An AI infrastructure company announces record demand.
Orders are up.
Backlog is growing.
The data-center market is “very strong.”
That sounds good.
But has the company actually delivered the equipment?
Did the revenue arrive?
Did profit margins improve?
Did the accounting profit turn into cash?
Orders show customer intent. Backlog shows unfinished demand. Revenue shows delivery. Margins and cash show whether execution created value.
By the end of this article, you will be able to open an AI infrastructure company’s earnings release, follow demand from orders to cash, understand terms such as backlog and book-to-bill when they first appear, and avoid treating one large order number as proof of future profit.
First Understand What the Company Actually Sells
Before looking at a financial table, identify the equipment.
AI data centers need more than GPUs.
Electricity has to enter the site, move safely to the racks, stay available through short interruptions, and support increasingly dense computing loads.
Heat has to leave the chips just as quickly.
That creates demand for several layers of equipment.
Grid Connection and Switchgear
Electricity usually enters through a utility connection and a substation.
A transformer changes voltage so electricity can move through different parts of the system efficiently and safely.
Switchgear is the equipment that switches, protects, and isolates electrical circuits. In simple terms, it helps control where power flows and disconnects equipment when something goes wrong.
UPS Systems and Batteries
A short grid interruption can stop servers.
A UPS, or uninterruptible power supply, is a battery-backed system that keeps critical equipment powered during short interruptions while the grid recovers or backup generation starts.
Busways and Power Distribution
A busway is a high-capacity enclosed conductor system that carries electricity through a building. Think of it as a modular electrical highway.
Power distribution units then send electricity toward rows of racks and help measure and manage the load.
Air and Liquid Cooling
Computing turns most of its electricity into heat.
Traditional facilities rely heavily on air cooling and chilled-water systems.
Dense AI racks increasingly add direct-to-chip liquid cooling, where liquid flows to cold plates attached directly to processors.
A CDU, or coolant distribution unit, manages the liquid loop between the building’s cooling system and the server equipment.
Monitoring and Service
Software and sensors monitor voltage, battery condition, temperature, and coolant flow.
Maintenance, replacement parts, and field service can create revenue long after the original construction project is finished.
The Contexta Demand-to-Cash Chain
Once the equipment is clear, follow the financial chain:
Pipeline → Orders → Backlog → Revenue → Margin → Cash
Each word answers a different question.
Pipeline: What Might Be Won?
A pipeline is a set of possible projects or sales opportunities.
A pipeline is not a signed contract.
A company can discuss a very large opportunity pipeline without eventually winning all of it.
Orders: What Did Customers Commit To?
Orders are customer commitments the company accepted during a period, according to that company’s accounting or reporting definition.
Definitions differ, so always check whether the company includes cancellable orders, service contracts, multi-year projects, or only firm purchase commitments.
Backlog: What Has Not Been Delivered Yet?
Backlog is accepted or contracted work that has not yet been fully recognized as revenue.
It is useful because it can show demand before that demand reaches the income statement.
But backlog is not money already earned.
Delivery can be delayed. Scope can change. Some orders can be canceled. Costs can rise before the project is completed.
Revenue: What Was Delivered?
Revenue is the sales amount recognized when products or services are delivered according to the applicable accounting rules.
Revenue is more concrete than pipeline or backlog because execution has moved further forward.
Margin: Did the Delivery Create Profit?
A margin is the percentage of revenue left after a defined set of costs.
For example, operating margin tells us how much operating profit remains from each dollar of sales after operating costs under the company’s reporting definition.
Revenue can grow while margin falls if labor, components, tariffs, freight, or fixed-price contracts become more expensive.
Cash: Did the Profit Become Money?
Operating cash flow is cash generated by normal business operations.
Free cash flow is a non-GAAP measure commonly used to estimate cash left after operating needs and capital spending, but the exact definition can differ by company.
This is the final reason to follow more than revenue.
A company can report higher profit while inventory, receivables, or factory expansion absorb cash.
Why Investors Get Backlog Wrong
A common question is:
If backlog is growing quickly, does that mean future profit is already visible?
No.
Backlog is evidence of unfinished demand, not guaranteed profit.
Imagine two companies each report $10 billion of backlog.
- Company A can deliver most of it in 18 months at stable prices and high margin.
- Company B needs four years, faces fixed-price contracts, rising component costs, and one customer represents half the backlog.
The headline number is the same.
The economic quality is not.
The Contexta Backlog Quality Test
- Is it really backlog?
Do not confuse a project pipeline with accepted customer demand. - What is the scope?
Whole company, one segment, or data centers only? - When will it convert?
Next quarter, next year, or several years later? - Can it be canceled or delayed?
Read the risk language around customer rights and project timing. - Is it concentrated?
A few hyperscalers can create very large orders. - Can the supplier deliver?
Factories, components, logistics, and skilled labor can constrain execution. - Are prices protected?
Fixed-price contracts can lose profitability if costs rise. - Does it become margin and cash?
That is the real test.
The best backlog is not simply the largest. It is backlog that can be delivered on time, at acceptable margins, and converted into cash.
Book-to-Bill: A Direction Indicator, Not a Profit Indicator
Book-to-bill = new orders ÷ revenue
Book-to-bill compares the pace of incoming orders with the pace of recognized revenue.
- Above 1.0: orders arrived faster than revenue was recognized.
- Below 1.0: revenue was recognized faster than new orders arrived.
A ratio above 1.0 can help backlog grow.
But it does not tell us whether those orders will have strong margins or generate cash.
One quarter can also be distorted by one huge contract.
That is why a rolling 12-month ratio—calculated over the most recent twelve months—can sometimes give a steadier view.
Four Companies, Four Different Exposure Profiles
Vertiv, Schneider Electric, Eaton, and ABB compete in overlapping areas.
But they should not be compared as if data centers mean the same thing to each company.
Vertiv: More Direct Data-Center Exposure
Vertiv sells critical digital-infrastructure systems including UPS equipment, power distribution, batteries, thermal systems, liquid cooling, modular systems, and services.
Its data-center exposure is comparatively direct, so changes in AI infrastructure spending can show up clearly in results.
Vertiv’s last headline backlog figure in the source article was $15.0 billion at December 31, 2025, up 109% year over year, with Q4 book-to-bill around 2.9.
That figure should not be silently relabeled as a current Q2 2026 backlog number.
What did update in Q2 2026?
- net sales: $3.274 billion, up 24% year over year
- adjusted operating margin: 22.6%
- operating cash flow: $1.1 billion
- adjusted free cash flow: $925 million
Vertiv said Q2 revenue included minor timing shifts caused by supply-chain congestion and multi-phase project execution.[1]
This is a useful real-world example:
strong demand can coexist with timing problems in reported revenue.
Schneider Electric: Strong Data-Center Demand Inside a Broader Portfolio
Schneider Electric sells electrical distribution, UPS systems, cooling, controls, prefabricated systems, software, and service.
But it also serves factories, buildings, utilities, and industrial automation.
At year-end 2025, the source article recorded group backlog of €25.362 billion and Energy Management backlog of €21.340 billion.
By H1 2026, Schneider reported record first-half revenue of €21.2 billion, adjusted EBITA margin of 19.3%, and free cash flow of €1.6 billion. It said demand remained very high in data centers and that the group had a record backlog.[2]
Adjusted EBITA is Schneider’s adjusted earnings before interest, taxes, and amortization measure. It is a profitability measure, not cash flow.
The lesson is scope:
strong data-center demand can be an important driver without making every Schneider number a data-center number.
Eaton: Backlog Growth Without One Data-Center-Only Backlog Number
Eaton sells switchgear, breakers, busways, UPS systems, distribution equipment, and monitoring systems.
Its Boyd Thermal acquisition also expanded its thermal-management exposure.
In Q2 2026, Eaton reported:
- Electrical Americas backlog up 33% year over year
- Electrical Global backlog up 103%
- rolling 12-month book-to-bill of 1.2 for the electrical businesses
- company free cash flow of $874 million
Eaton did not provide one absolute “data-center backlog” number in that release.[3]
That absence is important.
Do not manufacture a number the company did not disclose.
ABB: Read the Segment Before Reading the Headline
ABB sells electrification and automation equipment across data centers, utilities, industry, transport, and buildings.
Its Q2 2026 results were the latest group results available at this rebuild date.[4]
The source article recorded group backlog of about $30.0 billion at June 30, 2026 and Electrification backlog of $13.7 billion, with book-to-bill of 1.27 for the group and 1.39 for Electrification.
Data-center demand was an important Electrification driver.
But Electrification serves other end markets too.
So the correct statement is not:
“ABB has $13.7 billion of data-center backlog.”
It is:
“ABB Electrification had $13.7 billion of backlog, with data centers among the important demand drivers.”
The Contexta Scope Tag
Every number in your research sheet should carry a scope tag.
For example:
COMPANY— whole companySEGMENT— one reported business segmentEND MARKET— data-center-specific or another market if disclosedPERIOD— one quarter, year-end, or rolling 12 monthsCURRENCY— USD, EUR, or another currency
This small habit prevents many false comparisons.
A Better Tracking Table
| Company | Metric date | Demand signal | Scope | Execution signal |
|---|---|---|---|---|
| Vertiv | Backlog: Dec. 31, 2025; execution: Q2 2026 | $15.0B last disclosed headline backlog; Q4 B2B ~2.9 | company backlog | Q2 sales $3.274B; adj. op. margin 22.6%; adj. FCF $925M |
| Schneider Electric | Backlog absolute: Dec. 31, 2025; execution: H1 2026 | €25.362B group backlog at FY25; H1 record backlog | group / Energy Management | H1 revenue €21.2B; adj. EBITA margin 19.3%; FCF €1.6B |
| Eaton | Q2 2026 | Americas backlog +33%; Global +103%; B2B 1.2 | Electrical segments | company FCF $874M; segment sales/margins growing |
| ABB | Q2 2026 | group backlog ~$30.0B; Electrification $13.7B | group / Electrification | data centers important, but not sole segment driver |
The purpose of the table is not to declare a winner.
It is to keep dates and scopes attached to every number.
Why Strong Demand Can Still Disappoint Investors
A recent investor discussion around Vertiv captured a common confusion: the company raised guidance and reported strong profit and cash flow, yet investors still debated a revenue shortfall caused partly by timing.[5]
This happens because markets compare results with expectations, not only with last year.
A company can have:
- strong end-market demand
- growing backlog
- higher profit
and still disappoint if:
- delivery is slower than expected
- future growth expectations were even higher
- valuation already assumes near-perfect execution
- cash conversion weakens
This is why demand analysis and stock valuation are separate jobs.
How to Find the Numbers Yourself
Step 1: Start with Investor Relations
Search the company name plus investor relations.
Use the company’s official site before blogs or social posts.
Step 2: Open the Main Documents
- latest earnings release
- earnings presentation
- quarterly or annual report
- earnings-call transcript, when available
For U.S.-listed companies, SEC EDGAR also provides free access to filings.
Step 3: Search the Document
Use Ctrl+F or Command+F and search:
backlog, orders, bookings, book-to-bill, remaining performance obligations, margin, cash flow, and data center.
Remaining performance obligations, often shortened to RPO, are contracted future revenue obligations under accounting disclosure rules. They are not automatically identical to backlog, so do not merge the two unless the company explains the relationship.
Step 4: Record Scope, Date, and Currency
Write those three fields beside the number before copying it into a spreadsheet.
Step 5: Read the Risk Language
Search:
cancel, delay, fixed-price, customer concentration, capacity, supply chain, and working capital.
Working capital is the money tied up in short-term operating items such as receivables, inventory, and payables. Rapid growth can consume cash if inventory and customer receivables rise faster than supplier financing.
Step 6: Update the Sheet After Every Earnings Report
Never overwrite an old figure without keeping its date.
A year-end backlog number and a Q2 revenue number can coexist in the same research sheet if their dates are labeled clearly.
The Contexta Three-Pass Research Method
Instead of reading the entire earnings package randomly, use three passes.
Pass 1 — Demand: orders, backlog, book-to-bill, customer pipeline.
Pass 2 — Execution: revenue growth, delivery timing, factory capacity, margin.
Pass 3 — Cash: operating cash flow, free cash flow, inventory, receivables, capital spending.
This turns a long earnings package into a repeatable research process.
A Healthy Pattern
- orders grow across several customers and regions
- book-to-bill stays above 1 over time without depending on one huge order
- backlog converts near the expected schedule
- revenue grows
- margins remain stable or improve
- cash flow follows reported profit
Warning Signs
- one customer creates most of the order growth
- delivery dates keep moving outward
- backlog rises but revenue repeatedly misses the expected timing
- revenue rises while margins fall
- inventory and capital spending absorb cash faster than sales grow
- management repeatedly highlights cancellation, fixed-price, or supply-chain risk
- a segment backlog is described publicly as if it were data-center-only backlog
The Main Idea
AI data-center equipment sits between electricity and computing.
The demand can be real and enormous.
But demand reaches financial statements in stages.
Do not stop at backlog. Follow the chain until the demand becomes delivered revenue, acceptable margin, and real cash.
Continue Reading
- Who Gets Paid in the AI Data Center Boom? — map the value chain before analyzing individual suppliers.
- What Must Be Built to Power the AI Data Center Boom? — understand the physical equipment behind the financial metrics.
- Why Power, Not Chips, May Limit the AI Data Center Boom — see why bottlenecks can create orders and delays at the same time.
Key Terms
- pipeline: possible future projects or sales opportunities that are not yet necessarily signed
- orders: customer commitments accepted under a company’s reporting definition
- backlog: accepted or contracted work not yet fully recognized as revenue
- revenue: sales recognized after required delivery or performance conditions are met
- margin: the percentage of revenue remaining after a defined set of costs
- book-to-bill: new orders divided by revenue over the same period
- rolling 12 months: the most recent twelve-month period, updated each quarter
- free cash flow: a non-GAAP cash measure generally based on operating cash after capital spending; definitions vary
- RPO: remaining performance obligations, a contracted future-revenue disclosure that is not necessarily the same as backlog
- working capital: short-term operating capital tied to receivables, inventory, and payables
- CDU: coolant distribution unit that manages liquid flow between facility cooling and server equipment
- busway: an enclosed high-capacity electrical conductor system used to distribute power
Sources
- Vertiv — Q2 2026 results.
- Schneider Electric — H1 2026 financial results.
- Eaton — Q2 2026 results.
- ABB — Q2 2026 results.
- Reader question signal — Vertiv demand, delivery timing, and expectations.
- U.S. SEC — EDGAR company filings.
Status checked October 1, 2026. The Demand-to-Cash Chain, Backlog Quality Test, Scope Tag, and Three-Pass Research Method are The Contexta analytical frameworks. Company backlog and order definitions differ, and figures from different scopes or dates should not be compared without adjustment.