A DCF has only three real inputs — cash flows, growth, and risk (the discount rate). Its power is that it forces consistency: growth is not free; higher growth requires more reinvestment, and the two must be linked. Its limits lie not in the math but in the inputs. Damodaran stresses two iron rules: ① the perpetual growth rate cannot stay above the risk-free rate (≈ nominal GDP growth), or the firm eventually outgrows the whole economy; ② terminal value often makes up 60–80% of total value, meaning your assumptions about "ten years out" dominate today's conclusion. Faced with uncertainty, his stance is to confront it, not deny or freeze — "Don't hide from it. Don't go into denial."
A DCF is hypersensitive to its assumptions. Suppose a company is worth 100 per share at an 8% WACC and 2.5% perpetual growth. The table below shows how value swings when the discount rate and growth each move by just ±1% — exactly the trick in sell-side reports during the 2000 dot-com bubble: perpetual growth far above the economy plus minimal reinvestment, used to "compute" any price you like.
| WACC ↓ | Perpetual growth g → | ||
|---|---|---|---|
| 2.0% | 2.5% | 3.0% | |
| 7% | 110 | 122 | 138 |
| 8% | 92 | 100 | 110 |
| 9% | 79 | 85 | 92 |
Where it breaks: ① cyclical resource/shipping firms at the top, extrapolating today's high cash flows → severe overvaluation; ② early-stage loss-makers with no cash-flow baseline to extrapolate; ③ stuffing inconsistent assumptions (high growth + low reinvestment + high margins all at once) into the model → "precisely wrong."
Damodaran splits valuation in two: intrinsic valuation (DCF) answers "what is this company worth?"; relative valuation answers "is it priced reasonably versus peers?" The latter is what 90% of Wall Street "valuation" really is. The trap: multiples import the market's collective error wholesale. When the whole sector is in a bubble, "cheaper than peers" offers no safety — you are merely the relatively cheap one inside the bubble. He distinguishes the investor (who cares about value) from the trader (who cares about price): relative valuation is fundamentally a pricing game, earning money from "price reverting to the multiple," not from value creation.
Cisco, March 2000: around $80 a share, a market cap of about $555 billion — briefly the world's most valuable company — at a P/E above 130. Analysts argued "relative to other networking-equipment stocks, Cisco is fairly valued" — but the whole sector was in a bubble. After the Nasdaq peaked, Cisco fell to roughly $8–13 by 2002 (down over 85%) and did not retouch its 2000 high until around 2024. Fundamentals didn't save it — Cisco remains a fine company to this day, yet it died of "cheap relative to peers."
Relative valuation is not worthless — for early-stage or cyclical industries lacking reliable cash-flow forecasts, it is fast, comparable, and close to market reality. Where it breaks: ① the entire comp set is in a bubble or a panic together; ② the "comparables" actually differ wildly in growth, risk, and ROIC, making multiples incomparable; ③ a single multiple (e.g., P/E) ignores differences in capital structure and accounting.
Pure number-crunchers (eyes only on the model) are captured by the illusion of precision; pure storytellers (vision only) fall in love with the narrative and invent reasons for impossible growth. Damodaran's discipline binds the two: the story sets TAM, market share, margins, and reinvestment, which become cash flows; in return, the common-sense bounds of the numbers (e.g., perpetual growth ≤ risk-free rate) pull the story back to earth. The greatest danger is not the absence of a story, but falling in love with your own story — explaining away every contradicting fact as "the market just doesn't get it yet."
Uber, June 2014. Damodaran published a DCF valuing it at about $5.9 billion, on the story that "the urban car-service market is ~$100B and Uber takes a 10% share." Private markets already valued it at $17B. Investor Bill Gurley countered: Uber isn't grabbing share of existing taxis — it is expanding the whole market (displacing car ownership, creating new trips), so the TAM is far larger. The outcome: Uber went public in May 2019 at roughly $82 billion ($45 IPO price). Damodaran later admitted publicly — the error was not his math but that he had told the story too small, and he revised the narrative upward repeatedly. It is the most honest lesson in "the story drives the number" — even the Dean of Valuation is no exception.
Where it breaks: ① when the story can't be falsified ("will change the world in ten years"), the numbers are mere decoration; ② using the story to defend assumptions reality has repeatedly disproven (the flip side of a value trap — the "growth trap"); ③ treating each upward revision as "I see further" rather than admitting the original assumption was wrong.
"Value vs growth stock" is a false dichotomy to both Damodaran and Buffett — growth is simply a variable inside the value formula. The question is not "how fast is growth?" but "at what cost is it bought?" The formula is simple: to get growth of g, you must invest g/ROIC of capital. When ROIC > cost of capital, growth is positive; when ROIC < cost of capital, growth is burning money and destroying value — the faster it grows, the more it destroys. That is why some companies become worth less the more they "grow." As Buffett said in 1992, value and growth are "joined at the hip."
WeWork. In January 2019, after a SoftBank injection, it was valued at about $47 billion, with a dazzling growth story (members, cities, signed square footage all expanding fast). But unit economics were negative — it swapped long-term leases for short-term memberships, ROIC deeply negative, growth merely accelerating the cash burn. After the September 2019 prospectus exposed the real losses, the IPO collapsed; it went public via SPAC in 2021 at about $9 billion; in November 2023 it filed for bankruptcy protection. The growth was real, the value was negative — when growth has no returns behind it, expansion itself is destruction.
Boundary: an early-stage company with temporarily negative ROIC isn't necessarily bad (Amazon suppressed early profits to reinvest, but unit economics and eventual ROIC were positive). The difference: will reinvestment eventually convert into returns above the cost of capital, or is it a hole that never fills? Misuse: ① using "growth" to excuse any high valuation; ② watching only revenue growth, not return on capital; ③ mistaking cash-burning growth for a moat.