China-Related UDRP Disputes: Evidence, Procedure and Cross-Border Collaboration

A Study of Selected WIPO and CAC Decisions | 2026 Research Series

Version 1.1 — public edition. First drafted 13 September 2026; revised after independent review on 15 September 2026.

Published by the UDRP research project of Wei Bin, LONG AN LAW FIRM. The cited cases are public research materials, not representative matters handled by the lawyer. This is the English counterpart of the Chinese stage report, adapted for international practitioners; it is not a completed full-year statistical report.

Executive summary

The value of studying China-related domain name disputes lies partly in understanding how business facts become verifiable propositions in a cross-border record. A trademark owner, an operating company and a product discussed in the media may appear under different names. A domain name may have been acquired, used and offered for sale at different times. A Chinese name may overlap with a trademark without resolving the question of rights or legitimate interests.

This report develops four propositions from selected decisions. The connection between rights, parties and alleged conduct needs evidence. Impersonation, criticism and an asserted independent project require different factual enquiries. Chronology should distinguish contemporaneous material from later explanations. Language choice and participation should be recorded separately: Chinese-language contacts do not by themselves determine the language of the proceeding; silence is not an automatic outcome and should not be used to infer language ability.

The research corpus comprises ten selected decisions involving eleven domain names, with decision dates from 13 January to 4 September 2026. It is exploratory and non-random. It does not measure the incidence of disputes involving Chinese parties, compare provider performance or estimate the effect of legal representation. The report proposes an entity map, an evidence chronology and a bilingual evidence index as practical tools for collaboration.

1. Scope and legal coordinates

“China-related” needs an explicit basis. This study uses party locations expressly recorded in decisions as its core entry criterion, distinguishing the complainant and respondent sides. Mainland China, Hong Kong (China), Macao (China) and Taiwan (China) are coded and reported separately. Chinese-language content, a Chinese name, trademark protection in China or a registrar located there does not independently establish the location of a party. Such other connections may be recorded as separate subject tags but are not added to this denominator.

These categories describe the research record, not nationality or a presumption about intent. A proceeding may involve parties in several locations; the sum of location tags is therefore not necessarily the number of cases. Translations and transliterations also need reconciliation before counting entities.

The study has no matched set of non-China-related cases. It consequently does not suggest that the difficulties discussed here are unique to Chinese parties or more prevalent among them.

The Policy requires the complainant to establish all three elements in paragraph 4(a). Administrative remedies under paragraph 4(i) are cancellation or transfer; paragraph 4(k) preserves the possibility of independent court proceedings. An objection to online content and a case for transfer should therefore be analysed separately. A denial does not affirm every aspect of a domain holder’s conduct. Claims and legal consequences in other jurisdictions require separate factual and legal analysis. ICANN, UDRP

Current commentary uses ICANN’s updated Rules and WIPO Overview 3.1. ICANN records that contracted parties could implement the updated text from 21 August 2024 and were required to do so no later than 21 August 2025. Overview 3.1 was announced on 17 February 2026 and summarizes panel views rather than creating strictly binding precedent. The version actually cited in a historical decision remains part of that decision’s record. ICANN Rules; WIPO announcement; Overview 3.1

2. What the corpus does and does not show

The initial corpus contains five WIPO and five CAC decisions. Seven ordered transfer and three denied the complaint. At domain level, eight domains were ordered transferred and complaints concerning three domains were denied. These are an inventory of the selected material, not annual success rates.

Working response coding records eight cases without a response, one with a substantive response and one whose submission form remains unclear despite recorded arguments. All ten are provisionally coded as not addressing RDNH, which is different from expressly rejecting an RDNH allegation. These research records remain subject to independent review and lawyer sign-off.

Six cases receive closer discussion because they illustrate different evidentiary questions. All ten are listed in Appendix A so that the reader can see the wider input set. Neither provider balance nor selection of contrasting outcomes makes the corpus representative. Earlier search-return totals are not used as its statistical denominator.

3. Brand protection: making the evidentiary connections visible

Corporate identity and brand evidence

In CAC-UDRP-108500, concerning hailuo-02.com, the first two elements were established but bad faith was not. The point went beyond document organisation. A case under paragraph 4(b)(iv) required a basis for finding that the respondent knew or should have known of the complainants’ own use and reputation in HAILUO, but the record did not attribute the relevant use and goodwill to the complainant entities. Decision, Bad Faith

The analytical point is about the record, not a judgment on commercial strength. People familiar with a business may understand immediately why a product name, a media reference and a corporate entity belong together. A panel needs a basis for making that connection from the material before it.

The proposed entity map should identify each relevant name, the relationship asserted, the evidence for it and the period to which it relates. A translation can clarify a label but cannot create an affiliation, authorization or chain of title. Trademark records, business records and publicity may serve different purposes and should not be presented as interchangeable.

Showing conduct rather than only resemblance

D2026-3614, concerning luckincoffeefranchise.com, resulted in transfer. The panel considered the branded website and franchise communications, including requests for payment; the decision does not establish that payment was received. It also relied on an earlier adverse decision in finding a pattern of abusive domain-name registrations, and addressed consolidation of the joint complaint and the transferee issue. Decision, sections 6–7

The comparison is not a controlled explanation of why one complaint succeeded and another failed. It nevertheless helps frame a practical question: can the reader follow who did what, through which channel, and how that conduct relates to the complaining party?

For a brand team, the proposed enquiry has three parts: what association the website presents; what visitors are invited to do; and what the evidence shows actually happened. A screenshot, an email and a transaction record may support different propositions. An allegation of a completed transaction should not be substituted for evidence of a request or an enquiry, and a complainant’s characterisation should not be presented as a panel finding.

4. Domain holders: testing independent explanations

Chronology and asserted independent purpose

In CAC-UDRP-108841, concerning 1688.ai, the panel treated two sets of material differently. AI-project material created about two months after registration did not refer directly to 1688.ai; it supported only the respondent’s continuing involvement in AI-related activity shortly after registration. An archived AGI webpage dated 2 April 2025, before the dispute, was the more significant support for the second-element analysis. Numeric wordplay alone was not a sufficient basis for the explanation, and the later sale evidence did not establish bad-faith registration on that record. Decision, Principal Reasons C–D

The proposed research method distinguishes an event’s date from the date of the material offered to prove it. Metadata, archived pages and a later explanation are different sources. Their connection and limitations should remain visible rather than being back-filled to the registration date.

The .ai suffix and an AI-related business purpose are separate variables. One can be identified from the domain; the other requires evidence. Keeping them separate avoids classifying every .ai dispute as an AI-industry dispute or overlooking an AI-related dispute under another suffix. Cultural explanations should retain their source rather than being embellished in translation.

Development plans, names and inactive sites

CAC-UDRP-108373, concerning stdecaux.com, recorded explanations of personal use and project development but ordered transfer. D2026-0913, concerning 德高商标.com, did not treat the overlap with the registrant’s personal name as sufficient and ordered transfer on the circumstances before the panel. CAC decision; WIPO decision, sections 6.2–7

These outcomes do not establish that a project must already be trading or that a personal name can never matter. They invite a more precise enquiry into the gap between an explanation and evidence supporting it. What plans, designs, expenditure, trials or other material predate the dispute? Does the material concern this domain or a corresponding name? What contrary material remains unexplained?

An inactive page records a state at a particular time. The investigation should also consider prior use, documented preparations and the sequence surrounding inactivity. Neither “the site is blank” nor “the project is in development” should end the factual enquiry.

5. Criticism and the limits of the question being decided

In D2026-2234, concerning reliancehithiumleaks.com, the second element was not established, the third was not reached and the complaint was denied. The panel expressly distinguished neutrality about the truth of criticism from endorsement of it. Decision, sections 6.B–C

Reading that outcome alongside an impersonation case shows why “contains the mark” is not an adequate research category on its own. The record also needs to explain the relationship presented to visitors, the use being made and the conduct at issue. This is a framework for questions, not a safe harbour for anything described as criticism.

For a business team, defining the objective helps prevent distinct complaints from being collapsed into one. Obtaining a domain, addressing an apparent impersonation, challenging particular content and seeking financial relief are not interchangeable objectives. Their feasibility requires separate analysis.

A precise outcome summary preserves what was left undecided. “The second element was not established” should not become a finding that the website is lawful in every respect or in any other jurisdiction. “The third element was not reached” should not become a finding that bad faith was absent. These distinctions are particularly vulnerable when a decision is shortened or translated.

6. Procedure: separate language, participation and form

Unless the parties agree otherwise or the registration agreement specifies otherwise, Rule 11 starts with the language of that agreement while allowing a different determination in the circumstances. Under Rule 14, a panel shall proceed to a decision following procedural non-compliance and shall draw such inferences as it considers appropriate; this is not an automatic result rule. ICANN Rules, paragraphs 11 and 14

D2026-0913 proceeded in English despite a Chinese registration agreement. That individual determination is not a universal rule for Chinese-language agreements. Decision, section 6.1

The unresolved submission form in CAC-UDRP-108373 illustrates a coding issue: recorded substantive arguments and confirmed formal response status are different observations. A quotation about default in another case must not be attributed to the present respondent, and the presence of arguments must not silently establish compliance with every procedural requirement.

The proposed intake record separates notification, commencement, response deadline, language requests and subsequent communications. Party-reported dates should remain distinguishable from provider notices until checked by responsible counsel. Translation capacity and work allocation belong in the practical plan; a failure to respond does not establish inability to understand English.

This corpus cannot estimate the effect of language choice on outcomes. Such a study would require consistent variables, more observations and attention to comparability, including the nature and strength of the claims.

7. A practical collaboration model

Three proposed work products could make a cross-border record easier to review. They are recommendations from this study, not mandatory forms or a statement of work already performed in the cited cases.

An effective handover should identify propositions to be checked rather than simply requesting translation of a folder. Does a document connect the product to the party? Does a design record concern the domain? Does a passage claim official status or comment on the brand? Defining the question makes it less likely that fluent language will conceal an evidentiary gap.

Responsibilities also need separation. One person may verify local records, another develop the overall argument, and another maintain the procedural calendar. Translation accuracy, factual verification and legal sufficiency are different judgments. The collaboration record should show who has addressed each.

The same discipline should apply to brand owners and domain holders. Contrary facts belong in the internal assessment, with unresolved questions retained. The report does not determine the collection or submission requirements of any particular jurisdiction; actual matters require separate review.

8. Conclusions

The common theme in these selected decisions is the distance between a commercial explanation and a verifiable record. Brand owners need to connect rights, entities and alleged conduct. Domain holders need to connect independent explanations with relevant material. Language, names and geographic context help a reader understand facts; they should not substitute for those facts.

A useful annual study needs both transparent coverage and substantive analysis. Coverage explains which decisions were examined. Analysis explains why they matter. A larger case list cannot repair an undocumented selection process, and a persuasive narrative cannot replace a traceable source.

This stage report completes a comparative account organized around identity, use, chronology and procedure. It supplies a basis for research and professional discussion, not population estimates or predictions. A full-year statistical component requires a closed observation window and a subsequent search for late-published decisions. If coverage remains incomplete, the annual publication should continue to identify itself as a study of selected decisions.

Appendix A. Corpus register

Dates are decision dates. Links are to official decisions. The Unicode and ASCII forms of the Chinese domain identify one domain. “China-related side” records the party-location basis for inclusion and does not state nationality.

At case level, seven decisions ordered transfer and three denied the complaint. At domain level, eight domains were transferred and complaints concerning three domains were denied. CAC-UDRP-108272 involved two domains and accounts for the difference.

Appendix B. Sources, coding and limitations

Official full texts are the primary sources. Search results, lists and media references are discovery aids. The current HTML and an earlier saved JSON version of CAC-UDRP-108841 differ in coverage; the reason has not been established. The analysis uses the HTML containing the principal reasons, while both versions are retained.

The last source review for this stage report was 12 September 2026. The selected decisions run from 13 January to 4 September 2026. This report is not a full-year search population.

Year allocation follows the decision date, not the case-number prefix. Registration, acquisition, renewal, website appearance and sale are separate events. Unknown dates remain unknown. Element coding distinguishes established, not established, not reached and unclear. RDNH coding distinguishes an affirmative finding, an express rejection, no discussion and uncertainty; denial alone is not an RDNH finding. ICANN Rules, paragraph 15(e)

Cases are deduplicated by provider and case number; domains by case and normalized domain. The current inventory reconciles as 7 + 3 = 10 cases and 8 + 3 = 11 domains. One two-domain proceeding accounts for the difference. No percentages are reported. Response definitions must be fixed before further statistical use, unknowns must remain visible, and overlapping location tags must not be added as if mutually exclusive.

The material is small, non-random and limited to two providers. It is not a full-year search population; corporate concentration has not been treated as independence. It cannot support causal claims about language, counsel or outcomes. Some historical quotations require further checking. Fees, post-decision implementation and later litigation are outside this study.

For annual completion, the target window is 1 January–31 December 2026, followed by a documented supplementary search for late publications. The publication should disclose providers covered, inclusion and exclusion records, data version, missing information and the last search date. Incomplete coverage must remain a stated limitation.