A patent is a legal right. A moat is a business advantage. The two are related, but they are not the same thing, and confusing them is one of the more expensive mistakes a company can make about its own patents. The opening article described the right. This article asks when that right becomes a durable competitive advantage, an economic moat, and when it is merely an expensive certificate. Most patents are not moats. Understanding why is the difference between spending on patents that protect a business and spending on patents that protect nothing. As with the rest of the series, this is general information rather than legal or financial advice, and it assumes filing in the United States.

The Idea of a Moat

A moat, in business, is a durable competitive advantage, something that protects a company’s profits from the competition that would otherwise erode them. The term was popularized by Warren Buffett, who looked for businesses with an economic moat that competitors could not easily cross. The academic version of the same idea is the study of barriers to entry, the costs and obstacles that keep new rivals out of a market.

A patent is one classic barrier. By granting the right to exclude, it can in principle keep competitors out of a protected market for the life of the patent. That is the theory. Whether a particular patent actually does this depends on a set of conditions that are met far less often than patent holders hope. A patent is a candidate moat. The conditions decide whether the candidacy succeeds.

The right to exclude is necessary for a patent moat, but it is not sufficient. A patent can grant an airtight right to exclude from something no one wants, or from something a competitor can avoid at trivial cost, or from something the holder can never detect being copied. In each case the legal right is real and the moat is absent.

This is the gap where most patent disappointment lives. A company obtains a patent, sees the grant as a victory, and assumes the protection is now in place. The grant is a legal fact. The protection is an economic outcome, and it arrives only when the right to exclude attaches to something valuable, hard to avoid, and possible to enforce. The rest of this article is about those conditions.

The Conditions for a Patent Moat

Six conditions decide whether a patent functions as a moat. A weakness in any one of them can hollow out the whole.

The first is validity. A patent likely to be invalidated is a weak moat, because a challenger can remove it. Validity rests on the prior art, the subject of the previous article, and a patent that issued over a thin examination sits on uncertain ground.

The second is claim scope. The claims must be broad enough to cover the alternatives a competitor would use, yet narrow enough to remain valid over the prior art. This is a genuine tension. Broaden the claims to block more, and they become more vulnerable to invalidation. Narrow them to survive, and a competitor steps around them. The valuable patent lives in the band where the claims are both broad and survivable, and that band is often narrow.

The third is detectability. A right to exclude is only as good as the holder’s ability to tell when it is being violated. A patent on a visible product feature can be checked by buying the product. A patent on a process hidden inside a competitor’s own operations may be infringed for years without the holder ever knowing.

The fourth is the difficulty of designing around. If a competent rival can achieve the same commercial result by a different route the claims do not cover, and can do so cheaply, the patent protects little. The cost to design around sets a ceiling on what the patent is worth.

The fifth is enforceability. A patent is not self-enforcing, as the opening article noted, and enforcement is expensive, the subject of a later article on enforcement. A holder who cannot afford to litigate holds a right that a well-funded competitor can infringe with little fear.

The sixth is commercial relevance and timing. The patent must cover something the market actually values, and its remaining term must overlap the period in which that value exists. In a fast-moving field a patent can outlive the relevance of the thing it protects, granting exclusivity over a product the market has already left behind.

The Expected Value of a Patent Moat

These conditions can be combined into a simple model that makes their interaction explicit. The model is a tool for thought rather than a valuation method to defend in court, and its purpose is to show how a patent strong on one axis and weak on another ends up shallow.

Let $V$ be the present value of the exclusivity, the extra profit a firm would earn as the only seller in the covered market over the patent’s enforceable life. Let $p_v$ be the probability the patent survives a validity challenge. Let $p_d$ be the probability that infringement, when it occurs, is detected. Let $p_e$ be the probability of successful enforcement once infringement is detected. Let $C$ be the total cost to obtain, maintain, and enforce the patent. The expected value of the patent as a moat is then

\[E = p_v \, p_d \, p_e \, V - C.\]

The shape of this expression is the whole lesson. The three probabilities are each less than one, so they multiply downward. A patent that is likely valid, likely detectable, and likely enforceable still keeps only a fraction of $V$, and a patent weak on any single factor keeps almost none of it.

Consider a base case, with all values in millions of dollars. A patent protects exclusivity worth ten, it has a sixty percent chance of surviving a validity challenge, infringement is detectable seventy percent of the time, enforcement succeeds half the time it is attempted, and the lifetime cost is half a million. Then

\[E = (0.6)(0.7)(0.5)(10) - 0.5 = 2.1 - 0.5 = 1.6.\]

The headline figure was ten million. The expected moat value is 1.6 million, because the probabilities took nearly eighty percent of it before any money changed hands.

Now make the method hard to detect, as a process used inside a competitor’s own factory often is, so that $p_d$ falls to two tenths. Then

\[E = (0.6)(0.2)(0.5)(10) - 0.5 = 0.6 - 0.5 = 0.1.\]

The same patent, protecting the same market, is now barely worth its cost, because infringement that cannot be seen cannot be stopped.

Detectability is not the only factor that can collapse the value. The protected value $V$ is itself capped by the cheapest way a competitor can avoid the patent. A rational competitor will not pay more to be excluded than it costs to design around the claims. If that design-around costs $A$, the value the patent can actually protect is

\[V_{\text{eff}} = \min(V_{\text{market}}, A),\]

the lesser of the market value and the competitor’s avoidance cost. Suppose the market value is still ten million but a competent rival can design around the claims for one million. Then $V_{\text{eff}}$ is one million, and with the base-case probabilities

\[E = (0.6)(0.7)(0.5)(1) - 0.5 = 0.21 - 0.5 = -0.29.\]

The patent now costs more than it protects. It is not a shallow moat. It is a hole in the ground the company paid to dig.

The model is deliberately crude. Every input is an estimate, and reasonable people will disagree on each. Its value is not the precise number it produces but the structure it reveals, that the worth of a patent as a moat is a product of several fractions, and that one small fraction drags the whole product toward zero.

The Pharmaceutical Case, Where the Moat Is Real

Patents are a strong moat in pharmaceuticals, and the reason is that every condition holds at once. A small-molecule drug is often protected by a patent on the molecule itself. The claim is hard to design around, because the molecule is the product, and a competitor that changes it no longer has the same approved drug. Infringement is obvious, because the infringing pill contains the patented compound, which a laboratory can confirm. The value is enormous, because an approved drug can be a market of billions. And the long, expensive approval process aligns the patent term with the period of commercial value. Validity, scope, detectability, difficulty of designing around, enforceability, and relevance are all high together. This is the rare case in which a single patent is a deep and durable moat, and it is why the pharmaceutical industry organizes its entire strategy around patents.

The Software Case, Where the Moat Is Thin

Software is close to the opposite. A software patent often has narrow claims, both because broad ones founder on the eligibility limits the opening article described and because they founder on the prior art. The same function can usually be achieved by a different implementation the claims do not reach, so designing around is cheap. Infringement is hard to detect, because the accused code runs on a competitor’s own servers where no one outside can inspect it. And software moves quickly, so a patent that takes years to issue and enforce may protect a feature the market has already passed. Validity, scope, detectability, design-around difficulty, and timing are all weak together. This is why a software company’s real moat is rarely its patents, a point the closing section returns to.

The Portfolio and the Thicket

A single patent is seldom a moat, but a collection can be. A patent thicket is a dense set of overlapping patents that together cover a product from many angles, so that a competitor cannot design around one claim without running into another. The portfolio raises the avoidance cost $A$ far above what any single patent could, and it multiplies the litigation risk a competitor must weigh. The moat, in this case, is the thicket rather than the patent.

The portfolio effect can be stated precisely. A competitor entering the market is blocked if at least one patent in the portfolio proves valid, infringed, and enforceable against it. Let $q_i$ be the probability that patent $i$ alone would block, which is the single-patent product $p_v \, p_d \, p_e$ from the model above applied to that patent. If the patents hold or fail independently, the probability that the portfolio blocks is the complement of all of them failing,

\[P_{\text{block}} = 1 - \prod_{i=1}^{n} (1 - q_i).\]

This is why a thicket of individually weak patents can still be a strong barrier. Suppose a portfolio holds five patents, each with only a thirty percent chance of blocking on its own. No single one is frightening. Together,

\[P_{\text{block}} = 1 - (1 - 0.3)^5 = 1 - 0.7^5 \approx 0.83,\]

so a competitor faces better than an eighty percent chance of being stopped by at least one of them, and must pay to design around or invalidate every patent that blocks, a total cost far above any single $A$. The independence assumption flatters the result, since patents in a real thicket often share a specification and a prior-art exposure, so that one strong reference can fell several at once. The true blocking probability is therefore lower than the formula suggests, yet the qualitative lesson holds, that many shallow obstacles in series can form a deep barrier in aggregate.

The thicket has its own costs. Building and maintaining many patents is expensive, the strategy invites antitrust and policy scrutiny when it is used to block competition broadly, and numbers alone do not cure weakness, because patents that share a flaw can fall together to a single prior-art reference. A portfolio is a moat only when the patents in it are individually defensible and collectively hard to avoid.

When the Moat Is Not the Patent

For most companies, and nearly all startups, the durable advantage is not the patent at all. It is execution, distribution, accumulated data, network effects, brand, and the speed of a team that ships faster than rivals can copy. A patent may be a feature of such a moat, raising a competitor’s cost at one point, but it is rarely the moat itself.

This matters because founders consistently overestimate the protective value of patents and underestimate the cost of the conditions that make them work. A patent that is narrow, hard to detect, cheap to design around, and expensive to enforce is a poor substitute for a product customers prefer and a company that out-executes its rivals. The startup series develops this directly in its article on what it takes to succeed and where moats come from. The honest summary is that a patent can strengthen a moat but seldom creates one.

Epistemic State

The structural claims here are the six conditions and the multiplicative model that combines them. These are a framework for reasoning, testable in the sense that a competing analysis can dispute a condition or propose a different combination, and they should be read that way rather than as settled fact.

The numerical examples are illustrative. The probabilities and values are chosen to show the mechanism, not measured from a particular patent, and any real estimate of validity, detectability, or enforcement odds is uncertain and contested.

The contrast between pharmaceuticals and software is a generalization that holds broadly but admits exceptions in both directions, since some software patents are strong and some pharmaceutical patents are weak. Throughout, this is general information, United States centric, and not legal or financial advice.

Out of Scope

The formal methods of patent valuation, namely the income, market, and cost approaches, are a specialist discipline and are not treated here. The modeling of licensing and royalty income, the economics of standard-essential patents and their licensing commitments, and the operation of patent pools are likewise left aside. The mechanics of enforcement, which determine the probability $p_e$ in the model, belong to the enforcement article. The choice between patenting and secrecy, which determines whether a patent exists to be a moat at all, belongs to the article on patents and trade secrets.

Conclusion

A patent is a moat only when validity, claim scope, detectability, the difficulty of designing around, enforceability, and commercial relevance hold at the same time. Because these combine multiplicatively, a patent strong on five conditions and weak on the sixth is a shallow moat, and the headline value of a patent is almost never its real value. Pharmaceuticals are the case where every condition holds and the patent is a deep moat. Software is the case where the conditions fail together and the patent is thin. Most businesses sit between, and most of them hold patents that protect less than their owners believe. The next articles ask how a patent compares to keeping a secret, what patents mean for a founder, and what enforcing one actually requires.

References