Patent troll is a label, not a legal category, and the label does a lot of work that the facts sometimes do not support.

The neutral term is non-practising entity. An owner that does not make or sell anything covered by its patents.

Which immediately catches universities, research institutes, failed companies and individual inventors — few of whom anyone calls trolls. The label describes conduct people object to rather than a structural fact about the owner.

What the objection actually targets is cost asymmetry. A demand priced below the cost of defending is designed to be paid whether or not it has merit, and that is the practice the term was coined for.

Who does not practise their patents

Owner type Practises? Called a troll?
Universities No Almost never
Research institutes No Rarely
Individual inventors Usually not Rarely
Failed companies' successors No Sometimes
Defensive aggregators No No
Assertion entities No Usually
Operating companies asserting non-core patents Partially Sometimes

The structural fact is identical across every row. What differs is the behaviour: whether the assertion is grounded in analysis, priced against merit, and directed at a genuine infringement.

Which is why the useful question about any letter is not who sent it, but whether the claims read on your product.

The business model

Step What happens
Acquire Patents bought cheaply, often from failed companies or lapsing owners
Identify targets Products that plausibly read on the claims
Demand Letters at a price below defence cost
Settle Most recipients pay
Litigate A minority, to sustain credibility

Acquisition is cheap because the sellers have no alternative. A company winding down, or an owner facing a maintenance fee they cannot justify, sells for a fraction of what an assertion campaign might yield.

Nearly three in five US utility patents are abandoned before term — see the patent survival curve — and that pool of unwanted patents is where assertion portfolios come from.

Pricing below defence cost is the mechanism. A demand at a level clearly below what litigation would cost makes settlement the rational choice independently of merit, which is precisely the objection.

Why defending costs so much

Burden Falls on
Source code production You
Sales and financial records You
Engineer depositions You
Document preservation You
Business disruption You
Producing a patent and an assignment Them

The asymmetry is structural, not tactical. An operating company holds everything discoverable; an assertion entity holds almost nothing.

Which is why settlement pressure exists regardless of the patent's strength, and why the practice attracts the criticism it does. See patent litigation.

Where assertion portfolios come from

Source Why patents become available
Failed companies Assets sold in wind-down
Owners facing maintenance fees The fee exceeds perceived value
Individual inventors No means to enforce alone
Corporate portfolio pruning Non-core patents released
Bankruptcy estates Liquidation
Universities Licensing mandate, no enforcement capacity

The maintenance fee window is the main supply mechanism. An owner facing $8,280 on a patent generating nothing will take a fraction of that from a buyer, because the alternative is receiving nothing at all.

Which means assertion entities buy at the exact moment owners are least able to negotiate. See patent portfolio pruning.

It also means some acquired patents are genuinely strong. A failed company's patents may cover technology the market later adopted, which is a real asset regardless of who ends up holding it.

Reading a demand letter

The letter What it signals
Identifies specific claims Someone did analysis
Includes a claim chart Serious
Names the accused feature precisely Serious
Refers vaguely to "your products" Screening exercise
Cites a patent without a number Not credible
Demands a response in days Pressure tactic
Offers a low flat licence to many parties Volume campaign

A letter with no claim chart has usually not been analysed against your product. That does not make it safe to ignore, but it changes what you are dealing with.

Volume campaigns are identifiable. A modest flat fee, a short deadline and generic accusation language indicate a letter sent to many recipients rather than a considered assertion against you.

The first day's work

Step Time Why
1. Confirm the patent is in force 5 min It may have lapsed
2. Check current ownership 10 min The sender must own it
3. Read the independent claims 30 min Letters overstate scope routinely
4. Compare element by element 2 hours The all-elements rule
5. Read the prosecution history 1 hour Narrowing limits construction
6. Check for pending continuations 15 min Wider exposure
7. Preliminary prior art search 2 hours IPR viability

Do all seven before responding. They take a day and they frequently reveal that the claims are narrower than asserted, that the patent lapsed, or that the sender does not own it.

Ownership is worth checking specifically. The front-page assignee reflects grant only, and assertion entities frequently hold patents through shell entities whose chain of title has gaps. See patent assignment database.

Response options, by cost

Option Relative cost When it fits
Ignore Nil Only for clearly non-credible letters, and rarely
Request a claim chart Nil Always reasonable; tests seriousness
Design around Engineering time The claim covers a changeable detail
Negotiate a licence Low The claims read and the price is sensible
IPR petition Moderate You have strong documentary prior art
Declaratory judgment High You need the uncertainty resolved
Full defence Highest Everything else failed

Requesting a claim chart costs nothing and is informative. A sender with analysis will provide one; a volume campaign frequently will not.

Design-around remains the cheapest genuine resolution. If the claim recites a detail you can change, changing it ends the exposure without paying anyone.

Volume campaigns versus targeted assertions

Volume campaign Targeted assertion
Recipients Many Few
Demand size Below defence cost Substantial
Claim chart Usually absent Present
Product identified Generically Specifically
Deadline Short Reasonable
Willingness to litigate Low Real
Right response Analyse, then respond briefly Full analysis and strategy

A volume campaign is priced to be paid rather than fought, and identifying one changes the response. The analysis is the same; the escalation path is not.

Targeted assertions deserve the full treatment. Claim chart, prior art search, prosecution history, design-around costing and an opinion.

Both start with the same first day of work, which is why doing that work before responding matters regardless of which you have received.

The IPR route

District court Inter partes review
Invalidity standard Clear and convincing Preponderance
Grounds Any §102/§103 on documents only
Cost Seven figures Substantially less
Timeline 2–4 years ~18 months
Filing deadline 1 year from service
Estoppel Grounds raised or reasonably could have been

The preponderance standard is the reason IPR matters. The same prior art that would fail against a presumption of validity in court can succeed before the Board.

The one-year bar is absolute and starts on service of a complaint. Missing it forfeits the route.

Institution is discretionary and has been volatile — roughly 65% in October 2024, falling to around 37% by February 2026. Quote that with its date. See inter partes review.

What changed after the AIA

Change Effect
Inter partes review created Cheaper validity challenges
Joinder restricted Fewer mass multi-defendant suits
False marking claims narrowed Ended one abuse category
TC Heartland on venue Constrained forum selection
Octane Fitness on fees Fee shifting more attainable
Alice on eligibility Many software patents invalidated

Several of these together changed the economics substantially. Cheaper validity challenges plus constrained venue plus more available fee shifting altered what a marginal assertion is worth.

Fee shifting

Available, and not routine.

Statute 35 U.S.C. 285
Standard Exceptional cases
Test Octane Fitness — totality of circumstances
Practical effect More attainable than pre-2014, still uncommon
Requirement Litigating to the end rather than settling

Pursuing fees means not settling, which is the tension. The cases where fee shifting is most deserved are frequently the ones a defendant most wants to end cheaply.

It does shape behaviour at the margin, because a plainly meritless campaign now carries a real risk that did not exist before Octane Fitness.

Joint defence and shared cost

Volume campaigns hit many companies with the same patent, which creates an opportunity.

Mechanism Benefit
Joint defence group Shared prior art searching and costs
Shared IPR petition One challenge, many beneficiaries
Common counsel Reduced duplication
Information sharing Others may already hold the killer reference
Defensive aggregator Pooled acquisition and licences

A single IPR can resolve a campaign against dozens of companies, because cancelling the claims removes the patent from everyone's problem at once.

Finding the other recipients is usually possible. Litigation dockets show who has been sued on the same patent, and that is a public record.

Reducing exposure

Measure Effect
Monitor your CPC classes Advance warning of acquisitions and filings
Watch NPE litigation in your area Who is asserting, against whom
Maintain defensive prior art Ready material for any assertion
Keep FTO analysis current Opinions age
Defensive aggregator membership Pooled acquisition and licences
Publish defensively Stops patents issuing at all

Monitoring is the highest-value measure because assertion campaigns are visible before they reach you. An entity acquiring patents in your classification is telegraphing its next move. See patent monitoring.

Defensive publication prevents the patent existing. It is cheap, immediate, and the only measure that removes the problem rather than preparing for it. See can you patent something and make it free.

Worked example: two letters

Same week, two demand letters, very different responses.

Letter A Letter B
Claim chart included No Yes, element by element
Accused feature named "your software products" A specific named module
Patent status Lapsed 3 years ago In force, second fee paid
Ownership Recorded to a dissolved entity Clean chain
Demand Flat licence, low four figures Negotiable, six figures
Deadline given 10 days 45 days

Response to A

Step Outcome
Patent Center check Expired for unpaid maintenance fee
Response Short letter noting expiry
Cost Under an hour
Result No further contact

The patent had lapsed and the letter was sent anyway. Ten minutes in Patent Center ended it.

Response to B

Step Outcome
Element-by-element comparison Three of four elements present
Fourth element Arguably present under equivalents
Prosecution history Claim narrowed to overcome prior art — estoppel likely
Prior art search Two strong references, both documents
Assessment Real exposure, but IPR viable
Action Opinion obtained; IPR prepared; negotiation opened from a position

Letter B was a genuine assertion and was handled as one. The estoppel finding and the prior art moved the negotiation substantially without litigation.

The difference between the two was found in a day, and the day cost less than either settlement.

What the label obscures

Some assertions by non-practising entities are meritorious, and treating the category as uniformly abusive causes real errors.

Reality Consequence of ignoring it
Patents from failed companies can be strong Dismissing a valid claim
Individual inventors often cannot enforce alone Assertion is their only route
University patents rely on licensing The model funds research
Some campaigns are purely opportunistic Paying claims that read on nothing

The analysis is the same regardless of who sent the letter. Is the patent in force, do the claims read on the product, and is there prior art. Who owns it affects negotiation, not infringement.

Insurance and funding

Mechanism What it covers
IP defence insurance Defence costs, sometimes settlements
Indemnity from suppliers Claims arising from supplied components
Customer indemnity you gave Your exposure to customers' claims
Litigation funding Plaintiff-side, occasionally defence
Defensive aggregator membership Pooled licences

Indemnities you granted are the exposure people forget. A supply agreement promising to indemnify customers against IP claims means their demand letters become yours.

Check both directions at the outset. What you are owed by suppliers, and what you owe customers, determines who actually funds the response.

Patent trolls: the checklist

  1. Never settle before checking the patent is in force. Lapse ends the matter for nothing.
  2. Verify current ownership. The sender must own what they assert.
  3. Read the independent claims yourself. Letters routinely overstate scope.
  4. Compare element by element. One missing element defeats literal infringement.
  5. Request a claim chart if none was provided. It costs nothing and tests seriousness.
  6. Read the prosecution history for narrowing that estops equivalents.
  7. Diarise the one-year IPR bar from service of any complaint.
  8. Price the design-around before pricing the settlement.
  9. Get a written opinion if you are continuing with knowledge of the patent.
  10. Monitor your CPC classes. Campaigns are visible before they arrive, and advance warning is the cheapest defence available.