What Is CapEx? A Complete Guide for Enterprise Finance Teams
Ask most finance sites what CapEx is and you get a balance sheet answer. Capital expenditure is money spent to acquire or improve a long-term asset, capitalized rather than expensed, depreciated over its useful life. That definition is correct, and it stops being useful about ten minutes into the job.
It's useful if you're reading someone else's financial statements. It's close to useless if you're the person accountable for four hundred million dollars of capital moving through dozens of sites, and somebody has asked you how much of it is actually committed.
This guide covers the definition, the formula, the classification calls that cause the most arguments and the part almost nobody writes about: what happens to a capital dollar between the moment somebody requests it and the moment it lands in an asset. If you manage CapEx rather than analyze it from outside, that second half is the part that decides your year.
What is CapEx?
CapEx, short for capital expenditure, is spending that acquires, upgrades or extends the useful life of a long-term asset. A new production line, a fleet replacement, a plant expansion, a building fit-out, a major systems implementation.
Three characteristics separate capital spending from everything else on the ledger.
Common examples include land and buildings, machinery and production equipment, vehicles and mobile fleet, IT hardware and infrastructure, major software implementations and capitalized improvements to assets you already own.
Why the duration matters more than the dollar value
The interesting thing about that third characteristic is what it implies. Capital spending is governed differently because the decision is hard to reverse. You can cut an operating line in March and feel the benefit in April. A half-built facility doesn't offer that option.
That irreversibility is why capital deserves a control layer that operating spend doesn't need, and it's why treating CapEx as an oversized operating budget line tends to end badly. The controls quietly disappear, and the organization finds out at year end.
CapEx versus OpEx
The textbook distinction is straightforward. CapEx buys the asset. OpEx runs it. Buy the vehicle and it's capital; fuel it and it's operating.
| Dimension | CapEx | OpEx |
|---|---|---|
| Accounting treatment | Capitalized on the balance sheet | Expensed in the period incurred |
| Time horizon | Multiple years | Within the current period |
| Typical governance | Delegation of authority, Investment Committee, stage gates | Departmental budget authority |
| Reversibility | Low once committed | Relatively high |
| Cash timing | Often front-loaded | Spread across the period |
The classification calls that actually cause arguments
The vehicle-and-fuel example is easy because it was chosen to be easy. Finance teams rarely argue about vehicles. They argue about these:
| Grey area | The question your policy has to answer |
|---|---|
| Cloud software subscriptions | Are you acquiring an asset you control, or a service you consume? And is the implementation and configuration effort treated as a separate question from the subscription itself? |
| Major maintenance versus overhaul | Does the work restore the asset to its expected condition, or extend its life or capacity beyond the original specification? |
| Internal labour on capital projects | Is the time directly attributable to building or configuring the asset, and can you evidence it project by project? |
| Software development | Which phase of the work is this, and does your policy draw the line in the same place for internal-use software as it does for software built to be sold? |
| Small-value asset purchases | Does it clear your capitalization threshold, and is that threshold applied the same way at every site? |
Deliberately, those are questions rather than answers. The answers belong in your capitalization policy, they may differ depending on whether you report under IFRS or US GAAP, and they're the kind of judgement that deserves your auditor rather than a blog post. We take the software case further in capitalizing and expensing software costs, which is the row that causes the most disagreement.
What matters operationally is narrower and harder than getting the treatment right. It's that the same question gets the same answer in every part of the business.
When forty analysts across dozens of sites each make these calls slightly differently, the consolidated capital number isn't wrong in an obvious way. It's wrong in a way nobody can find, because each individual judgement looked reasonable to the person making it.
The CapEx formula, and what it can't tell you
The standard calculation derives capital expenditure from two consecutive balance sheets.
CapEx = (Current period PP&E − Prior period PP&E) + Depreciation for the period
A worked example. A company held $12.0 million in property, plant and equipment at the start of the year and $12.5 million at the end. Depreciation for the year was $900,000.
CapEx = ($12.5m − $12.0m) + $0.9m = $1.4 million
The half-million dollar increase understates what was actually spent, because depreciation reduced the book value of everything else at the same time. Adding it back recovers the real figure.
That formula is genuinely useful, and it has a limitation that most guides skip past. It's derived from published financial statements, which means it tells you what you spent after the period closed. It's an outside-in measure, built for analysts reading a company from the outside.
If you're inside the business, that number arrives too late to act on. What you need is the forward view: what has been requested, what has been approved, what has been contractually committed and what has actually landed. The formula gives you none of that, and no accounting standard requires it to.
The four states of a capital dollar
Here's the distinction that resolves more confusion than any other. A capital dollar is rarely a single number. It moves through four states, and each one answers a different question.
| State | What it means | The question it answers | Where it usually lives |
|---|---|---|---|
| Requested | Submitted in the plan or backlog, not yet approved | What does the business think it needs? | Site plans, intake forms, spreadsheets |
| Approved | Authorized under delegation of authority, funding released | What are we permitted to spend? | Approval system or email trail |
| Committed | Purchase orders raised, contracts signed | What can we no longer avoid spending? | ERP and procurement |
| Spent | Invoiced, paid, capitalized | What has actually left the business? | General ledger |
Two things follow from that table, and both cost organizations real money.
The first is that these four numbers are almost never equal, and the gaps carry meaning. A large gap between approved and committed suggests projects are stalling after authorization. A large gap between committed and spent suggests delivery is running behind the paperwork. Neither gap is visible if you only report one number.
The second is more practical. In most organizations, those four states live in four different systems owned by three or four different groups. Committed capital is the one that hurts most often, because it sits in procurement rather than finance. Money that's contractually locked in but absent from the capital forecast is the single most common source of the December conversation nobody wants to have.
Four numbers, one record
Requested, approved, committed and spent rarely sit in the same place. A short walkthrough shows what it looks like when they do, at project level and across the portfolio.
Request a demoThree types of CapEx, and why the mix matters
Total capital spend is a weak number on its own. What it's composed of is the more useful question.
A portfolio ranked purely on financial return tends to starve the second and third categories, because they compete poorly against growth projects on any NPV comparison. That deferral is invisible for a while, then it isn't. Tracking the mix year over year, and watching whether maintenance is being quietly deferred to fund growth, is a small piece of reporting with an outsized effect on how the next five years go.
How to forecast CapEx
Three methods appear in most guidance, and they're all reasonable starting points.
All three share one blind spot: they extrapolate from what you spent before. That's a reasonable proxy at portfolio level and a poor one for an organization running hundreds of active projects, because the strongest available signal isn't history. It's the commitments already in the system.
At enterprise scale, the more accurate forecast is built bottom-up. Take the approved portfolio, add committed values from purchase orders and contracts, layer in project schedules and phasing, apply the delivery pattern you actually observe rather than the one in the plan, then reconcile the total against the top-down envelope. Where the two disagree, that gap is worth more attention than either number alone.
Two habits make a bigger difference than method selection. Reforecast active projects monthly rather than quarterly, and treat a forecast that hasn't changed in two quarters as a forecast nobody is maintaining.
Four places enterprise CapEx breaks
Classification drifts across the business
The grey areas covered above are decided locally, and local decisions drift. One site capitalizes a refurbishment, another expenses it, and both are defensible in isolation. Consolidated, they produce a capital number that can't be compared year over year or site to site, and the reconciliation effort grows faster than the analysis.
Hood Companies reached that point with more than 120 spreadsheets across 125 sites feeding a single capital budget. The effort wasn't going into deciding where capital should go. It was going into making the numbers agree.
Commitments are invisible until they're actuals
Approved capital sits in the finance view. Committed capital sits in procurement. In many organizations the two only meet at month end, or later, which means the capital forecast reflects intention rather than obligation.
The practical fix is unglamorous and effective: bring contracted values and open purchase orders into the capital view, and report approved, committed and spent as three separate columns rather than one.
Carryover arrives without being re-examined
Approved capital that goes unspent has to go somewhere, and in most organizations it rolls into next year. What rarely happens is a fresh look at whether the original case still stands.
Carryover isn't automatically a problem. A project delayed two months for a permit is fine. The problem is that carryover typically enters the new year with its original approval intact, competing for delivery capacity against projects that were assessed on current assumptions. A project justified eighteen months ago on eighteen-month-old input costs may no longer clear the bar, and nobody has asked.
A useful discipline is to break the opening budget into its components: genuinely new requests, planned continuations of multi-year work and unplanned carryover from projects that didn't proceed as expected. The third category is the one that deserves a re-approval conversation, and it's often larger than leadership expects.
The gap between the sites and the centre
Requests originate where the assets are. Authority sits at the centre. Between those two points is where most of the delay lives.
That delay has a price, and it's rarely measured. Every week an Authorization for Expenditure spends in a queue is a week of deferred return, and in volatile input markets a delayed capital decision is often a repriced capital decision. Steel quotes expire. Contractor availability shifts. Equipment lead times stretch.
Macmahon Holdings reduced its AFE approval cycle from five to six weeks down to under one week. That change didn't improve a single business case on paper. It changed how much of each case survived contact with reality.
What good CapEx management looks like
If you're assessing your own process, or evaluating a system to support it, six things separate capital administration from capital management.
CapEx360 was built as a capital management system covering that full span, from intake through post-investment review, rather than as an approval workflow with reporting attached. That distinction is worth pressing every vendor on.
Frequently asked questions
- What is CapEx in simple terms?
- CapEx is money spent to buy, upgrade or extend the life of a long-term asset. It goes onto the balance sheet and depreciates over the asset's useful life, rather than being expensed in the period it's paid.
- What is the difference between CapEx and OpEx?
- CapEx buys assets that deliver value over multiple years and is capitalized. OpEx is the cost of running the business day to day and is expensed in period. The harder cases sit in between, including cloud subscriptions, major maintenance and capitalized internal labour, where the answer depends on your accounting policy and capitalization threshold.
- How do you calculate CapEx?
- Subtract prior period property, plant and equipment from current period PP&E, then add depreciation for the period. The result tells you what was spent after the period closed, which is why organizations managing capital actively also track requested, approved and committed amounts.
- What are examples of capital expenditures?
- Land and buildings, production equipment and machinery, vehicles and mobile fleet, IT hardware and infrastructure, major software implementations and capitalized improvements to existing assets.
- What is CapEx carryover?
- Approved capital that goes unspent in one budget year and moves into the next. Some carryover is normal timing. The risk is that it enters the new year carrying an approval based on assumptions that may no longer hold.
- What is the difference between CapEx and capital budgeting?
- CapEx is the spending itself. Capital budgeting is the process used to evaluate, prioritize and authorize it. CapEx management covers the full lifecycle around both, from intake through commitments, forecasting, actuals and post-investment review.
Where to go from here
The accounting definition of CapEx answers a reporting question. The operational definition answers a harder one: at any given moment, how much of your capital is requested, how much is approved, how much is contractually committed and how much has actually been spent.
Organizations that can answer those four questions on demand tend to spend their capital more deliberately, redeploy it faster when projects slip and enter each budget year knowing what they're carrying. Organizations that can't tend to find out in December.
See how CapEx360 manages the full capital lifecycle, from request through commitment, forecast and post-investment review.
Request a demo
