Damages

Patent damages are what a court awards to compensate for infringement, and 35 U.S.C. § 284 sets a floor: damages adequate to compensate, but in no event less than a reasonable royalty for the use made of the invention by the infringer. That floor is the single most important sentence in patent remedies. Once infringement of a valid claim is established, the award cannot be zero, even if the patent owner sold nothing, lost nothing, and would never have licensed anyone.

A reasonable royalty is the price a willing licensor and a willing licensee would have agreed to in a negotiation that never happened, held just before infringement began. Everything else in this area is machinery built to make that fiction produce a number a jury can defend.

The two damages theories, and why the choice matters

American patent law offers a patent owner two compensatory routes, and they pay very differently.

Lost profits ask what the patent owner would have earned but for the infringement. This is the larger recovery and the harder proof. The standard framework comes from Panduit Corp. v. Stahlin Bros. Fibre Works, 575 F.2d 1152 (6th Cir. 1978), which asks for four showings: demand for the patented product, an absence of acceptable non-infringing substitutes, manufacturing and marketing capacity to have exploited that demand, and a profit amount that can actually be computed. The second factor does most of the killing. If a customer denied the infringing product would have bought some third party’s lawful alternative rather than the patent owner’s, the sale was never the patent owner’s to lose.

Because lost profits require the patent owner to have been in the market, entities that only license, including universities and non-practicing entities, cannot reach them at all. For them the reasonable royalty is not a fallback. It is the whole case.

The hypothetical negotiation and the Georgia-Pacific factors

The reasonable royalty runs through a legal fiction called the hypothetical negotiation: a bargain reconstructed as of the date infringement first began, between a willing licensor and a willing licensee, with both sides stipulating that the patent is valid and infringed. That last assumption is what makes it hypothetical. A real negotiation would have discounted for the chance the patent would fall.

The framework courts use to populate that negotiation comes from Georgia-Pacific Corp. v. United States Plywood Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970), which listed fifteen evidentiary factors. They are not elements and they are not weighted. They are a checklist of what the parties would have argued about, including established royalties the patent owner has actually charged, rates the licensee pays for comparable patents, whether the license would be exclusive, whether the two parties are competitors, the patent’s remaining term, the profitability of the patented feature, the extent of the infringer’s use, and expert opinion on what a licensee would have paid.

The fifteenth factor is the one that carries the analysis: the amount the hypothetical parties would have agreed on. The other fourteen exist to constrain it.

In practice, factors 1 and 2 dominate, because comparable real-world licenses are the most credible anchor available. That is also where cases are won and lost. A license the patent owner signed for an unrelated portfolio, or one signed under litigation pressure, is routinely excluded as not technically or economically comparable.

Apportionment: the royalty base problem

Suppose a patent covers one feature of a laptop’s optical drive. Is the royalty base the laptop, the drive, or the chip?

The default answer is the smallest salable patent-practicing unit, and the general principle behind it is apportionment: damages must be tied to the incremental value the patented feature contributes, not to the value of everything it happens to ship inside. LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012), applied this to a patent on disc-type detection and held that the royalty could not run against the price of the whole laptop.

The entire market value rule is the narrow exception that lets a patent owner use the full product as the base, and it is narrow on purpose. The patented feature must be the basis for customer demand, meaning it must be what actually drives the purchase. LaserDynamics was explicit that a feature being necessary, or even important, is not enough. It must be the motivating factor.

The reason courts police this so hard is not arithmetic. A royalty on a $1,000 laptop at 0.1 percent and a royalty on a $20 drive at 5 percent produce the same dollar. But juries anchor on big numbers, and the Federal Circuit has treated the mere disclosure of a large revenue figure as prejudicial. VirnetX, Inc. v. Cisco Systems, Inc., 767 F.3d 1308 (Fed. Cir. 2014), reinforced this, rejecting a damages theory that used the entire iOS device as a base and separately rejecting the mechanical use of the Nash bargaining solution as a substitute for evidence.

The same skepticism killed the old shortcut. Uniloc USA, Inc. v. Microsoft Corp., 632 F.3d 1292 (Fed. Cir. 2011), rejected the 25 percent rule of thumb, which had assumed a licensee would part with a quarter of its expected profit. A rule of thumb that ignores the specific patent, the specific parties, and the specific market is not evidence.

Enhanced damages and willfulness

Section 284 also lets a court increase the damages up to three times the amount found or assessed. This is punitive, it is decided by the judge rather than the jury, and it is not part of the compensatory calculation.

Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U.S. 93 (2016), rewrote this area. A unanimous Court threw out the rigid two-part Seagate test, which had required clear and convincing proof of objective recklessness, and held that § 284 commits enhancement to the district court’s discretion, reserved for egregious cases of misconduct beyond typical infringement. It also lowered the burden of proof to a preponderance of the evidence and rejected the defense trick Seagate had enabled: an infringer could no longer escape enhancement by manufacturing a reasonable invalidity argument at trial that it had never actually relied on when it infringed.

Halo made enhancement easier to reach but did not make it automatic. Willfulness is a predicate, not a trigger. Many juries find willfulness and many judges then decline to enhance at all.

Marking, notice, and the damages clock

Two provisions quietly cap what is recoverable no matter how strong the theory.

35 U.S.C. § 287 conditions damages on notice. A patent owner who makes or sells a patented article must mark it with the patent number, or with a free public web address listing the number under virtual marking. Fail to mark, and damages start only when the infringer received actual notice, which usually means the date of the demand letter or the complaint. Years of infringement can evaporate this way. Section 287 does not apply to method claims, which is one reason patent owners value them.

35 U.S.C. § 286 bars recovery for any infringement committed more than six years before the complaint was filed. It is a damages limitation rather than a statute of limitations, so it does not bar the suit, only the older money.

Both of these sit upstream of every calculation above. The finest Georgia-Pacific analysis in the world runs against zero units if the clock never started.

Frequently asked questions

What is a reasonable royalty in patent damages? A reasonable royalty is the statutory floor for patent damages under 35 U.S.C. § 284. It is the payment a willing licensor and a willing licensee would have agreed on in a hypothetical negotiation held just before infringement began, assuming both sides knew the patent was valid and infringed. Courts most often structure the analysis around the fifteen Georgia-Pacific factors. It is what a patent owner recovers when lost profits cannot be proven, and it can never be zero once infringement is established.

How is a reasonable royalty calculated? Usually as a royalty rate applied to a royalty base. The base is the revenue attributable to the infringing sales, and the rate is the percentage the hypothetical negotiation would have produced. Both must be tied to the patented feature rather than the whole product, a requirement called apportionment. Courts also accept lump-sum structures and comparable license evidence. Rules of thumb untethered to the facts, such as the old 25 percent rule, were rejected in Uniloc v. Microsoft (2011).

What is the difference between lost profits and a reasonable royalty? Lost profits compensate the patent owner for sales it would have made but for the infringement, and they require proof under the four Panduit factors: demand for the patented product, absence of acceptable non-infringing substitutes, capacity to meet the demand, and a computable profit amount. A reasonable royalty compensates for the use of the invention itself and requires no such proof. Lost profits usually pay more, so patent owners plead them first and fall back to a royalty.

Can patent damages be tripled? Yes. Section 284 lets a district court increase damages up to three times the amount found or assessed. After Halo Electronics v. Pulse Electronics (2016), enhancement is discretionary and reserved for egregious infringement behavior, proven by a preponderance of the evidence rather than the clear and convincing standard the Federal Circuit had previously imposed. Enhancement is not automatic on a willfulness finding.

Authorities and sources

• 35 U.S.C. § 284: Damages (Cornell LII)
• 35 U.S.C. § 285: Attorney fees (Cornell LII)
• 35 U.S.C. § 286: Time limitation on damages (Cornell LII)
• 35 U.S.C. § 287: Limitation on damages and other remedies; marking and notice (Cornell LII)
• 35 U.S.C. § 283: Injunction (Cornell LII)
• Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U.S. 93 (2016) (Cornell LII)
• eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006) (Cornell LII)
• WesternGeco LLC v. ION Geophysical Corp., 585 U.S. 407 (2018) (Cornell LII)
• LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012) (PDF)
• USPTO Report to Congress on Virtual Marking (2014) (PDF)
• USPTO Manual of Patent Examining Procedure

Going further: What Is My Patent Worth? A valuation walkthrough .

This page is general legal information, not legal advice, and it does not create an attorney-client relationship.