AI data centers need more than GPUs.
Electricity must reach every rack without interruption. Heat must leave the chips just as quickly.
That creates demand for switchgear, UPS systems, batteries, busways, power distribution, chillers, coolant distribution units, and monitoring software.
Companies such as Vertiv, Schneider Electric, Eaton, and ABB sell different parts of this infrastructure.
Strong demand, however, does not appear in every financial number at the same time.
Backlog shows demand. Revenue shows delivery. Margins and cash show whether the demand created value.
This article explains the equipment first. It then follows the money from a possible project to an order, backlog, revenue, profit, and cash.
The Picks and Shovels Inside an AI Data Center
A data center power system has several layers. Each layer performs a different job.
Grid Connection and Switchgear
Electricity enters through a utility connection and substation. Transformers reduce the voltage. Switchgear protects the site and controls where power flows.
UPS Systems and Batteries
A short interruption can stop servers. An uninterruptible power supply, or UPS, keeps power flowing while the grid recovers or backup generation starts.
Busways and Power Distribution
Busways and power distribution units move electricity from the electrical room to rows of server racks. They also measure and control the load.
Air and Liquid Cooling
Computing turns electricity into useful work and heat. Traditional facilities use air cooling and chilled water. Dense AI racks increasingly add direct-to-chip liquid cooling.
A coolant distribution unit, or CDU, controls the liquid moving between the facility and the server equipment.
Monitoring and Service
Software and sensors track voltage, battery health, temperature, and coolant flow. Maintenance and replacement work can continue long after the original construction project ends.
Four Companies With Different Business Exposure
These companies compete in some markets. They are not the same type of business.
Vertiv
Vertiv has direct exposure to critical digital infrastructure. It sells UPS systems, power distribution, batteries, thermal systems, liquid cooling, modular infrastructure, and services.
Because data centers are central to its business, changes in AI infrastructure spending can appear clearly in orders and results.
Schneider Electric
Schneider Electric is a broader energy-management company. Its data center offer includes electrical distribution, UPS systems, cooling, prefabricated modules, controls, software, and field service.
Data centers can drive growth, but the company also serves factories, buildings, utilities, and other infrastructure.
Eaton
Eaton supplies switchgear, breakers, busways, UPS systems, power distribution, and monitoring equipment.
Its 2026 acquisition of Boyd Thermal expanded its position in thermal management and liquid-cooling components.
ABB
ABB supplies electrification and automation equipment. Its role can include switchgear, circuit protection, power distribution, controls, and higher-power electrical systems.
ABB’s Electrification business also serves utilities, transport, buildings, and industry. Its backlog is therefore not a pure data center number.
Follow the Financial Chain
A project moves through several stages before it creates cash for the supplier.
| Stage | Simple meaning | What can go wrong |
|---|---|---|
| Pipeline | Projects the company may win | No signed customer order |
| Orders | New customer demand accepted during the period | Terms and definitions differ by company |
| Backlog | Accepted orders not yet recognized as revenue | Delivery can be delayed, changed, or canceled |
| Revenue | Equipment or work delivered during the period | It records execution, not new demand |
| Margin | Profit left after producing and delivering the work | Labor, materials, tariffs, or fixed prices can reduce it |
| Cash | Money collected after operations and investment | Inventory and factory expansion can absorb cash |
Revenue tells us what has already been delivered. Backlog can reveal demand that has not reached sales yet.
That makes backlog especially useful when orders are rising faster than factories can deliver.
But backlog is not guaranteed profit. It is a bridge between demand and execution.
What the Latest Disclosures Show
The numbers below were checked in July 2026. They should not be compared without reading the scope.
Vertiv: Demand First, Then Execution
Vertiv ended 2025 with backlog of $15.0 billion, up 109% from one year earlier. Its fourth-quarter book-to-bill ratio was about 2.9.
In the first quarter of 2026, sales reached $2.65 billion, up 30%. Adjusted operating margin reached 20.8%, and adjusted free cash flow was $653 million.
The first-quarter release did not publish a new headline backlog figure. This is normal. Investors should keep the last disclosed backlog and then track whether sales, margins, and cash improve.
Schneider Electric: Backlog by Business and Timing
Schneider Electric ended 2025 with group backlog of €25.362 billion, up 18%.
Energy Management backlog was €21.340 billion, up 21%. The company said the largest increase came from Systems in North America, driven by faster data center demand.
It also reported strong growth in data-center-linked backlog scheduled for 2027 and later.
This is useful because it shows the business, region, and expected timing.
Eaton: Growth Rates Without One Absolute Number
Eaton reported that Electrical Americas backlog was 44% higher in March 2026 than one year earlier. Electrical Global backlog was up 73%.
The rolling 12-month book-to-bill ratio for the electrical businesses was 1.2.
Eaton did not give one absolute data-center-only backlog number. The correct approach is to record exactly what the company reports, not estimate a number that was never disclosed.
ABB: Strong Orders, but Check the Segment
ABB reported group backlog of about $30.0 billion at the end of June 2026, up 28% on a comparable basis. Group book-to-bill was 1.27.
Electrification reported book-to-bill of 1.39 and backlog of $13.7 billion. ABB said rapid data center investment remained a major catalyst for the segment.
The same segment also serves several other markets. The number shows strong electrification demand, not data center demand alone.
The Backlog Quality Test
A large number is not enough. A better analysis asks what is inside it.
-
Is it a signed order?
Do not confuse pipeline with accepted customer demand. -
Which business owns it?
Check whether it belongs to data centers, a broader electrical segment, or the entire company. -
When will it become revenue?
Next-quarter delivery is different from work scheduled for 2028. -
Can the customer cancel or delay it?
Read the risk language, not only the headline. -
Is the order concentrated?
One large customer can make growth look broader than it is. -
Can the supplier deliver?
Factory capacity, components, and skilled labor matter. -
Are prices protected?
Fixed-price orders can lose margin when costs rise. -
Does it produce margin and cash?
Backlog creates value only when execution is profitable.
The best backlog is not simply the largest. It is the backlog that can be delivered on time, at a healthy margin, and converted into cash.
How to Find Backlog Yourself
An expensive financial database is not required.
Step 1: Open the Official Investor Relations Page
Search the company name followed by
investor relations.
Start with the company’s own website.
Step 2: Open the Main Documents
- latest earnings release
- earnings presentation
- annual or quarterly report
- earnings-call transcript, when available
For U.S. companies, SEC EDGAR provides free access to Form 10-K and Form 10-Q filings.
Step 3: Search Several Terms
Use Ctrl + F on Windows or Command + F on a Mac.
Search for:
backlog,
order backlog,
orders,
bookings,
book-to-bill,
remaining performance obligations,
and data center.
Step 4: Record the Scope and Date
Write down whether the number covers:
- the whole company or one segment
- data centers or several end markets
- an absolute amount or only a percentage change
- one quarter or a rolling 12-month period
- the current year or later delivery years
Step 5: Read the Risk Language
Search for:
cancel,
delay,
fixed-price,
customer concentration,
capacity,
and supply chain.
Step 6: Build a Simple Tracking Table
| Company | Report date | Backlog | Growth | Book-to-bill | Scope and timing |
|---|---|---|---|---|---|
| Vertiv | Dec. 31, 2025 | $15.0B | +109% | 2.9, Q4 | Company backlog; retain until a newer figure is disclosed |
| Schneider Electric | Dec. 31, 2025 | €25.362B | +18% | Not stated here | Group; Energy Management €21.340B, with data-center-linked work extending beyond 2026 |
| Eaton | Mar. 31, 2026 | Not disclosed as one amount | Americas +44%; Global +73% | 1.2, rolling 12 months | Electrical segments, not data centers alone |
| ABB | Jun. 30, 2026 | $30.0B group; $13.7B Electrification | Group +28% | 1.27 group; 1.39 Electrification | Broad electrification exposure; data centers are one important driver |
Update the table after every earnings report. Keep the source link and the date beside each number.
Book-to-Bill Shows the Direction
Book-to-bill = new orders ÷ revenue
A ratio above 1 means orders arrived faster than the company recognized revenue. Backlog will usually rise, before cancellations and other adjustments.
A ratio below 1 means the company delivered more than it booked.
One quarter can be distorted by a large order. A rolling 12-month number usually gives a steadier view.
What a Healthy Pattern Looks Like
- orders grow across several customers and regions
- book-to-bill remains above 1 over time
- backlog converts into revenue near the expected schedule
- operating margins remain stable or improve
- free cash flow follows reported profit
What Should Make Investors Cautious
- one large order creates most of the growth
- delivery dates keep moving further into the future
- backlog rises while margins fall
- inventory and capital spending rise faster than revenue
- the company repeatedly warns about cancellations or fixed-price risk
A growing backlog can mean strong demand. It can also mean the company is struggling to deliver.
The Main Idea
AI data center equipment companies sit between grid electricity and computing.
Their products are essential, but demand reaches the financial statements in stages.
Orders show customer intent. Backlog shows unfinished demand. Revenue shows delivery. Margins and cash show whether the company created value.
Investors should follow the entire chain. The largest backlog alone does not provide the answer.
Related Articles
- Who Gets Paid When AI Data Centers Are Built?
- What Must Be Built to Power the AI Data Center Boom?
- Why Power, Not Chips, May Limit the AI Data Center Boom
Sources
- Vertiv FY2025 results: backlog and book-to-bill
- Vertiv Q1 2026 results: sales, margin, and free cash flow
- Schneider Electric FY2025 results and backlog
- Schneider Electric Q1 2026 revenue
- Eaton Q1 2026 orders, backlog growth, and book-to-bill
- ABB Q2 2026 backlog and Electrification results
- Investor.gov: Using EDGAR to research investments
The companies in this article are representative examples. Their inclusion is not a ranking or an investment recommendation. Backlog definitions differ by company and may not be directly comparable. Financial data checked in July 2026.