Growth capex is the money a company spends to expand capacity, enter new markets, or build capabilities it didn’t have before; maintenance capex is the money it spends to keep existing operations running at their current level. The distinction between growth capex vs. maintenance capex matters because it determines how much cash a business actually generates for its owners after paying to preserve what it already has. Companies almost never report the split cleanly, so investors have to estimate it.
What Each Bucket Actually Contains
Maintenance capex is non-discretionary. When a manufacturing plant’s conveyor belt wears out and gets replaced to keep the line moving, that’s maintenance. When a delivery company retires a ten-year-old van and buys an identical replacement for the same route, that’s maintenance too. The spending preserves what the business has without adding anything new.
Growth capex is discretionary. A manufacturer spending $50 million to build a new production wing that lifts output by 30% is investing for growth. A logistics company buying ten additional trucks to service a newly acquired national contract is doing the same. So is a retailer opening stores in a region where it previously had no presence, or a tech company adding data center capacity to support a new product line.
Why the difference matters: a company reporting strong net income while quietly underinvesting in maintenance is borrowing from its own future. The assets degrade, and eventually a large catch-up bill arrives that wipes out years of apparent profitability. A company spending heavily on real growth may look like it’s burning cash when it’s actually building future earnings. Treating all capex the same obscures both situations.
Why the Line Blurs in Practice
The clean two-bucket framework breaks down constantly. The most common headache is replacement with improvement. When a company retires an old machine and installs a newer model, the replacement almost always has better specs, lower energy use, or higher throughput. Is the full cost maintenance, or is the incremental improvement growth?
Accounting standards offer partial guidance. Under U.S. GAAP, costs that extend an asset’s life or increase its functionality can be capitalized, while routine repairs are expensed as incurred. IAS 16 draws a similar line internationally: day-to-day servicing is expensed immediately, but replacement of major components gets capitalized when the recognition criteria are met.1IFRS Foundation. IAS 16 Property, Plant and Equipment Those standards tell you what lands on the balance sheet. They don’t tell you what portion of the capitalized amount is growth versus maintenance. That split is left to the investor.
Technological obsolescence complicates the picture further. An automaker spending billions to retool plants for electric vehicles is technically maintaining competitive viability rather than expanding into a new business, but the magnitude of the spend looks nothing like routine maintenance. These are the judgment calls where knowing the underlying business gives an analyst a real edge over someone running formulas.
Where to Find the Numbers
Total capital expenditure appears on the statement of cash flows under investing activities, usually as “purchases of property, plant, and equipment” or “capital expenditures.” Depreciation and amortization appear on the income statement and again in the operating activities section of the cash flow statement, where they’re added back to net income as a non-cash charge.
For context, read the Management Discussion and Analysis section of the 10-K. SEC rules require public companies to describe their material cash requirements, including commitments for capital expenditures, the source of funds, and the general purpose of those requirements.2eCFR. 17 CFR 229.303 – (Item 303) Management’s Discussion and Analysis Some companies voluntarily break out growth spending from maintenance in their MD&A or on earnings calls. When they do, take it seriously but verify the logic, because management has an incentive to label spending as “growth” since that implies future returns rather than upkeep. Reading several years of these disclosures shows you whether the capex narrative holds together over time.
Three Methods for Estimating the Split
Since the split is almost never disclosed cleanly, you estimate it. Three approaches dominate.
The Depreciation Proxy
The most common method treats depreciation and amortization as a stand-in for maintenance capex. The logic is that D&A approximates the annual cost of wear and tear on existing assets, so subtracting it from total capex leaves the growth component.
Formula: Total Capex − D&A = Growth Capex.
The weakness is that depreciation is calculated on the historical purchase price of assets bought years ago. Replacing those same assets today almost always costs more because of inflation and rising material costs. Research from Morgan Stanley’s investment management division found that in aggregate, maintenance capital expenditures exceed depreciation by roughly 20%, with substantial variance across industries. The depreciation proxy therefore tends to overstate growth capex by understating what maintenance really costs. Analysts using this method should adjust D&A upward for inflation, and be especially skeptical for companies with old asset bases in industries facing high input-cost inflation.
The Historical Baseline
A second approach looks at what the company spent during periods of flat or zero revenue growth. The assumption is that during years the business wasn’t expanding, essentially all capex was maintenance. That level becomes your baseline, and anything above it in growth years is growth capex.
Period selection matters. A year of low spending because management deferred necessary maintenance gives you a falsely low baseline. You want periods of genuine operational stability, not neglect. Recessions can work as reference points, but only if the company didn’t slash necessary spending to protect short-term earnings. Comparing asset condition and efficiency ratios across the period helps confirm the business was actually being maintained.
The Revenue Regression
The third technique runs a regression of capex against revenue changes over a long time series, usually a decade or more. The portion of spending that moves with revenue increases represents growth capex. The residual amount the company spends regardless of whether revenue is rising or flat approximates maintenance capex.
This has statistical rigor the other methods lack, but it needs a long enough series to be meaningful and assumes the relationship between capex and revenue growth is reasonably stable. Companies going through strategic shifts, entering new industries, or making large acquisitions throw off the regression. It works best for mature businesses with consistent operating models.
How the Split Feeds Valuation
The whole exercise matters because of one number: free cash flow. The standard FCF calculation subtracts total capex from operating cash flow, which penalizes companies for investing in growth and makes an aggressively expanding business look worse than a stagnant one generating the same operating cash flow. For valuation, many analysts calculate an adjusted FCF that subtracts only maintenance capex, revealing the cash the business generates after keeping its existing asset base intact.
Warren Buffett popularized a version of this in his 1986 letter to Berkshire Hathaway shareholders as “owner earnings”: net income, plus depreciation and amortization, minus maintenance capex, minus working capital increases. The output represents the cash that could theoretically be extracted from the business each year without impairing its competitive position. The framework depends entirely on a credible estimate of maintenance capex.
High growth capex signals management confidence in expansion and can justify higher valuation multiples, but confidence alone doesn’t create value. The investment has to earn returns above the company’s cost of capital. When return on invested capital exceeds the weighted average cost of capital, each dollar of growth spending creates more than a dollar of enterprise value. When ROIC falls below WACC, growth actively destroys shareholder value: the company pours money into projects earning less than investors could get elsewhere.3Financial-Economics.nl. When Growth Destroys Value: Capital Intensity, ROIC, and the Cost of Capital The volume of growth spending matters far less than its efficiency.
Industry Context for the Numbers
How much a company spends on capex relative to revenue varies enormously by industry, and this context is essential when judging whether a growth capex figure is reasonable. As of early 2026, net capital expenditures as a percentage of sales run from under 2% for aerospace and defense to over 40% for water utilities. General utilities sit around 34%, power companies around 24%, and internet software companies around 26%. Traditional manufacturing sectors like auto and truck, chemicals, and machinery cluster between 2% and 5%. General retail is around 5%.
A 10% capex-to-revenue ratio would be alarmingly high for a machinery company and perfectly normal for a utility. The growth capex component also tends to be a larger share of total capex in faster-growing industries and smaller in mature, heavily regulated sectors where most spending goes toward maintaining aging infrastructure. Comparisons across industries without adjusting for capital intensity don’t produce useful conclusions; always compare within the same sector.