One patent or an entire portfolio? A framework for valuing biotech IP

When a drug earns billions of dollars a year, it is tempting to say that its key patent is worth billions. That sentence merges two different things. A blockbuster is a successful product. Its revenue reflects clinical benefit, manufacturing, regulatory approval, marketing, reimbursement, and competitive conditions, of which patents are only one part. The useful question for a founder, a technology transfer office, or a licensee is narrower: what do these particular rights add to this particular program?

The answer differs across cases. Sometimes a small number of rights account for much of an asset's commercial protection. Sometimes protection depends on several complementary rights that only work together. A platform patent may reach several products or licensees at once. Counting patents cannot distinguish these situations. This Note offers a practical framework for doing so. It is a synthesis of established valuation methods and the empirical literature, not a validated universal scoring formula. The illustration focuses on U.S. therapeutic biotechnology and pharmaceuticals; diagnostics, research tools, industrial biotechnology, and other jurisdictions can require different assumptions about regulation, pricing, and competition.

Three different objects of valuation

Before estimating anything, it helps to be clear about what is being valued. The first object is the whole technology or program: the molecule, its data, its development plan, and its expected commercial future. The second is a transaction package, such as an exclusive license that bundles patents with data, materials, and knowhow (the unpatented practical expertise needed to use a technology). The third is the incremental contribution of specified patent rights, meaning the difference those rights make to the program's expected economics.

These three quantities are related, but they are not interchangeable. A license price reflects what was bundled and who had bargaining power. A program valuation reflects everything the program owns. Only the third quantity speaks to the economic role of the patents themselves. WIPO's guide to valuation in this sector describes the two standard tools used here: the income approach, which values an asset by its expected future cash flows, and the market approach, which reasons from comparable transactions [1]. The incremental counterfactual below is this Note's application of those tools, not a rule quoted from that guide.

A six-step framework

1. Define the valuation question

A valuation answers a question asked on a date, in a jurisdiction, from a perspective. The value of rights to the owner who practices them can differ from their value to a licensee or to an acquirer with different products. The object must also be precise: a single patent, a patent family (applications and patents that share a common priority claim, although family definitions vary across databases), a broader portfolio, or rights granted under a license. A pending application is not an issued claim; its eventual scope is uncertain. Family membership helps identify related filings, but members can differ in claims, jurisdictions, and legal status. The analysis should map each member's incremental commercial effect rather than assume that every filing represents an independent invention or an equal unit of value. Finally, owning a patent gives the right to exclude others; it does not give permission to make the product. Freedom to operate, meaning the ability to commercialize without infringing other parties' rights, is a separate question (discussed in an earlier Note on patent strategy and business strategy).

2. Map claims to commercial activities and competitor alternatives

The claims, not the title or abstract, define a patent's legal scope. The second step is to map each relevant claim to what actually generates revenue: the molecule or sequence, the formulation, the approved indication or method of use, the delivery approach, the manufacturing process, or an underlying platform. The map is then compared with what a competitor could feasibly do instead. A narrowly worded claim covering a feature that every viable competitor needs may matter more than broad language that competitors can easily avoid. The question is always about the alternatives that remain open.

3. Evaluate the durability and practical effect of protection

A claim's expected value depends on the probability that it survives a challenge and can be enforced in practice, and even a claim later invalidated can affect value through the cost, delay, and uncertainty a challenge creates. The valuation should therefore model plausible legal outcomes and their practical consequences rather than assume that every listed claim will survive. The relevant factors include the scope that the specification actually supports, validity against prior art, enforceability, remaining term, the ability to detect infringement, the feasibility of designing around the claims, third party rights that may block the owner, and the cost of litigation. Supported scope deserves particular attention in biotechnology. In Amgen v. Sanofi, the Supreme Court held that the broad functional antibody claims at issue were not enabled across their full scope, because the patent did not teach skilled workers to make and use the full class of antibodies claimed [2]. The general lesson for valuation is that claim breadth on paper and defensible breadth are different inputs. This is not a conclusion about any other patent.

4. Translate competitive conditions into program cash flows

The core of the method is a comparison of two explicit scenarios. Each is valued as a risk adjusted net present value (rNPV): projected cash flows are weighted by the probability that the program reaches each stage, then discounted to the valuation date to reflect the time value of money [1]. The incremental contribution of a defined patent set P is then:

V(P) estimates an incremental contribution under a stated counterfactual. It is not necessarily a transaction price, which also depends on bundled assets and negotiation. The counterfactual needs care. Removing P means removing only the evaluated rights. The technology, the accumulated clinical data, the knowhow, any regulatory exclusivity, and other independent advantages remain. The product does not vanish, and other barriers to competition do not disappear. Both scenarios should also allow realistic strategy: in the world without P, the owner might rely on unaffected rights, pursue feasible alternative designs, adjust timing, or partner, while competitors still face their own development costs and delays. The removed rights should not be quietly recreated through a hypothetical new filing. The two scenarios can even produce the same entry timing if other barriers still bind; protection is not assumed to improve cash flows automatically.

Clinical risk should enter once and in the right place. Success probabilities determine which revenues and costs are expected to occur. If a development plan begins Phase 3 only after Phase 2 succeeds, Phase 3 costs are weighted by the probability of reaching Phase 3, not by the final probability of approval. Probabilities should be appropriate to the stage and indication. Discounting then reflects time and the appropriate cost of capital; once clinical probabilities are in the cash flows, the discount rate should not be raised again to count the same clinical risk.

5. Compare patent sets to test portfolio dependence

The same calculation can be repeated with different sets of rights included. This reveals how the rights relate. Some are complements: each matters mainly in combination with others. Some are substitutes: any one of them blocks the same competitor path. Some are redundant for the competitor path being modeled. These relationships are properties of claims and alternatives, so they cannot be read off a patent count.

6. Cross-check against deal evidence and report a range

The income approach should be checked against the market. Comparable licenses can show whether the modeled value is plausible, provided they are truly comparable in stage, indication, and setting. Shah, Vaughan and Ledley find that biotechnology license terms vary with institutional setting and development stage, comparing academic licenses with those between commercial firms [3]. Their study is openly available, but the underlying transaction database it draws on is commercial, and a sample median is not a recommended royalty for any particular deal.

Deal terms also need decomposition. Upfront payments are near certain; milestones and royalties are contingent on events that may never occur. Adding headline maxima produces the familiar large announced deal value, which is not a present value. Finally, the private economic value of exclusivity to its holder is a different quantity from the social or therapeutic benefit of the medicine. A valuation of the former says nothing about the latter.

A worked example: equal portfolios, different structures

Consider two hypothetical programs with the same economics in the absence of the evaluated rights. The table reports program rNPV under three scenarios. These are program values, not patent values.

Three incremental results follow from equation (1). The full portfolio contributes 230 minus 50, or 180, in both programs. The key patent alone, relative to the baseline, contributes 200 minus 50, or 150, for Program A, and 70 minus 50, or 20, for Program B. The additional contribution of the remaining rights, conditional on already holding the key patent, is 230 minus 200, or 30, for A, and 230 minus 70, or 160, for B.

The aggregate contribution is identical, yet the structures differ. In Program A, one patent delivers most of the protection. In Program B, the key patent does little on its own, and value depends on the rest of the portfolio. Two cautions apply. First, these numbers do not establish a unique fair allocation of value among individual patents, and they do not tell us what would be lost by removing the key patent from the full portfolio; that would require a fourth scenario (the portfolio without the key patent). Second, standalone appraisals of each patent cannot simply be added to value a complementary portfolio, because complements are worth more together than the sum of their separate values.

A credible version of this exercise reports a range and its sensitivities. The most influential assumptions are usually competitor entry timing, the speed and depth of price erosion after entry, clinical success probabilities, launch delay, the likely outcome of relevant legal challenges, and the feasibility of design alternatives. The numbers are hypothetical and are not drawn from any study cited here.

What the public evidence shows, and what it does not

Humira (adalimumab) illustrates why first patent expiry, product sales, and effective protection are separate variables. AbbVie's 2016 annual report states that the U.S. composition of matter patent for adalimumab, which covers the molecule itself, expired in December 2016, while other relevant patents remained [4]. AbbVie's 2023 report shows U.S. Humira net revenues of $18.619 billion in 2022 and $12.160 billion in 2023, a 34.7% decline, which the company attributes primarily to direct biosimilar competition following loss of exclusivity on January 31, 2023 [5]. These are U.S. figures, not global sales. The case shows that protection continued well after the first expiry. It does not isolate the dollar value of any single patent, and it does not show that patents alone explain the entire interval between 2016 and 2023.

Columbia University's Axel patents, which covered methods for introducing foreign DNA into host cells, illustrate platform value. Colaianni and Cook-Deegan report approximately $790 million in licensing revenues for Columbia and the inventors [6]. That figure is cumulative historical revenue earned across many downstream products and licensees. It is not the sale price or present value of one isolated patent.

Two recent studies show why portfolios deserve study and why counts are not values. Frakes and Wasserman examine patents associated with 515 FDA approved biologics and find denser patenting, and stronger barriers to biosimilar entry, than in their small molecule comparison [7]. Because their data link patents to drugs, a patent can appear for more than one product, so counting definitions matter, and the finding does not imply that every listed patent is individually necessary. Durvasula and Ouellette, analyzing small molecule drugs, find substantially more patents per drug alongside comparatively stable average measured market exclusivity [8]. Their measure uses first generic approval as a proxy for entry (or patent and regulatory expiry when no generic exists), and it is not a causal dollar appraisal. The two studies ask different questions with different data, so one does not refute the other. Together they caution against assuming that more patents automatically mean more valuable protection.

A further reading point concerns so called secondary patents. Kapczynski, Park and Sampat coded the claims of 1,304 Orange Book patents covering new molecular entities approved from 1988 to 2005, and found that, where present, independent formulation patents added an average of 6.5 years of nominal patent life [9]. Nominal patent life is not guaranteed market exclusivity, and "secondary" does not automatically mean weak or economically unimportant.

Data sources for model inputs

Public data can support a defensible range and transparent assumptions. There is, however, no comprehensive public table of the true value of each biotech patent. The following sources map to specific inputs, each with limits.

Future work

The framework suggests a modest research agenda. None of the following has been carried out; these are proposed studies, not results.

• Build a small claim to product map for a handful of approved therapeutics.
• Align listed patent expiry dates with actual competitor entry dates.
• Compare modeled removal of key families with removal of groups of secondary patents.
• Test whether incremental protection or licensing optionality explains observed value better than raw patent counts.

Conclusion

Asking what a patent is worth is really asking what changes if it is absent. Framing the question that way forces precision about the rights, the competitors, and the alternatives, and it keeps clinical, legal, and market risk in their proper places. The result is a range under stated assumptions rather than a single impressive number, which is exactly what makes it useful.

This Note is for educational purposes only and does not provide legal, financial, or investment advice.

References

1. [1] World Intellectual Property Organization. Intellectual Property Valuation in Biotechnology and Pharmaceuticals. WIPO web publication. See sections "2 The market approach" and "3 The income approach".
2. [2] Supreme Court of the United States. Amgen Inc. v. Sanofi, 598 U.S. 594 (2023). Slip opinion, No. 21-757.
3. [3] Prateet Shah, Gregory Vaughan and Fred D. Ledley. Comparing the economic terms of biotechnology licenses from academic institutions with those between commercial firms. PLOS ONE 18(3), e0283887 (2023). DOI 10.1371/journal.pone.0283887.
4. [4] AbbVie Inc. Annual Report on Form 10-K for the fiscal year ended December 31, 2016. U.S. Securities and Exchange Commission, EDGAR. Section "Patents, Trademarks and Other Intellectual Property".
5. [5] AbbVie Inc. Annual Report on Form 10-K for the fiscal year ended December 31, 2023. U.S. Securities and Exchange Commission, EDGAR. Product revenue table and Humira discussion.
6. [6] Alessandra Colaianni and Robert Cook-Deegan. Columbia University's Axel Patents: Technology Transfer and Implications for the Bayh-Dole Act. The Milbank Quarterly 87(3), 683 to 715 (2009). DOI 10.1111/j.1468-0009.2009.00575.x.
7. [7] Michael D. Frakes and Melissa F. Wasserman. Unveiling the Patent Landscape of Biologic Drugs. Northwestern University Law Review 120, 1315 (2026).
8. [8] Maya M. Durvasula and Lisa Larrimore Ouellette. Beyond the Pharmaceutical Patent Arms Race. Yale Journal on Regulation 43, 703 (2026).
9. [9] Amy Kapczynski, Chan Park and Bhaven Sampat. Polymorphs and Prodrugs and Salts (Oh My!): An Empirical Analysis of “Secondary” Pharmaceutical Patents. PLOS ONE 7(12), e49470 (2012). DOI 10.1371/journal.pone.0049470.
10. [10] U.S. Food and Drug Administration. Orange Book Data Files.
11. [11] National Bureau of Economic Research. Orange Book Patent and Exclusivity Data, 1985 to 2016.
12. [12] U.S. Food and Drug Administration. Purple Book Frequently Asked Questions. See especially Q5 to Q7 on patent information.
13. [13] U.S. Patent and Trademark Office. Patent Claims Research Dataset.
14. [14] BIO, Informa Pharma Intelligence and QLS Advisors. Clinical Development Success Rates and Contributing Factors 2011 to 2020. Biotechnology Innovation Organization (2021).