A patent dispute rarely ends in a courtroom, and the routes that resolve it cheaply are the ones available early.
Cost rises sharply once litigation starts. Discovery and expert work consume most of a seven-figure budget, and neither happens before a complaint is filed.
Which means the sequencing matters more than the merits at the margin. A dispute resolved before suit costs a fraction of the same dispute resolved after claim construction, even where the outcome is identical.
The routes, by cost
| Route | Relative cost | Ends the dispute? |
|---|---|---|
| Design-around | Engineering time | Yes, unilaterally |
| Do nothing (letter not credible) | Nil | Sometimes |
| Negotiate a licence | Low | Yes |
| Mediation | Low | Usually |
| Arbitration | Moderate | Yes, if agreed |
| PTAB challenge | Moderate | Validity only |
| Declaratory judgment | High | Yes |
| Full litigation | Highest | Yes |
Design-around is the only route that does not require agreement. If the claim covers a detail you can change, changing it ends your exposure without negotiating with anyone.
It is also the most overlooked, because a demand letter frames the question as pay or fight rather than as an engineering problem.
First: establish what you are dealing with
| Step | Time | Why it matters |
|---|---|---|
| 1. Confirm the patent is in force | 5 min | It may have lapsed |
| 2. Verify current ownership | 10 min | The asserting party must own it |
| 3. Read the independent claims | 30 min | Letters overstate scope routinely |
| 4. Compare element by element | 2 hours | One missing element defeats it |
| 5. Read the prosecution history | 1 hour | Narrowing constrains construction |
| 6. Check remaining term | 10 min | Changes every calculation |
| 7. Preliminary prior art search | 2 hours | Determines whether IPR is viable |
Do all seven before choosing a route. They take a day, and any one of them can end the dispute outright.
Remaining term deserves particular attention. A patent with two years left cannot support a case that takes three to try. See how long will a patent last.
Who has leverage, and why
| Factor | Favours |
|---|---|
| Discovery burden | The non-practising party |
| Ability to design around | The accused party |
| Business disruption from an injunction | The patent holder |
| Remaining patent term | Whoever benefits from delay |
| Funding | Whoever has more |
| Strong prior art | The accused party |
| Marking compliance | The patent holder, if compliant |
Leverage is rarely symmetrical, and identifying where it sits determines the route rather than the merits alone.
Time favours the accused party when term is short. A patent with two years left loses value every month a dispute continues, which is worth knowing before agreeing to an extended negotiation.
Discovery burden favours whoever holds fewer documents, which is why assertion entities have an advantage independent of the strength of their patent.
Design-around
| Question | If yes |
|---|---|
| Does the claim recite a specific, changeable element? | Design-around is likely available |
| Can the change be made without losing function? | Cost is engineering time only |
| Is the product already shipped in volume? | Retooling and field units complicate it |
| Are there pending continuations? | New claims may follow you |
Check the file history before designing around. A claim narrowed during prosecution has estopped the doctrine of equivalents for the surrendered scope, which makes the design-around safer than it looks.
Pending continuations are the risk. A family with live applications can write new claims toward whatever you build, which is precisely what continuations exist for. See patent is pending.
Document the design-around effort. It weighs against willfulness and it is evidence of good faith if the dispute continues.
Negotiation
| Position that helps | Why |
|---|---|
| A credible non-infringement analysis | Changes what the claim is worth |
| Prior art | Changes what the patent is worth |
| A costed design-around | Caps what you will pay |
| Prosecution history showing narrowing | Limits construction and equivalents |
| Willingness to litigate | Changes the other side's arithmetic |
A costed design-around sets the ceiling. No licence is worth more than the cost of avoiding the patent, and saying so with a number moves negotiations.
Prior art moves price more than argument does. A specific reference that predates the priority date and reads on the claim changes the risk both sides are pricing.
Mediation and arbitration
| Mediation | Arbitration | |
|---|---|---|
| Binding | No | Yes |
| Requires prior agreement | No | Usually yes |
| Cost | Low | Moderate |
| Speed | Fast | Faster than court |
| Private | Yes | Yes |
| Appeal | — | Very limited |
| Binds third parties | No | No |
Mediation works best once the technical position is understood by both sides. Mediating before either party has done the element-by-element analysis produces a negotiation about feelings rather than facts.
Arbitration usually requires an existing agreement — a licence, supply contract or joint development agreement containing a dispute clause. It is rarely available in a dispute between strangers.
The PTAB route
| District court | Inter partes review | |
|---|---|---|
| Decides | Infringement and validity | Validity only |
| Invalidity standard | Clear and convincing | Preponderance |
| Grounds | Any | §102/§103, documents only |
| Cost | Seven figures | Substantially less |
| Timeline | 2–4 years | ~18 months |
| Deadline | — | 1 year from service |
| Estoppel | — | Grounds raised or reasonably could have been |
The standard difference is the point. Prior art that would fail against a presumption of validity in court can succeed on a preponderance before the Board.
A stay changes the economics entirely. District courts frequently stay litigation pending IPR, which pauses spending and shifts leverage.
Institution is discretionary and volatile — roughly 65% in October 2024, falling to around 37% by February 2026. Quote it with the date. See inter partes review.
Declaratory judgment
| Who brings it | The accused party |
| Requires | A sufficiently concrete controversy |
| Advantage | Choose the forum and the timing |
| Cost | High — it is litigation |
| When it fits | The uncertainty itself is damaging |
It converts a letter into a lawsuit, which is sometimes exactly what is needed. A company unable to raise money, close a sale or launch a product because of an unresolved assertion may prefer a decision to a cloud.
Venue choice is the main benefit. Bringing the action first means it is heard where you chose rather than where the patent holder would have filed.
Willfulness while a dispute is open
| Conduct after notice | Effect |
|---|---|
| Continuing with no reasoned basis | Supports enhanced damages |
| Obtaining a written opinion | Principal defence |
| Genuine design-around attempt | Weighs against |
| Ignoring the letter entirely | Worst position |
| Reasonable belief of invalidity | Weighs against |
Ignoring a letter is the worst of both worlds. Knowledge is established and there is no reasoned basis for continuing, which is close to the definition of what enhanced damages target.
An opinion must predate the continued conduct to do its job. One obtained after suit addresses strategy, not the state of mind that mattered. See freedom to operate.
Worked example: four routes priced
A demand letter asserting one patent against one product line.
| Route | Cost | Time | Outcome |
|---|---|---|---|
| Design-around | $60,000 engineering | 4 months | Exposure ends permanently |
| Licence at the demanded rate | $400,000 over 3 years | Immediate | Exposure ends, ongoing cost |
| IPR | $300,000 | 18 months | Claims cancelled — or not |
| Litigate to judgment | $1,800,000 | 3 years | Uncertain |
What the analysis found
| Finding | Effect |
|---|---|
| Patent in force, 6 years remaining | Long enough to matter |
| Three of four elements clearly present | Real exposure |
| Fourth element is a specific geometry | Changeable |
| Prosecution history narrowed to that geometry | Estoppel — equivalents barred |
| Prior art found, moderate strength | IPR viable but not certain |
The estoppel finding made the design-around safe. Because the applicant narrowed the claim to that geometry to overcome prior art, changing it cannot be captured by the doctrine of equivalents.
Design-around at $60,000 against a $400,000 licence is a straightforward decision once the estoppel is confirmed.
The IPR was prepared but not filed. Holding a credible validity challenge in reserve changed the tone of the correspondence considerably.
When litigation is the right answer
| Condition | Required |
|---|---|
| Claims clearly read on the product | Yes |
| Design-around impractical | Yes |
| Damages justify the cost | Yes |
| Sufficient remaining term | Yes |
| Marking compliance, if asserting | Yes |
| Funding to see it through | Yes |
All six, not most of them. A well-founded case can fail commercially simply by outlasting the party bringing it, which is why funding belongs on the list alongside the merits.
Check marking before asserting. An unmarked patent holder recovers only from actual notice, which can remove most of the damages. See patent marking.
Settlement terms worth negotiating
Price is rarely the only term that matters.
| Term | Why it matters |
|---|---|
| Scope of the licence | Which products, which patents |
| Future products | Whether new versions are covered |
| Affiliates and customers | Whether downstream parties are released |
| Release for past infringement | Otherwise past exposure remains |
| Patents covered | The whole family, or one patent |
| Term | Life of patent, or shorter |
| Confidentiality | Whether the deal becomes a comparable |
A release for past infringement is essential and sometimes omitted. A licence going forward leaves the past claim alive unless it is expressly released.
Customer releases matter for suppliers. Settling your own exposure while leaving customers open invites the same claim to arrive through them.
Confidentiality cuts both ways. A disclosed settlement becomes a comparable licence in every future negotiation involving that patent.
Patent dispute: the checklist
- Confirm the patent is in force before doing anything else.
- Read the claims yourself and compare element by element.
- Check the prosecution history for narrowing that constrains construction.
- Calculate remaining term. It caps what any route is worth.
- Price the design-around first. It sets the ceiling on every other option.
- Search for prior art early, since it determines whether IPR is available.
- Diarise the one-year IPR bar from service of any complaint.
- Do not ignore a credible letter. Knowledge without a reasoned basis is the worst position.
- Get a written opinion if you continue with knowledge of the patent.
- Reach litigation last, and only when all six conditions above hold together.