What 'intrinsic value' actually means
Intrinsic value is the present value of every dollar a business will generate for its owners, discounted back to today at a rate that compensates for risk. That's it. Everything else — earnings multiples, EV/EBITDA, P/B ratio — is a heuristic shortcut to estimate the same number without doing the cash-flow work.
The DCF model is the only valuation method that doesn't rely on a comparable. It values a business on its own production. That's why the buy-side uses it for lower-coverage names where comparables are unreliable, and why it's the only valid valuation for a private business.
The two-stage model in 30 lines of math
Stage 1: explicit forecast period (years 1–5). Project free cash flow to firm (FCFF) each year. The standard FCFF = EBIT × (1 - tax rate) + D&A − CapEx − ΔWC. Apply a growth assumption — typically the 3-year trailing FCFF CAGR, capped at 15%.
Stage 2: terminal value at end of year 5. Use Gordon growth: TV = FCFF6 / (WACC − g), where g is the long-run sustainable growth rate (we use 2.5%, the long-run US real GDP+inflation blend).
Discount everything back to today at WACC: WACC = (E/V) × Re + (D/V) × Rd × (1 - t). Re is cost of equity from CAPM: risk-free + β × ERP. Risk-free we pull from the 10-year Treasury yield. ERP we use 5.5% (the Damodaran-published implied US ERP). Beta we calculate from 60-month rolling regression vs S&P 500.
Sum the discounted FCFFs plus the discounted terminal value. That's enterprise value. Subtract net debt. That's equity value. Divide by shares outstanding. That's intrinsic value per share. Compare to the current price. If intrinsic > price by >20%, undervalued. If intrinsic < price by >20%, overvalued.
A worked example: a real ticker
Take a mid-cap industrial. TTM FCFF = $400M. 3-year FCFF CAGR = 9%. Cap to 9%. Project years 1–5: $436M, $475M, $518M, $565M, $616M. Terminal FCFF = $616M × 1.025 = $631M. Terminal value = $631M / (8.5% − 2.5%) = $10.5B.
Discount each year at WACC = 8.5%. PV of years 1–5 sums to $1.96B. PV of terminal = $10.5B / 1.0855 = $6.99B. Enterprise value = $1.96B + $6.99B = $8.95B. Net debt = $1.2B. Equity value = $7.75B. Shares out = 110M. Intrinsic value per share = $70.45.
If the stock trades at $52, the model says +35% upside to fair value. If it trades at $90, the model says -22% to fair value. /dcf/ publishes this calculation on 1,001 names with a public methodology page and a sector-aware sanity cap (3× for Financials, 20× for everything else — keeps deposit-float-driven blowups from making banks look 1700% undervalued).
What can go wrong (and how /dcf/ guards against it)
Garbage in: FCFF inputs from misclassified line items (SBC, leases, capitalized R&D). We pull from the SEC EDGAR XBRL companyfacts feed, which uses the GAAP-tagged line items directly — no scraping vendor PDFs. The /methodology/ page lists every tag we use.
Terminal value sensitivity: the terminal accounts for 60–80% of total enterprise value. Small changes in g (1% vs 3%) move intrinsic by 30%+. We hold g constant at 2.5% across the entire universe so cross-comparisons stay apples-to-apples — and we publish the assumption.
Sector blowups: financials' OCF includes deposit float, which compounds to nonsense. Same for insurance reserves. We sector-cap upside at 3× for Financials/Insurance, 20× everywhere else. The cap fires on /dcf/ with a warning icon — the user sees both the raw model and the capped value.
Why this is on /dcf/ behind the Pro paywall
Bloomberg charges $24,000/year for the same calculation across the same universe. FactSet and Refinitiv are in the same range. /dcf/ is part of the Catalyst Edge Pro plan at $39/month — a 99.8% discount on the same data, with a published methodology page and an audit log on /trust/.
The free tier shows the top 3 undervalued names from the DCF run; the full table is Pro. The reason the table is paywalled isn't compute cost — it's that DCF intrinsic value is the highest-LTV signal we generate. Everyone who uses it heavily is a Pro upgrade lead. We'd rather give you the methodology in public and gate the live data than do it the other way around.
/preview/ drops you into the free tier; /pricing/ opens the Pro upgrade. If your trading workflow uses DCF intrinsic value at any frequency, $39/month is a rounding error against the alternative.
Get the same picks our paying readers get — free
Top 3 of /scanner/ and /jackpot/, the audited /trust/ log, and the daily catalyst email. No drip campaigns, no upsell sequences — one note per trading day.