How biotech companies are valued

The question that breaks a standard toolkit

Every generalist interview candidate learns the same three-legged valuation stool: a discounted cash flow, a set of trading comps, and a set of precedent transactions. Hand that same candidate a clinical-stage biotech company with no approved products, and all three legs wobble at once. There's no revenue to build a comp multiple from. There's no earnings for a P/E. A plain DCF technically still works in the sense that you can build a spreadsheet, but it produces nonsense if you simply project the company's future cash flows as if its lead drug candidate is certain to reach the market, because in most cases it isn't. The company might be worth a great deal of money today, priced by the market on the strength of its pipeline, while its income statement shows nothing but losses.

This is the single most common valuation question asked in healthcare interviews, precisely because it can't be answered by reciting the standard framework. The sector's real answer, risk-adjusted net present value, is a variation on a DCF that explicitly builds in the chance the future cash flows never happen at all, and it's worth understanding not just as a formula but as a way of thinking about uncertainty that shows up throughout healthcare interviews.

The core idea: risk-adjusted net present value

A standard DCF assumes the projected cash flows happen and discounts them back to reflect the time value of money and, through the discount rate, some general business risk. Risk-adjusted NPV, usually abbreviated rNPV, adds a second layer of adjustment on top: before discounting, each future cash flow is multiplied by the probability that it actually occurs.

For a clinical-stage drug candidate, "it actually occurs" means the drug clears every remaining hurdle between where it sits today and generating commercial revenue: the rest of its clinical trials, regulatory approval, and successful commercial launch. Rather than pretending that path is certain, an rNPV model assigns a probability to reaching each subsequent stage and multiplies the eventual cash flows by the combined probability of getting there. A drug still in an early clinical phase gets a heavily discounted value relative to the same drug's eventual peak commercial potential, because there's a real chance it never reaches the market at all. A drug that has already cleared its final clinical hurdle and is simply awaiting a regulatory approval decision gets valued much closer to its full potential, because most of the uncertainty has already resolved.

A simplified example

None of the figures below are real-world probabilities or industry data. They're illustrative round numbers to show the mechanic, the same way a $100 million EBITDA company is a stand-in for a real deal, not a statement about any actual company or drug class.

Suppose a hypothetical biotech has one drug candidate currently in a late-stage clinical trial, and suppose an analyst estimates the following, purely for illustration: a hypothetical 50% chance the trial succeeds and the drug is eventually approved, and if it is approved, the drug reaches peak annual sales of $500 million a few years after launch, generating a steady stream of after-tax cash flow for the company. An analyst building an rNPV would take each future year's projected cash flow from that drug, multiply it by the 50% probability of ever reaching the market, and then discount the result back to today using an appropriate discount rate. If the drug fails the trial, the company gets none of that cash flow, and the model reflects that by weighting the failure scenario at its own probability, typically effectively zero cash flow, in the overall expected value.

The company's total rNPV would then be the sum of that risk-adjusted value across its entire pipeline, plus any cash on the balance sheet, minus any debt. A company with several distinct drug candidates at different stages gets modeled the same way asset by asset, since each one carries its own separate probability and its own separate potential payoff, then summed together, the same sum-of-the-parts logic used for a large diversified pharma company's whole portfolio, a business model covered in pharma business models, pipelines, and patent cliffs. Cash on the balance sheet matters more here than in most valuation contexts, because it's what funds the company between now and its next value-defining trial result, a dynamic covered in biotech IPOs and follow-on offerings.

Why probabilities rise as a drug moves through development

The mechanism worth understanding, independent of any specific number, is that the probability of eventual success generally increases the further a drug candidate has progressed through clinical development, because each successfully completed stage resolves some of the uncertainty about whether the drug is safe and effective. A candidate that has only completed an early safety-focused trial still faces the much larger hurdle of proving it actually works at the scale needed for approval. A candidate that has already produced strong results in that larger efficacy trial has cleared the hardest part of the uncertainty and mostly faces a more mechanical regulatory review from there.

The qualitative pattern, independent of any specific number, looks like this moving through development:

This is exactly why a single trial readout can move a biotech's value so dramatically: the readout isn't just new information, it's a direct update to the probability term multiplying every future cash flow in the model. A positive result increases the probability, and every future cash flow the drug could generate suddenly gets weighted more heavily. A negative result can send that probability toward zero and erase most of the value attributed to that program in a single day.

Where the discount rate fits in

There are two schools of thought on how to handle risk in the model, and interviewers sometimes probe which one you understand. One approach builds nearly all of the risk into the probability-of-success weighting itself and then discounts the resulting risk-adjusted cash flows at something closer to a normal cost of capital, on the theory that the probability weighting has already done the heavy lifting of pricing the binary risk. The other approach uses a probability weighting that is more conservative and layers an additional risk premium into the discount rate on top, on the theory that even a probability-weighted cash flow still carries risk beyond what a normal business faces. Neither approach is universally "correct," and a candidate who can explain the tension, rather than just plugging in a number, demonstrates real understanding of the model rather than a memorized formula.

Comparable methods, adapted for a company with no earnings

rNPV isn't the only tool the sector uses, and it's worth knowing the others because interviewers sometimes ask how you'd triangulate a valuation rather than rely on one method alone.

Comparable transactions still work in biotech, but they compare deal economics rather than multiples of revenue or EBITDA, since the target usually has neither. A banker looks at what acquirers have paid, historically, for similar-stage assets in a similar therapeutic area: the upfront payment, the structure and size of any milestone payments, and the royalty rate, if the deal was a licensing transaction rather than an outright acquisition. That gives a market-based sanity check on the rNPV output, similar in spirit to how a normal precedent transactions analysis checks a DCF, even though the specific inputs being compared are completely different. The mechanics of how those deals are structured, and why the milestone and royalty pieces exist at all, are in healthcare deal structures: licensing and milestones.

Comparable companies work too, but the peer set has to be built around similar-stage assets rather than similar current financials, since a company's clinical-stage peer group is a much more meaningful comparison than any conventional financial peer group when nobody in the set has meaningful revenue yet. This is a genuinely different exercise than building a normal comps set, and it's part of why biotech-focused coverage teams develop deep therapeutic-area expertise: knowing which handful of companies are truly comparable requires understanding the clinical science, not just screening for similar size or geography.

Why trading comps and standard multiples still fail

It's worth being explicit about exactly why the conventional toolkit breaks down, because interviewers like a candidate who can articulate the failure mode precisely rather than just asserting that biotech is "different." EV/EBITDA and P/E require earnings, and a clinical-stage company has none. EV/Revenue requires revenue, and a clinical-stage company usually has none either, aside from occasional upfront or milestone payments from a partner that don't represent a repeatable operating business. Even a revenue multiple built from a small commercial-stage biopharma company is often unreliable, because that company's revenue could be driven by one recently launched product still ramping toward its eventual peak, which makes a current-year multiple a poor proxy for steady-state value. The whole reason rNPV exists is that it's built specifically to handle a cash flow stream where the central uncertainty is binary, whether an asset reaches the market at all, rather than the smoother, more continuous uncertainty a standard DCF or multiple is designed to handle. Matching the valuation method to the company's actual stage, rather than defaulting to whatever multiple you'd use on a normal company, is exactly the judgment tested in a "pitch me a stock" question, covered in pitching a healthcare stock.

The follow-up chain interviewers actually run

This topic rarely gets asked as a single question. The standard chain starts with "how would you value a pre-revenue biotech," and once you answer with rNPV, expect a follow-up on how probability of success changes across clinical phases, then a follow-up on how you'd choose a discount rate, then possibly a question about how a positive or negative trial result should flow through the model. A candidate who can walk that full chain, rather than defining rNPV and stopping, is demonstrating the kind of layered understanding the sector actually rewards.

Practice question

How would you value a biotech company that has no revenue and no approved products?

I'd use a risk-adjusted net present value approach rather than a standard DCF or multiples-based method, because those tools assume either earnings or reliable near-term cash flow, and a pre-revenue biotech has neither. The idea is to project the future cash flows the company would generate if its pipeline succeeds, then multiply those cash flows by the probability that the underlying drug candidate actually reaches the market, accounting for the clinical trials, regulatory approval, and commercial launch still ahead of it. That probability is generally higher for a drug further along in development, since each completed clinical stage resolves some of the uncertainty about whether it's safe and effective. You'd do this asset by asset across the company's whole pipeline, since each program carries its own separate risk and payoff, and sum the results together along with the company's cash and any debt. I'd also sanity-check the result against comparable transactions, meaning what acquirers have actually paid, in upfront payments, milestones, and royalties, for similarly staged assets, since that gives a market-based cross-check the way precedent transactions check a normal DCF.

What the interviewer is listening for: whether you can explain the mechanism, not just name-drop "rNPV." They want to hear that probability of success varies by clinical stage, that you understand why standard multiples don't apply, and ideally that you know a second method, like precedent deal economics, exists to cross-check the answer.