Apple, from the 10-K to a fair value estimate

Apple is the most-analyzed company on earth, which makes it the right place to start: every number below is public, and the disagreements between valuation methods are easy to see.

What this example teaches

• Why book-value methods carry little or no weight for asset-light compounders
• How a sector weighting decides which methods speak loudest in a blended estimate
• Why a wide spread between methods is itself information

Model snapshot - June 11, 2026. Price $290.83. Blended model fair value estimate $270.76, which sits 6.9% below the price. The live strip above shows where these numbers stand today.

Start where the filing starts

Apple's reported fundamentals feed every method the same inputs: trailing earnings put the stock at a P/E of about 35 (about 30x on forward analyst EPS estimates of $9.57), the balance sheet shows a price-to-book ratio above 40, return on equity screens at roughly 141%, and free cash flow yield is about 2.3%.

That ROE figure is the first lesson. Apple has spent two decades returning capital and shrinking its equity base, so "equity" on the balance sheet is tiny relative to the business. Any method that anchors on book value is measuring the accounting residue of buybacks, not the business.

What each method said

Method	Model fair value (6/11/26)	vs. $290.83 price
P/E (sector-calibrated)	$247.72	-15%
EV/EBITDA	$249.78	-14%
P/S	$200.25	-31%
PEG	$420.33	+45%
EPV (earnings power)	$78.83	-73%
DCF (single-stage)	$84.87	-71%
P/B	$65.32	-78%
Graham Number	$24.35	-92%

Two clusters jump out. The multiple-based methods that compare Apple to its sector (P/E, EV/EBITDA, P/S) land in the $200-250 range. The asset- and conservatism-anchored methods (P/B, Graham Number, EPV, single-stage DCF at modest growth) land between $24 and $85 - not because Apple is worth $24, but because those formulas were built for businesses whose value lives on the balance sheet. Graham's formula multiplies earnings by book value; give it a company with almost no book value and it returns a number that mostly measures that absence.

How the blend handles the disagreement

The model does not average all of these. Each sector has a calibrated weighting, and for Technology the weight sits with the earnings- and revenue-multiple family - P/E, P/S, and PEG carry the largest shares - while pure book-value methods carry little or none. The stock page's "Why this fair value" panel lists, for every stock, exactly which methods were blended at what weight and which were excluded for the sector with zero weight - Apple's P/B and Graham values are computed and shown, but they do not move the blend.

The result on the snapshot date was a blended estimate of $270.76 - below the price, by about 7%. Read precisely, that says: under the model's sector multiples and trailing fundamentals, the price already reflects somewhat more than those inputs justify. It is a description of the gap between price and a set of stated assumptions, not a forecast of where the price goes next.

What to do with a -7% reading

A single-digit gap in either direction is well inside the model's own uncertainty for a mega-cap - method dispersion here spans hundreds of dollars. The useful exercise is the one this page is built for: open the "Why this fair value" panel, look at which methods drive the blend, and use the assumption sliders to see what growth rate or discount rate would close the gap. If it takes implausible inputs to make the price look cheap, that tells you something; if modest ones do, that tells you something too.

Figures are model estimates computed from public fundamentals under stated sector assumptions, as of June 11, 2026. They are educational illustrations, not investment advice or a prediction of future prices.