June 5, 2026 · 13 min read
DCF, P/E multiple, EV/EBITDA, margin of safety — a step-by-step guide to figuring out what a stock is actually worth and when to buy it.
Intrinsic value is what a business is actually worth — independent of whatever price the stock market happens to be quoting today. It is the present value of all future cash flows the business will generate for its owners, discounted back at a required rate of return.
The stock market is famously Mr. Market: Benjamin Graham's fictional manic-depressive character who shows up every day offering to buy or sell your shares at a different price. Some days he's euphoric and offers too much. Other days he's panicked and will sell too cheap. Intrinsic value is the anchor that lets you evaluate whether Mr. Market's offer makes sense.
Markets are driven by sentiment, momentum, and short-term news cycles — not intrinsic value. A stock can trade 50% above or below fair value for years. Earnings surprises, index rebalancing, macro fear, and meme-stock frenzy all move prices in ways unrelated to the underlying business. The job of a fundamental investor is to exploit this gap.
Warren Buffett summarizes it simply: "Price is what you pay; value is what you get." The entire discipline of value investing is built on the idea that you can estimate intrinsic value well enough to know when the market is offering you a bargain.
No single method is universally correct — each makes different assumptions about growth, risk, and comparables. Smart analysts run two or three and look for convergence.
DCF is the gold standard of intrinsic value: project future free cash flows, choose a discount rate, and compute the present value of all those cash flows — including a terminal value that captures everything beyond the forecast window. It is theoretically rigorous but highly sensitive to assumptions. A 1% change in terminal growth rate can shift the output by 20% or more.
Multiply normalized (through-the-cycle) EPS by a 'fair' P/E — typically the sector median, company historical average, or a growth-adjusted PEG-derived multiple. Simple and widely used. Works poorly for unprofitable companies, highly cyclical businesses, or firms going through structural change. Example: NVDA at 35× forward EPS of $4.20 = IV of ~$147.
Enterprise Value / EBITDA strips out capital structure and tax differences, making it the preferred tool for comparing companies with very different debt loads or depreciation policies. A utility trading at 8× EBITDA vs. a sector median of 12× may be significantly undervalued. Subtract net debt from EV to get equity intrinsic value, then divide by shares outstanding.
Let's walk through a full DCF using a fictional company "StockCo": profitable software-adjacent business, $1B annual FCF, 8% FCF growth expected for 5 years, 10% WACC, 3% terminal growth rate, 500M shares, $1B net debt.
Start with the most recent annual FCF. Apply a realistic growth rate — e.g., StockCo earned $1B FCF growing 8%/yr. Year 1: $1.08B, Year 2: $1.17B … Year 5: $1.47B. Be conservative; the market already prices in optimism.
The discount rate reflects your required return. Most analysts use 8–12% for equities. Higher-risk businesses or rising interest rate environments warrant 10–12%. For StockCo (stable, low-debt), use 10% WACC.
After Year 5, assume growth stabilizes at 3% (GDP-level). Terminal Value = Year 5 FCF × (1 + 3%) ÷ (10% − 3%) = $1.47B × 1.03 ÷ 0.07 = $21.6B. This single number often drives 60–70% of total DCF value.
Divide each year's FCF by (1 + 0.10)^n. Year 1: $1.08B ÷ 1.10 = $0.98B … Year 5 TV: $21.6B ÷ 1.61 = $13.4B. Sum all discounted cash flows: ~$4.2B (5-yr FCF PV) + $13.4B (TV PV) = ~$17.6B total enterprise value.
StockCo has 500M shares and $1B net debt. Equity value = $17.6B − $1B = $16.6B. IV per share = $16.6B ÷ 500M = $33.20. If the stock trades at $25, margin of safety = 25% — potentially a buy.
If StockCo trades at $25, the margin of safety is 25% — within Graham's recommended buying zone. If it trades at $40, it is overvalued by ~20% on these assumptions.
NVIDIA is not a textbook DCF candidate — its growth is explosive and unpredictable — but it illustrates how to stress-test assumptions on a high-profile stock.
Running a standard 5-year DCF with these inputs produces an estimated intrinsic value of $120–$150 per share under base-case assumptions (15% FCF growth fading to 3% terminal). Bull case (20% growth, 3.5% terminal) stretches to ~$185. Bear case (10% growth, 2.5% terminal) compresses to ~$90.
At current prices (mid-2026), NVDA trades above the base-case DCF range, implying the market is pricing in supernormal growth beyond 5 years. This is not irrational — it reflects AI infrastructure demand — but it means there is no margin of safety on a strict DCF basis. Position sizing and scenario analysis become critical.
The P/E multiple approach is faster but requires a defensible "fair multiple." The standard approach: estimate normalized forward EPS, then apply a peer-comparable or historically grounded P/E.
At 35× (below NVDA's recent historical average of 45×+, above the sector median of 28×), the P/E method points to ~$147 intrinsic value — converging with the DCF base case. Convergence across methods increases confidence.
EV/EBITDA is the preferred valuation lens for capital-intensive sectors: utilities, telecom, industrials, REITs (using different metrics), and any business with heavy debt or significant D&A. It removes the noise of capital structure and tax treatment.
NextEra has EBITDA of ~$9B. The regulated utility sector median EV/EBITDA is ~12×. Fair EV = 9B × 12 = $108B. With $60B net debt and 2.05B shares: Equity IV = ($108B − $60B) ÷ 2.05B = ~$23.40/share. If NEE trades at $75, the EV/EBITDA method signals extreme overvaluation on pure earnings power — growth premium must justify the gap.
EV/EBITDA works well for comparisons within a sector. Cross-sector comparisons (e.g., tech vs. utility) are meaningless because appropriate multiples differ enormously.
Even the best intrinsic value estimate is an estimate. Analysts are wrong. Management outlooks are optimistic. Black swans exist. The margin of safety is the buffer between your estimated value and the price you actually pay — it is your protection against being wrong.
Price is close to or above intrinsic value. One wrong assumption and you lose money.
Acceptable for high-quality compounders with near-certain futures. Still risky for cyclicals.
Graham's recommended minimum. Absorbs moderate estimation error without causing permanent loss.
The fat pitch. Rare in bull markets but present in sector dislocations and earnings panics.
Buffett extended Graham's margin of safety concept: for truly exceptional businesses with durable moats, a smaller safety margin is acceptable because the compounding power of the business itself provides the buffer over time. For average businesses, stick to 25–30%+.
The table below shows how StockCo's intrinsic value per share changes across different discount rates (WACC) and terminal growth rate assumptions. The center cell ($33.20) is the base case from the walkthrough above.
| WACC \ Terminal Growth | 2% Terminal | 3% Terminal | 4% Terminal |
|---|---|---|---|
| WACC 8% | $41.10 | $48.80 | $59.60 |
| WACC 10% | $29.40 | $33.20 | $38.50 |
| WACC 12% | $21.70 | $24.10 | $27.30 |
Note how moving from 3% to 4% terminal growth rate at 10% WACC increases IV by $5.30 (+16%). Moving from 10% to 8% WACC with 3% terminal growth increases IV from $33.20 to $48.80 (+47%). Discount rate assumptions are the most powerful — and most contested — variable in any DCF.
The DCF is only as good as your FCF projections. Basing growth estimates on the last 3 exceptional years (pandemic boom, AI supercycle) without normalizing for reversion to mean produces wildly inflated values.
No company can grow faster than GDP indefinitely. A terminal growth rate above 3–4% implies the company will eventually be larger than the global economy. Use 2–3% for most businesses; 3.5% only for truly exceptional global franchises.
DCF models built during the 2021 zero-rate era with 6% WACCs produced drastically higher valuations than the same models run today at 10% WACC. Rising interest rates mechanically reduce the present value of long-duration assets. Always run your model with current risk-free rates.
A DCF that outputs $33.20/share is not a precise target — it is the midpoint of a wide range. Think in ranges: $25–$45 is the honest output. Point estimates create false precision and encourage anchoring.
Comparing a company to the average P/E of a sector index ignores that failed companies are removed from indices. The surviving companies deserve premium multiples. Adjust peer comparison accordingly.
DCF requires positive, predictable free cash flows. For many interesting investment opportunities, it is the wrong tool entirely.
Companies like early-stage SaaS or pre-profitability biotech have no FCF to discount. Use EV/ARR or EV/Revenue with a sector growth premium, or TAM penetration analysis to estimate terminal-state earnings.
A biotech's value is a portfolio of binary outcomes (drug approval / no approval). Probability-weight each drug's potential peak sales by its phase success rate. This is a risk-adjusted NPV, not a traditional DCF.
Oil majors, miners, steelmakers, and autos have highly variable earnings. Never run DCF off a peak-cycle year. Normalize earnings over a full commodity cycle and apply a mid-cycle multiple. Buying cyclicals near peak earnings at 'cheap' P/E is the classic value trap.
You do not need to build a DCF spreadsheet from scratch. These tools streamline the process:
Full DCF models, Graham Number, and Peter Lynch value for thousands of stocks. Free tier available.
Visual DCF charts with bull/bear scenario toggles. Great for beginners.
Historical P/E, P/FCF, EV/EBITDA, and FCF data going back 20+ years. Excellent for normalizing.
Our AI-powered intrinsic value tool runs DCF, P/E, and Graham Number in one click — no spreadsheet needed.
Is the intrinsic value estimate based on conservative, realistic FCF projections (not analyst best-case)?
Does the current market price offer at least a 20% margin of safety vs. your IV estimate?
Have you stress-tested with multiple discount rates and terminal growth rates?
Does the company have a moat (competitive advantage) that justifies a low discount rate?
Have you accounted for net debt/cash so you're comparing equity value — not enterprise value — to the share price?
Intrinsic value is not a magic number — it is a disciplined framework for separating price from value. No single method is definitive, and all require honest assumptions about growth, risk, and competitive durability.
The most powerful insight from intrinsic value analysis is not the number itself: it is the process of understanding why a business generates cash, how durable that cash generation is, and what rate of return justifies owning it. That understanding is what separates investors from speculators.
BriMindInvest's Intrinsic Value tool runs DCF, P/E-based, and Graham Number estimates for any publicly traded stock — no spreadsheet needed. Get the AI-powered BriMind Score alongside the numbers.
Get AI prediction signals, unlimited stock comparisons, portfolio analytics, and personalized watchlists — free for 14 days, no credit card required.
14-day free trial · No credit card required · Cancel anytime