Discounted cash flow (DCF): a step-by-step valuation on Apple

A discounted cash flow (DCF) valuation estimates what a business is worth today by forecasting the cash it will generate in the future and converting each future dollar into today's dollars. A dollar received in five years is worth less than a dollar now, both because money can earn a return in the meantime and because the future dollar might not arrive. The DCF makes that discount explicit, year by year.

It is the most widely taught intrinsic valuation method, and also the most misunderstood. A DCF does not reveal a hidden true price. It is a disciplined way of writing down what you believe about a company's growth, profitability, reinvestment and risk, and seeing what those beliefs imply per share. Change the beliefs and the answer changes, sometimes dramatically.

This guide builds a complete DCF on Apple Inc. using its fiscal 2025 numbers (the year ended September 27, 2025, from the 10-K filed October 31, 2025). Every step is shown: the starting free cash flow, the five-year forecast, the discount factors, the terminal value, the bridge from enterprise value to equity value, and the per-share result. Then we stress-test it with a sensitivity grid, run it backward to see what the market price implies, and cover the cost of capital, the exit-multiple alternative, the mid-year convention and the mistakes that most often wreck a model. Throughout, the Apple model is an educational outline, not a fair value or a view on the stock.

What a DCF is

The idea goes back at least to John Burr Williams's The Theory of Investment Value (1938): an asset is worth the present value of the cash it will pay its owners. For a bond, those cash flows are written in the contract. For a company, they have to be forecast, which is where all the difficulty lives.

A DCF has four moving parts:

1. The cash flow you forecast, usually free cash flow.
2. The explicit forecast period, typically five or ten years, where you model each year.
3. The terminal value, a single number standing in for all the years after the forecast.
4. The discount rate, which converts future dollars to present dollars and reflects risk.

Step 1: choose the cash flow

Most equity DCFs discount unlevered free cash flow (cash available to all capital providers, before interest) at the weighted average cost of capital, which produces enterprise value. A simpler variant, used here, starts from reported Free cash flow (FCF): definition, formula and Apple example: cash from operations minus capital expenditure. For Apple the difference is small, because its interest costs are tiny relative to its cash flow, but in a heavily indebted company it matters, and you should match the cash flow to the discount rate (unlevered with WACC, levered with the cost of equity).

Two judgment calls hide inside that number. First, Apple recorded $12,863 million of stock-based compensation in fiscal 2025. It is a non-cash expense added back in operating cash flow, but it is a real cost to shareholders, paid in new shares instead of cash. Some analysts subtract it from free cash flow (giving $85,904 million for Apple); others leave it in and use a diluted share count. Either is defensible; doing neither understates the cost. Second, the starting year should be representative. If a year includes a one-off tax payment or an unusually large working-capital swing, normalize it before you grow it.

Step 2: forecast the explicit period

Next, project free cash flow for each year of the forecast. A careful model forecasts revenue, margins, taxes, capital expenditure and working capital separately, because each has its own drivers. The illustrative version here uses one growth rate to keep the mechanics visible: 6% a year for five years. That is an assumption chosen for teaching, not a forecast of Apple. For context, Apple's net sales grew from $391.0 billion in fiscal 2024 to $416.2 billion in fiscal 2025, about 6.4%.

How long should the explicit period be? Long enough for the company to reach a steady state, where growth, margins and reinvestment have settled to levels you would expect to last. Five years suits a mature business; a fast grower might need ten or more, with growth fading year by year.

Step 3: discount each year

Each forecast year is divided by its discount factor, (1 + r) raised to the number of years. We use 9% here, again an illustrative input; the section on the cost of capital below shows where such a number comes from.

Step 4: the terminal value

A company does not stop at year five, so the model needs a value for all the cash flows after it. The most common approach is the Gordon growth (perpetuity growth) formula, named for Myron Gordon's 1959 work: it assumes free cash flow grows at a constant rate g forever after the forecast period.

That 76% is typical for a mature company and it is the single most important thing to understand about DCFs. Most of the answer comes from a number that rests on two assumptions, the perpetual growth rate and the discount rate, applied to a cash flow five years out. A model that looks rigorous because it has detailed annual rows can still be driven almost entirely by its last line.

The terminal growth rate needs a hard ceiling. A company cannot grow faster than the economy it operates in forever, or it would eventually become the economy. Analysts usually pick a rate near long-run inflation or nominal GDP growth, often 2% to 3% for a US company, and never one at or above the discount rate, where the formula divides by zero or a negative number.

Step 5: from enterprise value to value per share

Discounting free cash flow gives the value of the operating business to all its capital providers. To get the value that belongs to shareholders, subtract debt and add back cash and investments that the forecast cash flows do not already count.

Read that gap carefully. It does not mean the stock was mispriced by half. It means the market price was consistent with very different assumptions from the ones typed into this outline: faster growth, a lower discount rate, a higher terminal growth rate, or some mix. The DCF's real job is to make those differences visible.

A note on the bridge: we counted Apple's non-current marketable securities as cash. They are mostly high-grade fixed-income holdings that Apple could sell, so many analysts do the same. Counting only cash and current securities ($54,697 million) would cut the net cash to a net debt of $43,960 million and lower the per-share result by about $5.26. State whichever definition you use. Other items that sometimes belong in the bridge include minority interests, pension deficits, lease liabilities and preferred stock; for Apple they are not material to this outline.

Step 6: stress-test it with a sensitivity grid

Because a DCF is only as good as its inputs, every DCF should come with a sensitivity table that shows how the answer moves when the key assumptions move.

The grid shows three things.

• The discount rate matters more than near-term growth. At 6% growth, moving the discount rate from 9% to 8% lifts the value from $132.87 to $159.33 (+20%); moving it to 10% cuts it to $113.98 (−14%). Changing five-year growth from 6% to 8% at 9% adds only about 9% (to $144.40).
• The effect is asymmetric. Lowering the discount rate raises the value more than raising it by the same amount lowers it, because the terminal value's denominator (r − g) shrinks.
• The range is wide. Across this modest grid the result runs from about $105 (4% growth, 10% discount) to about $189 (10% growth, 8% discount).

Two more tests outside the grid: cutting five-year growth to 4% and terminal growth to 2.5% gives $114.75; extending the 6% growth phase to ten years before the 3% terminal gives $146.63. A model whose conclusion survives most of its sensitivity table is telling you something. One whose conclusion flips inside the table is telling you the conclusion depends on an input you cannot know precisely. That is the bridge from DCF to the Margin of safety: Benjamin Graham’s rule for valuation errors idea.

Step 7: run it backward (reverse DCF)

Instead of asking "what is the stock worth given my assumptions?", a reverse DCF asks "what assumptions does the current price imply?" You hold the discount rate and terminal growth fixed and solve for the growth rate that makes the model equal the market price.

A reverse DCF is often more useful than a forward one because it starts from a fact (the price) and turns it into a question you can judge against history: has the company ever grown free cash flow that fast, for that long? Apple's free cash flow was $99.6 billion in fiscal 2023, $108.8 billion in fiscal 2024 and $98.8 billion in fiscal 2025. Whether the implied rates are plausible is exactly the debate investors have, and this page takes no side in it.

Where the discount rate comes from: WACC and CAPM

The discount rate is the return investors require for bearing the company's risk. For an enterprise DCF it is usually the weighted average cost of capital (WACC): the blended required return of shareholders and lenders, weighted by how much of each the company uses.

The cost of equity is usually estimated with the capital asset pricing model (CAPM), formalized by William Sharpe in 1964 and others. It adds to a risk-free rate, typically a long-term Treasury yield, a premium for risk: the stock's beta (its sensitivity to market moves) times the equity risk premium (the extra return investors expect from stocks over Treasuries). The cost of debt is the yield the company would pay on new borrowing, reduced by the tax shield because interest is tax-deductible.

Each CAPM input is itself an estimate. Beta depends on the period and index you measure it against; different data sources publish different betas for the same stock. The equity risk premium has no single agreed value, and published estimates vary by several points. That is why the discount rate deserves the same sensitivity treatment as growth, and why many practitioners pick a round rate and test a range rather than claiming precision to a decimal.

The alternative terminal value: an exit multiple

Instead of the perpetuity formula, many bankers value the terminal year by applying a market multiple, for example EV/EBITDA or EV/FCF, to the final year's figure. This exit multiple method is quick and anchored to how similar companies trade today.

The lesson is that the two methods should cross-check each other. Convert an exit multiple into its implied perpetual growth rate, and a perpetual growth rate into its implied multiple. If an exit multiple implies growth far above the economy's, the multiple is borrowing today's market optimism and importing it into a valuation that is supposed to be independent of the market.

The mid-year convention

The basic model assumes each year's cash arrives on the last day of the year. In reality it arrives through the year. The mid-year convention discounts each year's cash flow by (1 + r) raised to t − 0.5 instead of t, as if it all arrived at mid-year.

The adjustment is small relative to the uncertainty in the other inputs, but it is a common source of mismatch when comparing two analysts' models. State whether you use it.

Common mistakes

• Terminal growth near the discount rate. As g approaches r, the terminal value heads toward infinity. A 4% gap between them is very different from a 1% gap.
• Inconsistent growth and reinvestment. Growth requires investment. Forecasting rising free cash flow while capital spending stays flat assumes growth for free. A rough check: the implied return on new investment should be plausible for the business.
• Mismatched cash flow and discount rate. Unlevered free cash flow goes with WACC; cash flow to equity goes with the cost of equity. Mixing them double-counts or omits debt.
• Double-counting cash. If your cash flow includes interest income from a cash pile, adding the cash pile in the equity bridge counts it twice.
• Ignoring dilution. Use a share count that reflects options and restricted stock, or subtract stock-based compensation from free cash flow.
• Hockey-stick forecasts. Years one to four look like the past; year five jumps; the terminal value multiplies the jump forever.
• False precision. Reporting $132.87 suggests a certainty the inputs don't support. Report a range and the assumptions behind it.
• Reverse-engineering a target. Adjusting inputs until the model agrees with a price you wanted to see. A reverse DCF does this honestly, out loud.

DCF compared with its neighbors

The Valuation methods: intrinsic, relative and asset-based compared guide compares the wider set of approaches, and the Comparable company analysis: how to build and read a comps table page covers the multiples side in depth.

How Finabase does it

Finabase does not run DCFs or publish intrinsic values. It gives you the inputs. The Financials tab has annual and quarterly cash flow statements with operating cash flow, capital expenditure, share-based compensation and dividends, and balance sheet lines for cash, current and non-current investments and current and non-current debt; click any row to chart its history, which is how you check whether a starting year is representative. Free cash flow itself is not a statement row: subtract capex from operating cash flow yourself. The Valuation tab charts ten fiscal years of P/FCF, FCF yield, EV/EBITDA and free-cash-flow growth, which help you sanity-check a terminal multiple against the company's own history. The summary page's key statistics show the data provider's enterprise value, total cash, total debt, shares outstanding and beta. The valuation tab's "Financial health" card also shows an FMP rating with a DCF sub-score; that is the data provider's own score, not a Finabase valuation.

Frequently asked questions

What is a DCF in simple terms?

A DCF values a company as the sum of all the cash it is expected to generate in the future, with each future year's cash reduced to today's dollars at a rate that reflects risk. It is the present value of future free cash flow, plus net cash, divided by the number of shares.

Why is the terminal value so large in a DCF?

Because it stands in for every year after the forecast period, which for a long-lived company is most of its life. In the Apple outline it was 76% of enterprise value, which is common for mature businesses and is why terminal assumptions deserve the most scrutiny.

What discount rate should I use in a DCF?

Most analysts use the weighted average cost of capital for unlevered free cash flow, built from a CAPM cost of equity and an after-tax cost of debt. Because its inputs are estimates, test a range, for example 8% to 10% for a large established company, and see how much the result moves.

What is a reverse DCF?

A reverse DCF solves for the growth rate that makes a DCF equal to today's share price, holding the discount rate and terminal growth fixed. For Apple at $255.46, it implied about 22.4% annual free-cash-flow growth for five years at a 9% discount rate and 3% terminal growth.

Is a DCF accurate?

A DCF is precise but not accurate in any guaranteed sense: its arithmetic is exact, while its inputs are forecasts. Its value is in making assumptions explicit and testable, not in producing a single correct price.

Summary

A discounted cash flow valuation adds up the present value of a company's forecast free cash flow and a terminal value, adjusts for debt and cash, and divides by the share count. Built on Apple's fiscal 2025 free cash flow of $98.8 billion with 6% growth, a 9% discount rate and 3% terminal growth, the outline gives $132.87 a share, three-quarters of it from the terminal value, against a $255.46 market price. A sensitivity grid shows the same model spanning roughly $105 to $189, and a reverse DCF shows the price implied about 22.4% a year of five-year growth. Those numbers are an educational outline, not a fair value. The lasting lessons are to match cash flows to discount rates, keep terminal growth below the economy's, cross-check exit multiples against their implied growth, and always report a range with the assumptions that drive it.