T-Mobile v. KAIFI: When a Patent Settlement Becomes a Contract Case

A patent case can end without ending the parties’ risk. If a settlement ties a later payment to the outcome of a U.S. Patent and Trademark Office proceeding, a few words in the agreement may determine whether the deal works as intended—or produces a second lawsuit.

The Federal Circuit’s August 28, 2026 decision in T-Mobile US, Inc. v. KAIFI LLC illustrates that problem. The underlying dispute involved a patent, ex parte reexamination, claim scope, and alleged inequitable conduct. Yet the Federal Circuit did not decide the contract merits. It held that the appeal belonged in the Fifth Circuit because KAIFI’s claim was fundamentally a state-law contract claim that did not necessarily raise a substantial patent-law issue.

For patent owners, accused infringers, licensees, and businesses negotiating around a pending USPTO challenge, the decision is more than a jurisdictional ruling. It is a drafting warning: define the post-grant event that triggers payment with enough precision that a court does not have to reconstruct the parties’ business deal later.

The deal hinged on whether a patent claim “survives” reexamination

KAIFI sued T-Mobile in the Eastern District of Texas on August 28, 2020, alleging infringement of U.S. Patent No. 6,922,728, which concerns Wi-Fi calling technology. T-Mobile requested ex parte reexamination of the patent on July 13, 2021, and the USPTO ordered reexamination on August 19, 2021.

With the reexamination still pending and trial approaching, the parties settled. Their January 12, 2022 Settlement and Patent License Agreement gave T-Mobile a license, required an upfront payment, and provided for an additional payment “if any Asserted Claim survives the EPR.” The agreement identified claims 1, 2, 3, 5, 7, 10, 11, and 20 as the asserted claims. It also included a covenant restricting T-Mobile from challenging the validity or enforceability of the licensed patents, while excepting its already-filed reexamination request.

The USPTO later issued a reexamination certificate confirming claims 1, 2, 3, 5, 10, 11, and 20 as patentable without amendment. Claim 7 was determined patentable as amended. T-Mobile did not make the contingent payment. Instead, it filed a declaratory-judgment action contending that none of the asserted claims had survived within the meaning of the agreement.

T-Mobile’s theory did not rest solely on the formal result printed in the reexamination certificate. It argued that KAIFI had changed its claim-scope positions during reexamination and had engaged in inequitable conduct, so the infringement case the parties had settled did not meaningfully survive. KAIFI counterclaimed for breach of contract. The district court granted summary judgment to KAIFI and ordered T-Mobile to make the additional payment.

The Federal Circuit’s precedential opinion in T-Mobile v. KAIFI did not affirm or reverse that merits ruling. It transferred the appeal.

Why the Federal Circuit transferred the appeal

The Federal Circuit has exclusive appellate jurisdiction over district-court civil actions arising under federal patent law. But a lawsuit does not arise under patent law merely because a patent appears in the facts, the contract is a patent license, or the disputed payment depends on a USPTO proceeding.

Under the Supreme Court’s framework in Christianson v. Colt Industries Operating Corp. and Gunn v. Minton, a state-law claim may fall within federal arising-under jurisdiction only when a federal issue is necessarily raised, actually disputed, substantial to the federal system as a whole, and capable of resolution in federal court without disturbing the congressionally approved federal-state balance.

The Federal Circuit found at least two problems with treating KAIFI’s breach-of-contract claim as a patent case.

First, a patent-law issue was not necessarily raised. The district court could resolve the contract claim by reading “survives” in its ordinary sense and comparing the agreement with the reexamination certificate. The certificate confirmed several specified claims as patentable without amendment. That path did not require a court to decide prosecution disclaimer, inequitable conduct, or another question created by patent law.

Second, any embedded patent question was not substantial in the jurisdictional sense. The dispute was backward-looking and specific to one private contract. Resolving it would not control numerous other cases or threaten the uniform development of patent law. The government also had no direct interest in which private party prevailed on the payment dispute.

The panel therefore transferred the appeal to the Fifth Circuit, the regional circuit with appellate jurisdiction over cases from the Eastern District of Texas. That transfer leaves the contract merits for another court. It also prevents readers from treating the district court’s construction of “survives” as though the Federal Circuit had adopted it as a rule of patent law.

The central drafting lesson: describe the event, not the shorthand

The agreement used a compact phrase that may have sounded clear during negotiations. But “survives the EPR” left room for competing conceptions of success. Did a claim survive if the USPTO confirmed its patentability? Did it need to remain unamended? Did it need to preserve the same claim scope or infringement position that existed in the district-court case? What if the claim remained formally intact but later faced an enforceability challenge?

The Federal Circuit observed that future parties can avoid uncertainty by using different language. Its example referred to an asserted claim “confirmed to be patentable in the reexamination proceeding.” That formulation anchors the payment trigger to an identifiable agency result rather than an undefined characterization of what survived.

The broader point applies to settlements tied to ex parte reexamination, inter partes review, post-grant review, reissue, or ordinary prosecution. The agreement should identify the precise agency event and its legal effect. A drafter may need to specify whether the trigger is a final written decision, a reexamination certificate, a notice of intent to issue a certificate, termination of the proceeding, expiration of rehearing rights, or completion of appellate review.

That is especially important when the settlement is part of a broader patent licensing arrangement. The payment language, release, license scope, covenant not to sue, and challenge restrictions should operate together rather than point to different definitions of the same event.

Amended claims need their own treatment

The facts also show why a settlement should address amended claims directly. Seven asserted claims were confirmed without amendment, while claim 7 was patentable as amended. The parties’ drafting history reflected disagreement about whether an amended claim would survive, but that disagreement did not decide the Federal Circuit’s jurisdictional analysis because multiple unamended claims had been confirmed.

Future agreements should not leave this issue implicit. An amended claim might be treated as satisfying the trigger, failing the trigger, or satisfying it only if the amendment meets an agreed standard. For example, the parties may care whether the amendment materially narrows claim scope, preserves coverage of an accused product, creates intervening-rights issues, or changes a royalty base. The right answer depends on the negotiated business objective, not on a universal meaning of “survive.”

Patent owners preparing for PTAB and post-grant proceedings should consider those consequences before accepting a contingent settlement structure. Accused infringers should do the same before agreeing that an agency outcome automatically releases additional money or changes a continuing royalty.

Separate patent status, claim scope, and enforceability

T-Mobile v. KAIFI also exposes the danger of asking one undefined phrase to carry several legal concepts.

A claim’s formal status after USPTO review is one question. The scope of that claim in light of the intrinsic record is another. Enforceability based on alleged inequitable conduct is yet another. A settlement can make one or more of those concepts relevant, but it should say which ones control the parties’ obligations.

The distinction is particularly important when an agreement contains a no-challenge provision. Here, T-Mobile’s theory that alleged inequitable conduct meant no claim survived appeared to collide with its covenant not to challenge validity or enforceability. A carefully integrated agreement should make clear which contentions remain permitted, whether they may be raised only defensively, and whether a permitted USPTO proceeding has consequences beyond the proceeding itself.

Businesses involved in patent litigation should review those provisions as a single allocation of risk. A payment trigger should not silently preserve a theory that another section appears to waive.

Forum provisions cannot solve every jurisdiction problem

The case offers a second strategic lesson. Parties often assume that a patent-centered dispute will be reviewed by the Federal Circuit. That assumption can be wrong when the pleaded claim is created by state law and patent issues are not necessary and substantial under the federal jurisdiction test.

A forum-selection clause and governing-law clause remain important, but private agreement cannot create federal subject-matter jurisdiction or appellate jurisdiction that Congress has not supplied. Counsel negotiating a patent settlement should therefore consider where a future contract claim may be filed, what substantive law will govern, and which regional circuit may review the resulting judgment.

This matters because regional circuits may apply contract doctrines and procedural rules differently. The expected appellate path can affect litigation strategy, preservation of arguments, and the value of clarity in the agreement itself.

A practical framework for contingent patent settlements

Before signing an agreement tied to a USPTO outcome, the parties should be able to answer a few concrete questions from the text alone:

• Which patent claims and which administrative proceeding control the trigger?
• What document or procedural event establishes the outcome?
• Are confirmed, amended, substituted, and newly added claims treated differently?
• Must the result be final after rehearing and appeal, or does an earlier event control?
• Does claim scope, product coverage, validity, or enforceability matter in addition to formal claim status?
• How do the trigger and any no-challenge covenant interact?
• When is payment due, how is notice given, and what happens if the result is mixed?

Those are drafting questions, but they are also valuation questions. The parties cannot price the settlement intelligently until they agree on what outcome they are buying and what future risk they are retaining.

What patent owners and technology companies should take away

The Federal Circuit did not announce a universal interpretation of “survives” and did not decide who ultimately wins the appeal on the contract merits. Its holding was jurisdictional: the state-law contract claim did not necessarily raise a substantial patent-law issue, so the appeal belonged in the Fifth Circuit.

Still, the opinion delivers a concrete message. Patent settlements should convert the parties’ business understanding into objective language tied to identifiable events. That requires coordination among the people evaluating infringement, claim scope, USPTO procedure, licensing economics, and contract enforcement.

Tucker Law advises inventors and businesses on patent strategy and related intellectual-property disputes and transactions. If a settlement, license, or post-grant proceeding may affect important payment or enforcement rights, contact Tucker Law to discuss the agreement before ambiguity becomes the next case.

This article provides general legal information and is not legal advice. The interpretation and enforceability of any agreement depend on its language, the governing law, and the specific facts.