Patent Damages: Reasonable Royalties, Lost Profits, Apportionment, and Enhancement

Summary. How much a patent is worth, decided by a jury, disciplined by doctrine.

The statute is one sentence, and it has caused thirty years of argument

35 U.S.C. § 284 provides that upon finding infringement, the court shall award "damages adequate to compensate for the infringement, but in no event less than a reasonable royalty for the use made of the invention by the infringer, together with interest and costs as fixed by the court."

That is the whole of it. Everything else — the hypothetical negotiation, the fifteen Georgia-Pacific factors, apportionment, the smallest salable unit, the entire market value rule — is judge-made architecture built on top of two ideas embedded in that sentence: compensation and a floor.

The compensation principle means damages restore the patentee to the position it would have occupied but for the infringement. The floor means that even a patentee who lost nothing recoverable gets something, because the infringer used property it did not own.

Those two ideas generate the two paths every patent damages case runs on:

Lost profits — what the patentee would have earned had the infringer not been in the market. Available only when the patentee can prove it would have made the sales.
Reasonable royalty — what the parties would have agreed to in a hypothetical license negotiation. Always available, and by far the more common.

A patentee may recover lost profits on the sales it can prove it would have made and a reasonable royalty on the rest. Mixing the two is standard practice.

Lost profits: the Panduit framework

Lost profits are the larger number when they are available, because a patentee's margin usually exceeds any plausible royalty rate. They are also the harder proof.

The dominant test comes from Panduit Corp. v. Stahlin Bros. Fibre Works, Inc., 575 F.2d 1152 (6th Cir. 1978), which the Federal Circuit adopted. A patentee must show:

Demand for the patented product. Usually easy — sales by either party demonstrate demand.
Absence of acceptable non-infringing substitutes. The hard element. If a customer denied the infringing product would have bought something else that does not infringe, the patentee did not lose that sale.
Manufacturing and marketing capability to have made the sales.
The amount of profit the patentee would have made.

Element two is where lost profits cases die. Defendants build the record early: alternative designs, competitor products, the customer who testifies she would have bought the cheaper option anyway. A "substitute" need not be as good — it must merely be acceptable to the purchasers who bought the infringing product. Courts consider whether the substitute has the advantages of the patented features, whether purchasers specifically sought those features, and whether the substitute was actually available during the damages period. A design-around that the defendant could have built but did not is generally not an available substitute, though it may bear on the royalty analysis.

Several refinements matter in practice:

Market share apportionment. Where the market has multiple competitors, a patentee may recover lost profits on the share of infringing sales it would have captured — the State Industries v. Mor-Flo approach — rather than all-or-nothing.

Price erosion. A patentee who lowered prices to compete with an infringer may recover the difference, but must account for the volume it would have lost at the higher price. Economists call this the elasticity problem, and it defeats many price erosion claims.

Convoyed sales. Profits on unpatented items sold with the patented product are recoverable only where the items are functionally related, not merely sold together for convenience. The classic contrast: a patented printer and its required proprietary cartridge (recoverable) versus a patented printer and a desk sold in the same transaction (not).

Lost profits require the patentee to be in the market. A non-practicing entity that licenses rather than sells cannot show lost sales, which is why NPE cases are royalty cases.

Reasonable royalty: the hypothetical negotiation

When lost profits are unavailable or incomplete, the measure is a reasonable royalty. The dominant construct is a hypothetical negotiation: a fictional bargain between a willing licensor and a willing licensee, conducted at the time infringement began, assuming both parties knew the patent was valid and infringed.

That last assumption is deliberately counterfactual and does real work. In the real world the defendant would have argued invalidity and non-infringement, driving the price down. The hypothetical negotiation removes those arguments, which is why hypothetical royalties exceed real-world ones.

The framework for conducting the negotiation comes from Georgia-Pacific Corp. v. United States Plywood Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970), which listed fifteen factors. They are not a formula, and courts have grown impatient with experts who march through all fifteen mechanically. The ones that decide cases:

Factor 1 — Established royalties the patentee has received. The single most persuasive evidence, if genuinely comparable.
Factor 2 — Rates paid by the licensee for comparable patents.
Factor 4 — The licensor's policy of maintaining a monopoly by not licensing. A patentee that never licenses argues for a higher rate.
Factor 8 — Established profitability and commercial success of the patented product.
Factor 9 — Advantages over old modes. What the invention adds over the prior art — the heart of apportionment.
Factor 11 — Extent of the infringer's use.
Factor 13 — The portion of realizable profit attributable to the invention as distinguished from other features, manufacturing process, business risks, and improvements added by the infringer.
Factor 15 — The amount a willing licensor and willing licensee would have agreed upon.

Factor 13 is doing the most work in modern practice, because it is the statutory root of apportionment.

Apportionment: the doctrine that reshaped royalty practice

Here is the problem apportionment solves. A patent covers one feature of a smartphone. The phone sells for $900. If the royalty base is the phone and the rate is 2%, the award is $18 per unit for a feature that may be worth pennies. Multiply by fifty million units and a modest invention produces a catastrophic verdict.

The Federal Circuit's answer, developed across a line of cases including Uniloc USA v. Microsoft (2011), LaserDynamics v. Quanta (2012), VirnetX v. Cisco (2014), and Ericsson v. D-Link (2014), is that damages must be apportioned to the incremental value contributed by the patented feature.

Two mechanisms accomplish this:

The smallest salable patent-practicing unit. Rather than using the whole product as the base, the analysis should begin with the smallest component that practices the invention — the chip rather than the phone, the module rather than the machine. Even then, if that unit contains unpatented features, further apportionment is required.

The entire market value rule. Using the whole product as the base is permissible only where the patented feature drives customer demand for the entire product. This is a demanding standard. Evidence that the feature is valuable, or that customers like it, is not enough; the patentee must show it is the basis for the purchasing decision.

Two practical consequences follow.

First, the base matters more than the rate. Juries anchor on the base. A patentee who gets the entire product admitted as the base has effectively won the damages case even if the rate is small — which is exactly why the Federal Circuit polices the base so carefully and why defendants file motions in limine to exclude the total product revenue from the courtroom.

Second, comparable licenses have to be genuinely comparable. Experts routinely offer real-world licenses as evidence of the royalty. Courts require the expert to account for differences in technology, scope, timing, and structure. A portfolio license covering 3,000 patents is not evidence of the value of one, absent an explained methodology for extracting the single patent's contribution. A settlement license entered to avoid litigation costs may be excluded entirely or admitted with heavy caveats.

The Daubert gate

Damages is the phase of patent litigation where expert exclusion is most consequential, because a patentee whose damages expert is excluded may recover nothing at all. Federal Rule of Evidence 702, as amended in 2023 to emphasize that the proponent must establish admissibility by a preponderance and that the expert's opinion must reflect a reliable application of the methodology, governs.

Recurring grounds for exclusion:

A royalty rate with no tie to the evidence. The infamous "25% rule of thumb" was rejected in Uniloc, and any rate derived from a general heuristic rather than case-specific evidence is vulnerable.
Comparable licenses without a comparability analysis. The expert must explain why the license informs the hypothetical negotiation, not merely that it exists.
A base that violates apportionment. Using total product revenue without satisfying the entire market value rule.
A survey or conjoint analysis with methodological defects. These have become common tools for isolating feature value and are frequently attacked.
Failure to apportion within the smallest salable unit.
A hypothetical negotiation conducted at the wrong date.

Because exclusion can be case-ending, damages experts should be stress-tested internally long before reports are served, and the Daubert record should be built during the expert's deposition rather than assembled afterward from the report alone.

The limits that operate independent of the merits

Three doctrines reduce or eliminate damages regardless of how strong the infringement case is.

Marking

35 U.S.C. § 287 provides that a patentee who makes or sells a patented article must mark it with the patent number — or, since 2011, with a web address providing that information (virtual marking). A patentee who fails to mark recovers damages only from the date of actual notice, which usually means the filing of the complaint or a specific pre-suit notice letter identifying the patent and the accused product.

The consequences are severe and routinely underappreciated:

Marking is required of licensees too, and the patentee bears responsibility for their compliance. A single unmarked licensee product can cut off pre-suit damages entirely.
The requirement applies only to patents with apparatus claims. A patentee asserting only method claims need not mark, which is one reason asserted claim selection has damages consequences.
Once raised by the defendant, the patentee bears the burden of proving marking or actual notice.
Actual notice must be specific. A letter offering a license to a portfolio without identifying the patent and the accused product does not start the clock.

The six-year window

35 U.S.C. § 286 bars recovery for infringement occurring more than six years before the complaint. This is not a statute of limitations on the claim — a patentee can sue over old infringement — but a limit on the damages period.

Note the interaction with laches: in SCA Hygiene Products Aktiebolag v. First Quality Baby Products, LLC, 580 U.S. 328 (2017), the Supreme Court held that laches cannot bar damages within the § 286 window, following the copyright analysis of Petrella v. Metro-Goldwyn-Mayer, Inc., 572 U.S. 663 (2014). Delay may still support equitable estoppel, which is a different doctrine with a different showing.

Extraterritoriality

Damages generally cover only domestic infringement under 35 U.S.C. § 271. But WesternGeco LLC v. ION Geophysical Corp., 585 U.S. 407 (2018) held that a patentee who proves domestic infringement under § 271(f) — supplying components from the United States for combination abroad — may recover foreign lost profits flowing from that domestic act. The decision rests on the focus of the statute rather than on any general rule permitting foreign damages, and its reach beyond § 271(f) remains contested.

Willfulness and enhancement

35 U.S.C. § 284 permits the court to increase damages "up to three times the amount found or assessed." The standard governing that discretion was rewritten in Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U.S. 93 (2016).

Before Halo, the Federal Circuit's Seagate test required objective recklessness proven by clear and convincing evidence — a standard so demanding that a defendant could avoid enhancement by developing a reasonable defense during litigation, even one it had never considered before infringing. The Supreme Court rejected that:

"The subjective willfulness of a patent infringer, intentional or knowing, may warrant enhanced damages, without regard to whether his infringement was objectively reckless."

The current framework:

Willfulness is a jury question about the infringer's state of mind at the time of the challenged conduct. Knowledge of the patent is necessary but not sufficient; the conduct must be egregious — "willful, wanton, malicious, bad-faith, deliberate, consciously wrongful, flagrant, or . . . characteristic of a pirate."
The standard of proof is preponderance, not clear and convincing.
Enhancement is discretionary even after a willfulness finding. District courts apply the Read factors: deliberate copying, whether the infringer investigated and formed a good-faith belief, litigation conduct, size and financial condition, closeness of the case, duration of misconduct, remedial action, motivation to harm, and concealment.
A defense developed only in litigation does not defeat willfulness, though it bears on whether enhancement is appropriate.

Opinions of counsel occupy an interesting position. 35 U.S.C. § 298 provides that failure to obtain or present an opinion may not be used to prove willfulness or intent to induce. But an opinion the defendant did obtain and relies on is powerful evidence of good faith — at the cost of waiving privilege over the subject matter. That trade-off is one of the more consequential decisions in a patent defense.

For induced infringement under § 271(b), the knowledge requirement comes from Global-Tech Appliances, Inc. v. SEB S.A., 563 U.S. 754 (2011) (knowledge of the patent and that the induced acts infringe, satisfiable by willful blindness) and Commil USA, LLC v. Cisco Systems, Inc., 575 U.S. 632 (2015) (a good-faith belief in invalidity is not a defense to inducement).

Attorney fees

35 U.S.C. § 285 permits fee awards to the prevailing party "in exceptional cases." Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014) rejected the Federal Circuit's rigid prior test and defined an exceptional case simply as:

"one that stands out from others with respect to the substantive strength of a party's litigating position (considering both the governing law and the facts of the case) or the unreasonable manner in which the case was litigated."

District courts decide this case-by-case, by a preponderance, considering the totality of circumstances, and the companion decision in Highmark Inc. v. Allcare Health Management System established that the determination is reviewed for abuse of discretion. The practical effect has been a meaningful increase in fee awards, particularly against plaintiffs asserting weak patents against many defendants and against defendants who litigate abusively. Fees are available to prevailing defendants as well as patentees — this is not a one-way statute.

Related tools: 28 U.S.C. § 1927 permits fee-shifting against counsel who multiply proceedings unreasonably and vexatiously, and courts retain inherent power to sanction bad-faith conduct.

Interest and costs

Prejudgment interest is the rule, not the exception. General Motors Corp. v. Devex Corp., 461 U.S. 648 (1983) held that prejudgment interest should ordinarily be awarded absent some justification for withholding it, because it is necessary to make the patentee whole. Disputes concern the rate (prime, treasury, or the patentee's borrowing cost), whether it compounds, and whether the patentee's delay justifies reduction.

Post-judgment interest runs automatically under 28 U.S.C. § 1961 at the statutory rate.

Costs are awarded to the prevailing party under Federal Rule of Civil Procedure 54(d), though the categories are narrow.

Injunctions and the shadow they cast on damages

Before 2006, a patentee that won generally got an injunction, and that prospect drove settlement values. eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006) ended the presumption, holding that the traditional four-factor equitable test applies to patent injunctions under 35 U.S.C. § 283: irreparable harm, inadequacy of legal remedies, balance of hardships, and the public interest.

The practical consequence was a bifurcation of the patent world. Competitors who practice their patents still frequently obtain injunctions, because lost market share and price erosion are hard to compensate with money. Non-practicing entities rarely do, because a patentee whose only interest is licensing revenue can be made whole with damages.

Where an injunction is denied, courts may award an ongoing royalty for future infringement. That rate is typically higher than the pre-verdict royalty, because the hypothetical negotiation now occurs against a backdrop of adjudicated validity and infringement.

What a damages case is worth before it is filed

Sophisticated parties value patent cases before filing, and the analysis has a recognizable shape.

Identify the accused revenue. Which products, over what period, at what volume. Then reduce it by the § 286 six-year window and by any marking failure.

Identify the apportionment story. What is the smallest salable unit? Is there any credible entire-market-value theory? What is the technical contribution over the closest prior art — the Georgia-Pacific factor 9 and 13 analysis that will drive the rate?

Identify the comparables. Existing licenses, industry rates, and their genuine comparability. This is usually the most important single input, and its absence is the most common reason a damages case is weak.

Assess lost profits availability. Is the patentee in the market? Are there acceptable non-infringing substitutes? If substitutes exist, the case is a royalty case and its value drops substantially.

Assess willfulness. Was there pre-suit notice? Copying? An opinion? A trebled verdict changes everything about settlement posture, and so does its absence.

Assess the fee exposure in both directions. Under Octane Fitness, a weak case can cost the plaintiff the defendant's fees.

Discount for the gauntlet. A damages number is worth its face value multiplied by the probability of surviving claim construction, summary judgment, an IPR under 35 U.S.C. § 311, trial, and appeal. Each stage has meaningful failure rates. A $50 million damages model with a 15% path-to-judgment probability is a $7.5 million case before litigation costs.

A worked example: two damages theories, one patent

Meridian Filtration, Inc. holds a patent on a backwash valve assembly used in industrial water treatment skids. Ardent Process Systems sells competing skids that, the jury finds, infringe.

The complete skid sells for $340,000. The valve assembly inside it sells, when sold separately, for $6,200. Meridian sells skids too — it is a competitor, not a licensing entity — and it holds roughly 40% of the domestic market. Two other companies, Hollis Fluidics and Brightwater Systems, hold most of the rest, and both sell skids using a different valve design that does not infringe.

Ardent sold 480 infringing skids over the damages period.

Meridian's lost profits theory. Meridian's expert, Dr. Ines Barros, applies Panduit. Demand is established: 480 units sold. Capability is established: Meridian's plant ran at 62% capacity. Profit is established: $71,000 incremental margin per skid.

Element two is the fight. Ardent argues that Hollis and Brightwater sold acceptable non-infringing skids throughout the period, so most of Ardent's customers would simply have bought from them. Meridian responds that its patented backwash design reduces downtime by 18%, that three of Ardent's largest customers specified that capability in their procurement documents, and that Hollis and Brightwater skids do not provide it.

How the court is likely to handle it. Neither extreme survives. The procurement documents support lost profits for the customers who specified the capability — perhaps 90 units. For the rest, Ardent's substitutes argument prevails, but Meridian invokes market share apportionment: of the remaining 390 units, Meridian would have captured roughly its 40% share, or 156 units. Total lost profits: about 246 units × $71,000 = $17.5 million, subject to the jury's view of the market share evidence.

The royalty theory for the remainder. For the other 234 units, Meridian seeks a reasonable royalty. Here apportionment bites hard.

Dr. Barros initially proposed a 3% royalty on the $340,000 skid price — $10,200 per unit, $2.4 million total. Ardent moved to exclude under Federal Rule of Evidence 702, arguing the entire market value rule was not satisfied: the backwash valve is one of roughly forty subsystems in a skid, and no evidence showed it drove purchasing decisions for the skid as a whole.

The court agreed and excluded the opinion. Dr. Barros's revised analysis begins with the smallest salable patent-practicing unit — the $6,200 valve assembly — and then apportions within it, because the assembly includes conventional actuators and housings that are not part of the invention. She estimates the patented backwash geometry accounts for 45% of the assembly's value, producing an apportioned base of $2,790, and applies a 12% rate derived from two prior Meridian licenses to smaller manufacturers. Result: about $335 per unit, or $78,000 across 234 units.

The lesson. The apportionment ruling reduced the royalty component by 97%, and Meridian's recovery now rests almost entirely on lost profits. This is the ordinary shape of modern patent damages: the lost profits case, when available, dwarfs the royalty case, and the royalty case is disciplined until it is small.

Willfulness. Meridian sent Ardent a notice letter three years before filing, identifying the patent and the accused skid line by model number. Ardent's engineering director testified that he read it, forwarded it to the general counsel, and heard nothing further. No opinion of counsel was obtained; none was required, and under 35 U.S.C. § 298 the absence cannot be used to prove willfulness. But the jury also heard that Ardent's product manager had written, two months after the letter, "we should keep selling until they actually sue." The jury finds willfulness. The court, applying the Read factors, enhances by 1.5× rather than trebling — the case was close on infringement, and Ardent's litigation conduct was unremarkable.

Final tally. $17.5 million lost profits + $78,000 royalty = $17.58 million, enhanced to about $26.4 million, plus prejudgment interest under General Motors Corp. v. Devex Corp., 461 U.S. 648 (1983) and post-judgment interest under 28 U.S.C. § 1961.

And the injunction. Meridian, a direct competitor losing market share, satisfies the eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006) factors and obtains a permanent injunction — which, in commercial terms, is worth more than the verdict.

Damages discovery: what to ask for and what to produce

Damages discovery is its own project, and it should begin early because the data is often in systems nobody has queried before.

What a patentee needs from the infringer:

Unit sales and revenue by product, by month, for the full damages period
Cost data sufficient to compute margins, including standard cost build-ups
Pricing histories, discount schedules, and rebate programs
Product marketing materials and technical documentation showing how the accused feature is described to customers
Customer communications referencing the patented capability
All patent licenses the infringer has taken or granted in the field
Design-around analyses, including projects considered and abandoned
Documents concerning knowledge of the patent — the willfulness file
Market studies, win/loss analyses, and competitive intelligence

What an accused infringer needs from the patentee:

All licenses to the patent or related patents, and all license negotiations
The patentee's own sales, capacity, and margin data (to test the Panduit elements)
Documents about non-infringing alternatives and competitor products
Valuations of the patent for any purpose — acquisition, financing, accounting, tax
Litigation and settlement history involving the patent
Marking compliance evidence for the patentee and every licensee
Communications establishing when actual notice was given

Two structural points. First, license agreements are often subject to third-party confidentiality obligations, and producing them requires notice to the counterparty and usually a protective order — build the schedule accordingly. Second, financial data produced natively in a usable format saves months; produced as PDFs it costs a fortune to rebuild. Negotiate format in the ESI protocol, not later.

Why juries and judges see damages so differently

Patent damages is one of the few areas of law where the fact-finder and the reviewing court operate on visibly different logics, and understanding the gap explains a great deal of what happens after a verdict.

Juries reason from anchors. A number placed in front of a jury exerts gravitational pull regardless of the instructions. This is why the fight over the royalty base is really a fight over what number the jury hears. A patentee who gets $340,000 per skid into evidence has changed the jury's sense of scale even if the court instructs that the base is $6,200.

Juries reason from fairness. Evidence that the defendant copied, ignored a warning letter, or profited enormously moves damages numbers even when it is legally relevant only to willfulness. Bifurcation exists partly to manage this.

Appellate courts reason from methodology. The Federal Circuit does not reweigh evidence; it asks whether the expert applied a reliable method to sufficient facts, whether the base was proper, and whether substantial evidence supports the verdict. A verdict supported by a methodologically sound expert survives even when it is large. A verdict resting on an expert whose apportionment was hand-waved does not, however modest the number.

The practical instruction for both sides is the same: build the damages case for the appellate court, and present it to the jury. A damages theory that is intuitively appealing but methodologically thin is a verdict you will lose eighteen months later, after paying for the appeal.

Special damages contexts

Design patents. 35 U.S.C. § 289 provides a remedy unique in American law: the infringer's total profit on the article of manufacture to which the design is applied, with no apportionment. In Samsung Electronics Co. v. Apple Inc., 580 U.S. 53 (2016), the Supreme Court held that the "article of manufacture" may be a component rather than the entire end product, which returned some discipline to the analysis — but § 289 remains the most patentee-favorable damages provision in the statute, and it explains a great deal of design patent enforcement strategy.

Standard-essential patents. Where a patent has been declared essential to a standard and the holder committed to license on fair, reasonable, and non-discriminatory terms, the royalty analysis changes. Courts adjust the Georgia-Pacific factors to exclude value attributable to the standard's adoption rather than to the technology, and to account for royalty stacking across the many patents essential to the same standard. The hypothetical negotiation is conducted against the backdrop of the FRAND commitment, not a free market.

Pharmaceutical and ANDA cases. A Hatch-Waxman case brought under § 271(e)(2) often involves no marketed infringing product and therefore no past damages at all — the remedy is an order delaying approval. Damages arise only where the generic launched at risk, in which case lost profits on brand sales are enormous and the Panduit substitutes analysis is dominated by the regulatory framework rather than by ordinary market competition.

Government infringement. Suits against the United States proceed under 28 U.S.C. § 1498 in the Court of Federal Claims, where the remedy is "reasonable and entire compensation" — a reasonable royalty — and injunctive relief is unavailable. Contractors performing for the government with authorization are covered too, which occasionally moves an entire dispute out of district court.

What defense counsel should do first

If you represent an accused infringer, five moves early in the case do more for the damages exposure than anything done later:

Test marking compliance immediately. Identify every product the patentee and its licensees sell that practices an apparatus claim, and determine whether it was marked. A marking failure can eliminate years of damages, and the patentee bears the burden once you raise it.
Fix the damages start date. Calculate the 35 U.S.C. § 286 six-year boundary and determine the date of actual notice with precision. Notice letters that fail to identify the patent and the accused product do not count.
Build the non-infringing alternatives record. Identify competitor products, design-arounds, and prior art embodiments available during the damages period. This is the Panduit element that defeats lost profits, and the evidence lives with engineers who will leave the company.
Collect every license the patentee has ever granted. Comparable licenses are the most persuasive royalty evidence there is, and a low prior license is worth more than a good expert.
Audit the willfulness file. Find every internal document mentioning the patent or the patentee. Decide early whether to rely on an opinion of counsel, understanding that doing so waives privilege over the subject matter, and understanding that under 35 U.S.C. § 298 the absence of an opinion cannot be used against you.

Bifurcation and trial structure

How a damages case is tried changes what it is worth, and the structural decisions are made months before trial.

Bifurcating damages from liability. Some courts try infringement and validity first and reserve damages for a second phase. Defendants generally favor this, because it keeps the patentee's large revenue numbers away from the liability jury and because most cases settle after a liability verdict without a damages trial ever occurring. Patentees generally resist, because damages evidence — the copying, the notice letter, the profits — colors the liability narrative favorably.

Bifurcating willfulness. More common and less contested. Evidence that the defendant knew about the patent and kept selling is prejudicial on infringement in a way that is hard to cure with an instruction. Many courts try willfulness with damages or reserve it entirely.

Trying enhancement to the court. After Halo, willfulness is a jury question but enhancement is not. The Read factor analysis happens on post-trial briefing before the judge, which means the record supporting enhancement — litigation conduct, remedial steps, the closeness of the case — is partly built after the verdict.

Ongoing royalties. Where an injunction is denied, the court sets a rate for continuing infringement. This is decided post-trial, often on supplemental briefing and sometimes with a further hearing. Parties who have not thought about it before the verdict find themselves negotiating from a standing start.

Sealing the numbers. Damages evidence is the most commercially sensitive material in a patent case, and both sides usually want portions of the trial closed or the exhibits redacted. Courts are increasingly reluctant to seal, given the public right of access, and a party that has not built a specific, document-by-document justification will find its financial data read into an open record.

Frequently asked questions

Who bears the burden on marking? The patentee, once the defendant identifies products it contends were sold unmarked. The defendant's initial burden is one of production only — it must point to unmarked products — after which the patentee must prove compliance or actual notice. This allocation surprises patentees regularly, and it is why marking audits belong in the pre-filing workup rather than in the middle of expert discovery.

Does an ongoing royalty have to match the jury's rate? No, and it usually does not. The pre-verdict royalty reflects a hypothetical negotiation in which validity and infringement were uncertain. After a verdict, they are not, and courts routinely set ongoing rates above the jury's figure to reflect the changed bargaining position.

Can a patentee recover both lost profits and a reasonable royalty? Yes — lost profits on sales it proves it would have made, and a reasonable royalty on the remaining infringing units.

Does the royalty base have to be the smallest salable unit? Not always, but using anything larger requires satisfying the entire market value rule, and even the smallest salable unit often requires further apportionment if it contains unpatented features.

Is the 25% rule of thumb still usable? No. It was rejected as inadmissible in Uniloc because it bears no relation to the facts of the specific case.

How does virtual marking work? Mark the product with "Patent" or "Pat." and a web address that associates the product with the patent numbers. The page must be freely accessible and kept current — a stale page can defeat constructive notice.

Does an infringer's own profitability set a ceiling? No, but it is relevant. A royalty that exceeds the infringer's margin invites the argument that no rational licensee would have agreed to it in the hypothetical negotiation.

When is willfulness decided? Usually by the jury at trial, with enhancement decided afterward by the judge. Some courts bifurcate willfulness to avoid prejudicing the liability case.

Can a defendant recover fees? Yes. Section 285 is symmetrical, and prevailing defendants have recovered substantial fees in cases found exceptional under Octane Fitness.

What happens to damages if some asserted claims are invalidated in an IPR? Damages attributable to those claims disappear. If the surviving claims are narrower, the apportionment analysis usually changes too — which is why parallel proceedings affect damages models as much as liability.