeks, generator step-up units 144 weeks
  3. [3] Terrapin Construction Group (June 2026, updated November 2026), Switchgear, Transformer, and Generator Lead Times in 2026medium-voltage switchgear 52–80 weeks, diesel generators over 3 MW 90–110, UPS systems 30–60
  4. [4] DC Atlas (February 2026), Data Center Supply Chain Lead Times: The Reality Behind Delivery Delayschilled-water systems 48–60 weeks (citing CAE), chillers up to 85 (citing SourceBlue)
Request a demo