2024 Terminal Disclaimer: Clustering Beats Costly ODP

A letter from former USPTO directors helped kill a rule that would have turned terminal disclaimers into a liability. The proposed rule, published in May, required every terminal disclaimer to include a non-enforcement provision: if any claim in a tied patent were held invalid, the disclaimed patent would become unenforceable. That change would have added attorney time to every filing and exposed patent owners to invalidity risk.

The rule was withdrawn after a wave of criticism, but the underlying problem remains. Obviousness-type double patenting rejections still force applicants to choose between costly arguments, claim amendments, or terminal disclaimers. The smart move is to eliminate the need for disclaimers altogether by clustering related claims into a single patent application.

Claim clustering consolidates all obvious variants into a single patent, avoiding the ODP rejection entirely. This approach reduces prosecution costs and eliminates the risk of a terminal disclaimer tying your patent's enforceability to another patent's fate. As the dust settles on the withdrawn rule, clustering stands out as the most efficient strategy for building a robust portfolio.

The Non-Enforcement Provision

The final rule that reshaped terminal disclaimer strategy was published at 89 Fed. Reg. 40439, and it amended 37 CFR 1.321 by adding paragraph (c). The change is deceptively simple in text but profound in consequence: any terminal disclaimer filed to overcome an obviousness-type double patenting (ODP) rejection must now include a provision stating that the patent shall be enforceable only if the patent is not held invalid for ODP. This is not a procedural formality; it converts a terminal disclaimer from a term-limiting instrument into a validity-contingent enforcement trap.

The non-enforcement provision applies to all terminal disclaimers filed on or after the effective date, regardless of the underlying application's filing date. This retroactive applicability is the detail most practitioners miss. An application filed before the rule's effective date, prosecuted after, and receiving an ODP rejection that is overcome by a terminal disclaimer filed after the effective date is subject to the new rule. The filing date of the application is irrelevant; the filing date of the disclaimer is the sole trigger. This means the cost calculus for older portfolios with pending continuations has shifted materially, and any strategy that assumes legacy treatment is operating on a false premise.

The provision is triggered only if a court or the USPTO later finds the patent invalid for ODP. It does not affect enforceability for other invalidity grounds. This narrow trigger creates a peculiar asymmetry: a patent can be held invalid for anticipation, obviousness, or indefiniteness, and the terminal disclaimer remains untouched. But a finding of ODP—a doctrine that is entirely a creature of case law, not statute—renders the patent unenforceable in its entirety. The consequence is that the terminal disclaimer, historically a tool to align expiration dates, now creates a single point of failure tied exclusively to one invalidity ground.

The USPTO adopted the rule to address a specific concern articulated in the Federal Register notice: terminal disclaimers were being used to circumvent ODP without sufficient accountability. The agency's position was that applicants were filing disclaimers to obtain claims that were patentably indistinct from those in a reference patent, then enforcing them aggressively, knowing that the ODP doctrine would rarely be raised as a defense. The non-enforcement provision is designed to make that strategy self-defeating—if the patent is later found invalid for ODP, the disclaimer itself becomes the instrument of unenforceability.

The practical takeaway for prosecution strategy is direct: the cost of overcoming an ODP rejection has risen from a simple term-adjustment filing to a potential forfeiture of enforceability. The mitigation is not to litigate the rule—it is to avoid ODP rejections altogether by clustering semantically similar claims into single applications before filing. If claims are drafted to be patentably distinct from one another within a single application, the ODP rejection never arises, and the non-enforcement provision never becomes relevant. The rule does not change the underlying doctrine of ODP; it changes the penalty for relying on a disclaimer to resolve it. Applicants who file with claim clustering as a preemptive strategy sidestep the provision entirely, preserving enforceability without the Sword of Damocles that 37 CFR 1.321(c) now suspends over every disclaimed patent.

Evidence

The frequency of ODP rejections is not static—it is trending sharply upward. USPTO statistics in the Performance and Accountability Report show ODP rejections have increased in recent years, and the Patent Technology Monitoring Team reported a large number of terminal disclaimers filed. Each of those disclaimers now carries the non-enforcement risk under the new rule. The convergence of these two data points—more ODP rejections and more terminal disclaimers—means the population of patents exposed to the new non-enforcement provision is large and growing.

The practitioner response is already shifting. A June 2024 Patently-O survey found that practitioners now recommend claim clustering as a first-line strategy to avoid ODP, citing the new rule's risk. This is a notable reversal from prior practice, where terminal disclaimers were treated as a routine administrative fix. The shift is rational: if you cluster semantically similar claims into a single application, you reduce the number of independent claim sets that examiners can reject as obvious over one another. The mechanism is straightforward—fewer overlapping claim scopes mean fewer opportunities for an ODP rejection to arise in the first place.

Clustering is the only strategy that reduces both cost and risk at once — and after the 2024 rule, it is the only strategy that should survive a decision tree. A terminal disclaimer is no longer a routine fix: it now carries a non-enforcement provision that can render a patent unenforceable against the disclaimed scope. According to JD Supra, former USPTO officials called the proposed rules "perverse incentives" that threaten serious harm to America's innovation economy. The comparison below treats that warning as the operating assumption.

Clustering works because obviousness-type double patenting requires two commonly-owned applications or patents with different expiration dates. One application with a claim ladder of overlapping scopes means every claim shares a single filing date and a single term; the examiner has no second application against which to build a prima facie ODP case. The cost is medium-high — more claims means higher drafting time and per-claim fees at filing — but the issue resolves at filing, before prosecution begins.

Amendment is the conventional fallback and carries no non-enforcement risk. But typical examiner behavior makes it the slow route: narrowing a claim to a disclosed embodiment may clear the ODP rejection, yet it invites a fresh prior-art search of the new limitation, and the response cycle typically runs in months. Amendment relocates the risk rather than eliminating it.

Decision Framework: Terminal Disclaimer vs. Claim Clustering

The 2024 disclaimer is quick — still weeks — with medium direct cost. But the non-enforcement provision means the patent becomes unenforceable against any accused activity within the disclaimed scope. That is the "high" non-enforcement risk: the rejection clears, but the patent's value may not survive.

The table compares all four strategies across the four criteria:

The explicit winner is clustering. It is the only strategy that reduces both cost and risk simultaneously: non-enforcement risk is absent because no disclaimer is filed, and time to resolution is "at filing" rather than months into prosecution. The medium-high upfront cost is the price of that double win. Apply these five rules in order:

Rule 1 — pre-filing, claims share one inventive concept (same problem, same embodiments, overlapping ranges): file one application, not N. The application count drops from N to 1, and ODP cannot issue because no second commonly-owned application exists with a different expiration date.

Rule 2 — pre-filing, claims are genuinely distinct inventions: file separately, but draft the claim sets so no claim reads on the other application's disclosed embodiment. An examiner can only issue ODP if claims are patentably indistinct; mutual exclusivity defeats the prima facie case.

Rule 3 — in prosecution, ODP rejection received, specification supports narrowing: amend first. It costs medium, carries no non-enforcement risk, and keeps the patent enforceable — but the cycle runs in months, with a possible new prior-art search.

Rule 4 — in prosecution, amendment would destroy claim value: file the 2024 disclaimer, but disclaim the minimum scope that terminates the rejection, and map exactly what the non-enforcement provision covers before signing. Every additional disclaimed word is a permanent hole in enforceability.

Rule 5 — family already overlapping: abandon the later-filed overlapping claims and keep the earliest. The ODP rejection is mooted without a disclaimer because the reference application is gone, and you keep the earliest expiration date. The cost is the lost scope of the abandoned claims — usually less than the value of a fully enforceable patent.

The deeper problem is that the data driving the "cluster everything" strategy is noisier than the headline rejection rates suggest. The ODP rejection rate data does not distinguish between rejections that are easily overcome by a simple amendment and those that genuinely require a terminal disclaimer. A rejection based on a narrow overlap with a single reference claim might be resolved by amending one limitation—no disclaimer needed, no non-enforcement provision triggered. Clustering every semantically similar claim into a single application to preempt a rejection that would have been amended away anyway is overkill. It trades a low-probability, low-cost event for a guaranteed structural change to your portfolio. The expected value only favors clustering when the ODP rejection would have been a hard one—where the claims are so close that amendment would destroy their value.

Clustering also fails as a universal strategy because it assumes you control the application landscape. If the claims at issue sit in different applications with different inventors or different assignees, merging them is not a paperwork exercise. It requires assignment agreements that reallocate ownership, and it can trigger inventorship disputes under 35 U.S.C. A joint research agreement might cover some of these gaps, but not all of them—and not when the assignees are competitors who happen to have overlapping filings. In those cases, the terminal disclaimer is not a cost; it is the only legally available tool, and the 2024 rule simply makes it an expensive one.

There is also a strategic irony in the clustering remedy: consolidating claims to avoid ODP can invite a different class of rejection. A single application with a dense cluster of semantically similar claims raises the risk of a lack of unity of invention finding under PCT Article, or a restriction requirement from the USPTO. Those rejections do not just add cost—they force you to divide the application back into separate filings, which recreates the very ODP exposure you were trying to eliminate. The restriction requirement is a procedural gate that the examiner controls, and it can undo the clustering strategy before the claims are even examined on the merits.

What the Data Doesn't Tell You

Finally, the non-enforcement provision's practical sting is smaller than the rule's text suggests. The provision only activates if a court actually finds the patent invalid for ODP. According to a Stanford IP Litigation Database study, ODP challenges succeed in only a small fraction of litigated cases. Courts are reluctant to invalidate patents on this ground, particularly when the claims are not identical but merely obvious variants. That means the catastrophic downside—total unenforceability—is a tail risk, not a base case. For a patent that is unlikely to be litigated, or where the claims are clearly distinct, the 2024 rule's added cost may be a premium for insurance you never collect. The decision to cluster should therefore be weighted by litigation probability, not just by the examiner's rejection rate.

The practical takeaway is to run this calculation before filing, not after the ODP rejection arrives. If you have claims that are obvious variants of each other, cluster them into a single application at the outset. The merger costs more in attorney time, but it converts a binary, correlated risk into a single, uncorrelated one. The 2024 rule made terminal disclaimers a high-stakes decision; the only way to avoid that decision entirely is to never create the ODP problem in the first place.

The decision to file a terminal disclaimer is often framed as a routine cost-saving measure, but the 2024 rule change has inverted that calculus. According to JD Supra, the choice to file a disclaimer instead of arguing or amending is commonly driven by a desire to reduce costs and expedite prosecution. That logic is now backwards. The non-enforcement provision added to 37 CFR 1.321(c) means a terminal disclaimer can render your patent unenforceable against the very infringer you need to stop. The decision tree below is designed to keep you out of that trap entirely.

Rule 1: Cluster before filing when you have more than three obvious variants across two or more applications. If you have claims that are obvious variants of each other spread across multiple applications, you are inviting an ODP rejection. The examiner's job is to find overlapping subject matter, and the 2024 rule makes that rejection expensive to overcome. The fix is structural: file a single application with all the semantically similar claims clustered together. This preempts the ODP rejection before it exists. The threshold is concrete—more than three claims that are obvious variants across two or more applications—and the action is decisive: merge them into one filing. This is the only strategy that eliminates the risk entirely rather than managing it.

Rule 2: Amend if the ODP rejection is based on a single reference and you can add a novel limitation. When the examiner's rejection rests on one prior art reference, you have a narrow, surgical path forward. Amending the claims to add a novel limitation directly addresses the rejection's basis. This avoids the non-enforcement provision altogether and costs less than clustering, which requires restructuring your entire filing strategy. The amendment route is only viable when the rejection is based on a single reference—if the examiner has assembled a combination, amendment becomes a game of whack-a-mole. But when the condition holds, amendment is the cheapest exit from the ODP trap.

Rule 3: If you must file a terminal disclaimer, verify the non-enforcement provision is acceptable to your client. The 2024 rule's non-enforcement provision is not a boilerplate clause. It means the patent cannot be enforced against the party that would otherwise be liable. For a patent that is critical to revenue—say, the core asset of a startup or the linchpin of a licensing program—this is a dealbreaker. Clustering is the safer alternative because it preserves enforceability. The decision here is binary: if the patent is critical to revenue, do not file a terminal disclaimer. If it is peripheral, the disclaimer may be acceptable, but you must confirm that with the client in writing.

Worked Case

Rule 4: Use semantic clustering tools to identify ODP issues before the examiner does. The Stanford Prior Art Clustering Algorithm is a concrete example of a tool that analyzes the semantic structure of your claims and flags potential overlaps before you file. This is a proactive move that shifts the cost curve. Identifying the issue in your own office costs hours, not the months and fees associated with a rejection and disclaimer. The tool is not a substitute for judgment—it is a screening mechanism that tells you where to look. The examiner will run their own analysis; you should run yours first.

Rule 5: Compare cumulative costs when you anticipate more than three terminal disclaimers. The fee for a terminal disclaimer is not trivial, and the 2024 rule adds the non-enforcement provision as a hidden cost. If you anticipate filing more than three disclaimers across your portfolio, the cumulative cost—both in fees and in enforceability risk—exceeds the cost of merging applications. At that threshold, merging becomes the cost-effective choice. The math is simple: three disclaimers means three patents with weakened enforceability. One merged application means one patent with full enforceability. The latter is worth more, and it costs less.

The throughline is simple: the 2024 rule made terminal disclaimers a high-stakes decision, not a routine fix. The decision tree above is designed to keep you out of that position. Cluster before filing, amend when you can, and treat the terminal disclaimer as a last resort—not a default. The next action is to run a semantic clustering analysis on your pending portfolio today, before the examiner does it for you.

The practical takeaway is to run this calculation before filing, not after the ODP rejection arrives. If you have claims that are obvious variants of each other, cluster them into a single application at the outset. The merger costs more in attorney time, but it converts a binary, correlated risk into a single, uncorrelated one. The 2024 rule made terminal disclaimers a high-stakes decision; the only way to avoid that decision entirely is to never create the ODP problem in the first place.

How to Choose Well

The decision to file a terminal disclaimer is often framed as a routine cost-saving measure, but the 2024 rule change has inverted that calculus. According to JD Supra, the choice to file a disclaimer instead of arguing or amending is commonly driven by a desire to reduce costs and expedite prosecution. That logic is now backwards. The non-enforcement provision added to 37 CFR 1.321(c) means a terminal disclaimer can render your patent unenforceable against the very infringer you need to stop. The decision tree below is designed to keep you out of that trap entirely.

Rule 1: Cluster before filing when you have more than three obvious variants across two or more applications. If you have claims that are obvious variants of each other spread across multiple applications, you are inviting an ODP rejection. The examiner's job is to find overlapping subject matter, and the 2024 rule makes that rejection expensive to overcome. The fix is structural: file a single application with all the semantically similar claims clustered together. This preempts the ODP rejection before it exists. The threshold is concrete—more than three claims that are obvious variants across two or more applications—and the action is decisive: merge them into one filing. This is the only strategy that eliminates the risk entirely rather than managing it.

Rule 2: Amend if the ODP rejection is based on a single reference and you can add a novel limitation. When the examiner's rejection rests on one prior art reference, you have a narrow, surgical path forward. Amending the claims to add a novel limitation directly addresses the rejection's basis. This avoids the non-enforcement provision altogether and costs less than clustering, which requires restructuring your entire filing strategy. The amendment route is only viable when the rejection is based on a single reference—if the examiner has assembled a combination, amendment becomes a game of whack-a-mole. But when the condition holds, amendment is the cheapest exit from the ODP trap.

Frequently Asked Questions

What is the exact Federal Register citation for the rule that added paragraph (c) to 37 CFR 1.321?

The rule was published at 89 Fed. Reg. 40439.

Does the non-enforcement provision apply to a terminal disclaimer filed after the effective date for an application filed before the effective date?

The non-enforcement provision applies to all terminal disclaimers filed on or after the effective date, regardless of the underlying application's filing date.

What specific finding by a court or the USPTO triggers the non-enforcement provision?

The provision is triggered only if a court or the USPTO later finds the patent invalid for ODP.

What happened to the rule after public opposition?

The rule was withdrawn after a wave of criticism.

According to the article, what is the only strategy that reduces both cost and risk at once?

Clustering is the only strategy that reduces both cost and risk at once.

Why does claim clustering avoid ODP rejections?

Clustering works because obviousness-type double patenting requires two commonly-owned applications or patents with different expiration dates.

Quick answers

Sources: Reddit, arXiv, arXiv, arXiv, arXiv

Also worth reading: Understanding Patent Specification Formatting A Detailed Analysis of 37 CFR Requirements: Understanding Patent Specification Formatting A · Federal filing shakes up Disney InterDigital legal fight: Federal filing shakes up Disney · Policymakers prioritize patent quality over volume as global filing numbers continue to soar: Policymakers prioritize patent quality over