Decision for dispute CAC-UDRP-108413

The Panel is not aware of any other legal proceedings which are pending or decided and which relate to the disputed domain name.

Complainant is the owner of, amongst others, the registrations for trademarks TONGYI QWEN and QWEN.

The most relevant registrations in this case are mentioned here below:

- Hong Kong TM No. 306741423 – QWEN – Nice Cl. 9, 42, registered on November 27, 2024;

- Philippines TM No. PH4202400532848 – QWEN – Nice Cl. 9, 42, registered on January 18, 2025;

- Malaysian TM No. TM2024038102 - QWEN – Nice Cl. 9, 42, registered on December 5, 2024.

Complainant has submitted evidence of the above-mentioned registrations by copies of the trademark registrations or print outs from the official databases of the respective trademark registers.

Complainant is Alibaba Innovation Private Limited, a company with its domicile in Singapore. Complainant belongs to the Alibaba Group, a leading Chinese multinational conglomerate founded on June 28, 1999, in Hangzhou, Zhejiang, China.

Complainant states that the Alibaba Group operates globally across a wide range of sectors including e-commerce, retail, Internet, and technology. It offers consumer-to-consumer (C2C), business-to-consumer (B2C), and business-to-business (B2B) sales services through both Chinese and international marketplaces. In addition, the Group provides services in digital media and entertainment, logistics, and cloud computing. Alibaba Innovation Private Limited is part of this global ecosystem, which also includes entities such as Alibaba (China) Co., Ltd. and Alibaba.com Corporation.

Further Complainant alleges that Alibaba is one of the world’s largest and most influential companies in the fields of retail and e-commerce and is active in over 190 countries. In addition to its leadership in online commerce, the company was ranked in 2020 as the fifth-largest artificial intelligence enterprise globally. Through its fintech subsidiary, Ant Group, Alibaba also operates as the world's second-largest financial services provider after Visa. Furthermore, it is recognized as one of the most prominent venture capital and investment firms worldwide. Moreover, Alibaba operates the world’s largest online marketplace across all major segments: B2B (Alibaba.com), C2C (Taobao), and B2C (Tmall). The Group has also significantly expanded into the media, entertainment and artificial intelligence sectors, with revenues in these areas growing at a triple-digit rate year over year. Notably, during the 2018 edition of China’s Singles’ Day – the world’s largest online and offline shopping event – Alibaba set a global sales record.

In addition, as part of its AI strategy, Complainant informs that Alibaba has developed QWEN, a family of cutting-edge large language models (LLMs) designed for multilingual understanding, reasoning, and enterprise applications. These models are accessible via Alibaba Cloud and form the foundation of the Group’s generative AI ecosystem.

Complainant has systematically promoted its Qwen models through a series of public releases, both internationally and domestically. Between 2023 and 2025, Alibaba Cloud issued numerous official announcements – via its global press room, Alzilla, and Chinese platforms such as Baidu – highlighting the open-sourcing, technical upgrades, and enterprise adoption of Qwen models. These releases demonstrate a coordinated communication strategy aimed at establishing Qwen as a leading open source. In order to substantiate this Complainant has enclosed a list with links to press releases of its models, being overseas release and domestic release, as well as ranking lists and photos taken on exhibitions.

According to Complainant, the QWEN model has achieved top-tier recognition across several international and Chinese AI benchmarks, including OpenCompass, the Open LLM Leaderboard, Compass Arena, and Chatbot Arena. Specifically, Qwen-72B, Qwen2-72B, and Qwen2.5 consistently ranked first or among the top performers in evaluations published between 2023 and 2024, with scores demonstrating superior performance compared to other open-source models. These results have been widely documented by official sources and leading technological media outlets, as Complainant states. In this respect Complainant has enclosed a WORD document with a ranking list with links to articles in Chinese in which QWEN is mentioned. Also, the date of the article can be read in English.

Complainant continues by alleging that Alibaba has actively showcased its Qwen models at major industry events, including the APSARA Conference held annually in Hangzhou. The Qwen models were exhibited in the 2023, 2024, and 2025 editions of the conference, as evidenced by official event schedules and photographic documentation. These exhibitions serve as further proof of the public disclosure and promotion of the Qwen models to both domestic and international audiences. Attached to the complaint are indeed photos on which QWEN is depicted and taken on exhibitions. No objective dates can be perceived though.

Further, Complainant explains that Complainant’s AI model family was originally launched under the name “Tongyi Qianwen”, which is the Chinese designation of the technology. In international and technical contexts, the Complainant adopted the abbreviated name “Qwen” to refer to the same suite of large language models (LLMs) and multimodal models (MLLMs). This equivalence is explicitly confirmed in official communications by Alibaba Cloud, which states that “Qwen (also known as Tongyi Qianwen)” refers to the same proprietary AI model family. Accordingly, the term “Qwen” has become the globally recognized identifier of the Complainant’s technology.

Complainant registered the domain names:

- <qwen.ai> on August 27, 2023, which resolves to Complainant’s official website https://qwen.ai/; and

- <qwenchat.com> on December 1, 2024.

A screenshot of Complainant’s official website is enclosed to the Complaint.

Complainant further states that it has consistently promoted the “QWEN” denomination across major online platforms, including the official website https://qwen.ai/, as well as verified accounts on leading social media and developer environments such as GitHub, Hugging Face, X, Discord, LinkedIn. Complainant encloses a screenshot of the aforementioned platforms, which show substantive amounts of followers, at least on some in connection with QWEN: GitHub 11,3K, X 117,9 K, Discord 19.050 Members, LinkedIn 30 K followers.

Complainant concludes that, within the broader domain of multimodal AI technologies, QWEN has emerged as a distinctive identifier closely associated with Complainant’s brand and product suite.

Respondent registered the domain name <qwenchat.ai> on January 29, 2025. The domain name is used for a website for conversation services.  

Complainant contends that the requirements of the Policy have been met and that the disputed domain name should be transferred to it.
No administratively compliant Response has been filed.

According to the Policy paragraph 4(a)(i) it needs first to be established that:

(i) The domain name is identical or confusingly similar to a trademark or service mark in which the complainant has rights;

Complainant states that it has rights in QWEN by virtue of its trademark registrations.

Further, the disputed domain name incorporates Complainant’s trademark “QWEN” in its entirety, followed by the descriptive term “chat”, resulting in a domain name that is significantly similar to Complainant’s trademarks. The addition of “chat” does not diminish the confusing similarity; on the contrary, it reinforces it, as the term directly relates to conversational AI functionalities — an area closely connected to the Complainant’s QWEN models, which operate in the field of artificial intelligence and natural language processing. Notably, the disputed domain name replicates the structure of the Complainant’s domain qwenchat.com, registered on December 1, 2024, further underscoring confusing similarity.

In this respect Complainant refers to a previous UDRP decision in which it was decided that including a descriptive term that is intrinsically linked to Complainant’s core technology increases, rather than mitigates, the likelihood of confusion. (see, e.g., WIPO Case No. D2024‑1807, Equifax Inc. v. Domain Admin, TotalDomain Privacy Ltd).

Also, Complainant states that panels have also confirmed that top‑level domains (“TLDs”), including “.ai”, are normally disregarded when assessing identity or confusing similarity under the first element of the Policy (see WIPO Case No. D2005‑0540, RX America, LLC v. Mattew Smith; WIPO Case No. D2006‑0582, Sanofi‑Aventis v. US Online Pharmacies; and WIPO Case No. D2007‑0768, Playboy Enterprises International, Inc. v. Zeynel Demirtas).

The relevant analysis therefore focuses on the second‑level portion of the disputed domain name, which reproduces Complainant’s trademark in a manner that clearly creates a likelihood of confusion.

In light of the foregoing, Complainant concludes that the disputed domain name is highly and confusingly similar to Complainant’s prior registered trademarks, in accordance with Paragraph 4(a)(i) of the Policy.

The Panel has the following observations.

The UDRP procedure is a fast procedure in which rights and facts should be clearly displayed so that the Panel can make a quick assessment as well as render a well-founded quick decision.

Complainant has submitted a lengthy list of trademarks TONGYI QWEN and QWEN registered or filed in countries throughout the world. However most invoked trademarks are irrelevant.

First of all, when assessing the identity or similarity between trademark TONGYI QWEN and the disputed domain name <qwenchat.ai>, the Panel concludes that the two are not identical.

As for similarity, the Panel finds that the assessment is complex. The trademark TONGYI QWEN consists of two elements that at first sight appear to be equally dominant. The weighing of them in relation to the disputed domain name needs an extra layer of legal assessment. Moreover, it needs to be defined according to what legal equipment that assessment will be used. In fact, such a comparison is normally only assessed according to (national) trademark law and not clear is here what trademark law should be used, if any. Indeed, the Policy does not provide a guidance what trademark law is applicable. In fact, paragraph 1.7 of the WIPO Jurisprudential Overview 3.0 states on the question: ‘What is the test for identity or confusing similarity under the first element?’:

The standing (or threshold) test for confusing similarity involves a reasoned but relatively straightforward comparison between the complainant’s trademark and the disputed domain name.

The Panel concludes that the test cannot be made straightforward and in line with the ratio of the UDRP procedure, which procedure should rule the obvious cases with a clear and quick ruling, will disregard the trademark registrations for TONGYI QWEN as being too complex to compare them with the disputed domain name.

The Panel will accordingly base the identity and similarity assessment between trademark and disputed domain name on Complainant’s trademark registrations for QWEN.

In this respect, the Panel finds that only trademarks that have been registered and predate the disputed domain name are relevant because these trademarks constitute rights that can be invoked against Respondent.

For this reason, the Panel has only mentioned the trademarks registered in Hong Kong, The Philippines and Malaysia as relevant in the section Identification of Rights.

Thus, the assessment that needs to be made is between QWEN on one hand and <qwenchat.ai> on the other. The Panel finds that the disputed domain name is similar to the invoked trademark as it comprises the trademark followed by the descriptive term ‘chat’.

Further, it is commonly accepted amongst panelists that the Top Level Domain, in this case, ai, can be disregarded.

Based on those three registrations for the trademark QWEN the Panel concludes that Complainant has, to the satisfaction of the Panel, shown the disputed domain name is confusingly similar to the trademarks in which Complainant has rights (within the meaning of paragraph 4(a)(i) of the Policy).

According to the Policy paragraph 4(a)(ii) it needs further to be established that:

(ii) The respondent has no rights or legitimate interests in respect of the domain name.

Paragraph 4 (c) of the Policy provides circumstances that could demonstrate that Respondent has no rights to and legitimate interests in the disputed domain name. These circumstances are not exclusive.

Circumstance that are providing rights or legitimate interests to the domain name are:

(i) before any notice to the respondent of the dispute, the use of, or demonstrable preparations to use, the domain name or a name corresponding to the domain name in connection with a bona fide offering of goods or services; or

(ii) respondent (as an individual, business, or other organization) has been commonly known by the domain name, even if though it has acquired no trademark or service mark rights; or

(iii) respondent is making a legitimate noncommercial or fair use of the domain name, without intent for commercial gain to misleadingly divert consumers or to tarnish the trademark or service mark at issue.

Further, according to the WIPO Case No. D2003-0455, Croatia Airlines d. d. v. Modern Empire Internet Ltd., Complainant is required to make out a prima facie case that Respondent lacks rights or legitimate interests. Once such prima facie case is made, Respondent carries the burden of demonstrating rights or legitimate interests in the domain name. If Respondent fails to do so, Complainant is deemed to have satisfied paragraph 4(a) (ii) of the Policy.

Complainant first of all states, as a preliminary observation, that it is well established in prior UDRP decisions – including WIPO Case No. D2000-0134, Pharmacia & Upjohn Company v. Moreonline, and WIPO Case No. D2000-0118, Chargers Football Company v. One Sex Entertainment Co., a/k/a chargergirls.net – that the mere registration of a domain name does not, in itself, confer rights or legitimate interests upon the Respondent.

Further Complainant states that Respondent is neither a licensee nor an authorized distributor of Complainant and has not been granted any permission to use Complainant’s trademarks – nor any variation, misspelling, or derivative thereof – in the disputed domain name or in any other manner.

Complainant is not aware of any evidence indicating that Respondent is commonly known by the disputed domain name, or by any name corresponding to it, whether as an individual, business, or other organization. Furthermore, the publicly available Whois records, as attached to the Complaint, do not disclose any identifying information. In accordance with the Policy, this absence of evidence supports the conclusion that Respondent has not acquired rights or legitimate interests in the disputed domain name.

Moreover, although Complainant has made its QWEN AI model series available on an open‑source basis, this does not grant third parties any right or authorization to register domain names incorporating Complainant’s trademarks. Respondent has not been licensed, authorized, or otherwise permitted by Complainant to make use of the “QWEN” mark in any form, including within a domain name.

Respondent is not making a bona fide offering of goods or services, nor is it engaged in any legitimate non‑commercial or fair use of the disputed domain name. On the contrary, Respondent is using the disputed domain name to promote AI‑related conversation services under the “QWEN” name, without any disclaimer or indication of the lack of affiliation with Complainant. This conduct misleads users, exacerbates the likelihood of confusion, and interferes with Complainant’s ability to secure and use a domain name that aligns with its brand identity and reputation.

The Panel follows above reasoning.

Any further arguments of Complainant in this respect which the Panel summarizes as:

• Respondent registered the distinctive trademark of Complainant without legitimate connection or authorization and Respondent should have known about it due to the distinctive nature of the trademark;
• The disputed domain name was registered well after the filing of the trademark applications of Complainant and Respondent should have known about the trademark applications;
• Complainant had already publicly launched TONGYI QWEN in 2023, which is before Respondent’s legal activities began and thus Respondent should have known about the launch.

Will be set aside as the Panel can decide on the previous mentioned reasoning and concludes that Complainant has, to the satisfaction of the Panel, shown Respondent to have no rights or legitimate interests in respect of the disputed domain name (within the meaning of paragraph 4(a)(ii) of the Policy).

According to the Policy paragraph 4(a)(iii) it finally needs to be established that:

(iii) The domain name has been registered and is being used in bad faith.

In this case Complainant asserts that it is reasonable to consider that Respondent could not have been unaware of the existence of Complainant’s trademarks. It has several arguments for that:

1. Disputed domain name was registered after promotion of Complainant’s trademarks

When registering the disputed domain name, on January 29, 2025 Complainant had already adopted and publicly promoted the “QWEN” and “TONGYI QWEN” trademarks, which had garnered significant attention in the AI industry.

Specifically, Complainant states that it has consistently and publicly promoted its QWEN AI models across major international and domestic platforms since 2023, including through official announcements, benchmark rankings, and industry events. The QWEN models have achieved top-tier recognition in global AI evaluations and have been prominently featured at conferences such as APSARA. In this respect Complainant refers to enclosed evidence, displayed in a WORD document with links to articles in English and Chinese.

Also, as further illustration Complainant encloses VPN-based searches conducted from the country where Respondent appears to be geolocated, and which show that Complainant’s “QWEN” trademark appears among the top search results. According to Complainant this confirms that Respondent could not have been unaware of Complainant’s brand when registering the disputed domain name.

2. QWEN trademark is only associated with Complainant

Complainant continues with alleging that the term “QWEN” is neither a generic nor a commonly used word in English, nor does it possess an independent meaning outside its association with the Complainant’s AI model series. It is a stylized transliteration of the Chinese name “通义千问” (TONGYI QIANWEN), which identifies Complainant’s original product line and has been consistently used throughout the Complainant’s domestic and international branding efforts.

3. Identical domain name composition <qwenchat.com> (Complainant_ vs <qwenchat.ai> (Respondent)

The disputed domain name mirrors the structure of the domain name <qwenchat.com>, already registered and owned by Complainant since December 1, 2024, and therefore predating Respondent’s registration. This duplication of Complainant’s existing domain format provides additional evidence of Respondent’s bad faith, clearly indicating an intention to imitate and exploit Complainant’s established online presence.

4. Use of Complainant’s device mark on Respondent’s website

Another argument of Complainant is the use of the Complainant’s trademark on the website, combined with the presentation of services that appear to replicate or relate to Complainant’s QWEN AI models: this confirms that Respondent was clearly aware of Complainant’s business and is intentionally seeking to benefit from its goodwill. Furthermore, the website is presented in English, which indicates an intent to target an international audience and to exploit the global recognition of the Complainant’s QWEN brand. The website also includes a dedicated language-selection section, offering multiple interface languages that are official languages in countries in which the Complainant’s trademark is registered and where the QWEN product series is widely recognized. This feature shows a clear intention to target users in markets where Complainant’s brand enjoys strong visibility and commercial value.

5. Similar UDRP with the same Respondent

A recent precedent is Mistral AI v. Domain Administrator, WIPO Case No. D2025-113. In that proceeding, the Respondent was the same Respondent involved in the present case. The Complainant was again a leading European artificial intelligence company specialized in large language models (LLMs). The fact that Respondent has engaged in the same conduct against another major LLM provider demonstrates a clear pattern of abusive behavior within the meaning of paragraph 3.1.2 of the WIPO Overview. This recidivism significantly reinforces the conclusion that, in the present case as well, Respondent intentionally targeted a well-known AI brand and registered the disputed domain name in bad faith. The decision in this case has been attached to the Complaint.

6. Complainant’s QWEN has gained visibility through trademark applications

Respondent’s registration of the disputed domain name occurred at a time when Complainant’s QWEN brand had already gained visibility through formal trademark filings and sustained public promotion. By early 2024, Complainant had taken concrete steps to secure its brand identity, including multiple trademark applications for “QWEN” and “TONGYI QWEN”, as well as the earlier filing and registration of “TONGYI”, a mark closely tied to the QWEN project. It cannot be reasonably viewed as coincidental. Instead, it reflects a calculated attempt to benefit from Complainant’s growing reputation and to mislead users by appropriating a brand that had already achieved recognition in the relevant commercial and technological landscape.

7. No response from Respondent in cease and desist procedure

Before starting this UDRP procedure Complainant had instructed its authorized representative to notify Respondent of the infringement and to request the immediate cessation of the unauthorized use and the transfer of the disputed domain name. However, all contact attempts proved unsuccessful:

- The e-mail address listed in the Registrar’s WHOIS for the disputed domain name was found to be non‑functional;

- The notification sent to it was returned as undeliverable.

For this reason, the Complainant submitted a Whois Inaccuracy Report to the Registrar.

An additional e-mail address identified on Respondent’s website likewise proved to be non‑operational, with the message sent to it also returning as undeliverable.

Despite these efforts, Respondent provided no means of communication and failed to respond to any attempt to establish contact.

Respondent’s failure to provide valid contact information and its complete lack of response further reinforce the conclusion that the disputed domain name was registered and is being used in bad faith, as such conduct demonstrates an intention to avoid accountability and to disregard the Complainant’s trademark infringement concerns.

Complainant’s conclusion

Complainant concludes that all above taken together, these circumstances clearly demonstrate that Respondent registered and is using the disputed domain name in bad faith, with the intent to exploit Complainant’s brand recognition and goodwill. The factual context leaves no doubt that Respondent was aware of Complainant’s activities and deliberately sought to benefit from its commercial success.

Panel findings

Based on the above the Panel has the following findings.

Bad faith circumstances occur mainly when a Respondent sees financial gain on riding on the coat tails of Complainant’s success. This is mostly profitable when Complainant has a reputation that Respondent should have been aware of. Therefore, first it needs to be established whether Complainant had such a reputation on the moment of the registration of the disputed domain name.

Complainant alleges that it already had been publicly known, at least in the circles of AI language models experts, since and immediately after launch, which was in 2023. It substantiates this with a list of links.

Within the light of a UDRP procedure which is a quick procedure and where the facts and arguments should be straightforward and therefore be displayed in an easy and clear manner the Panel founds this substantiation insufficient:

• The Panel is forced to open links, click through many documents and investigate them;
• The origin and relevance of the links is not explained;
• Many times, the links are in Chinese;
• Many times, these links are not dated or dated after registration of the disputed domain name;
• Many of the links originate from Complainant’s own sources rather than from objective third party sources;
• More specifically, the photos of conferences and fairs do not have objective dates;
• The referral to the APSARA conference, a conference for AI language experts, should have been further explained why it is relevant as it is not a commonly known conference for the public at large to which the Panel belongs;
• Lastly, the VPN search showing that Complainant’s QWEN is the most common result in Respondent’s location when conducting a search date of 2026 which is well after the registration of the disputed domain name in early 2025.

Further, it should be noted that the registration of the disputed domain name was very quickly after the launch of QWEN by Complainant and therefore it should be accurately explained that the trademark had already have a reputation in the specialized circles of language AI models.

Further, Complainant’s argument that Respondent should have known of the trademark applications for QWEN as indication of bad faith cannot be a standing alone argument as there is no indication that it is a common practice to check trademark registers before registering a domain name.

Nevertheless, there are arguments that the Panel finds convincing and straightforward with respect to the allegation that Respondent knew about Complainant’s use of QWEN:

• The argument that the trademark QWEN is distinctive and that it is not likely that two parties create the same trademark for the same purpose;
• The argument that the composition of the domain name of Complainant and that of the disputed domain name is the same;
• The fact as shown that Respondent has used the device mark of Complainant on its website;
• The argument that a similar UDRP case about the use of a domain name containing a trademark of a company active in language AI services and where the disputed domain name was registered by the same Respondent had already been decided;
• The fact that Respondent was not revealed in the whois, neither answered on the cease and desist letter of Complainant, nor answered in this Complaint.

The combination of those arguments establishes sufficiently that use and registration of the disputed domain name is in bad faith.

Consequently, the Panel concludes that Complainant has shown the disputed domain name has been registered and is being used in bad faith (within the meaning of paragraph 4(a)(iii) of the Policy).

The Panel is satisfied that all procedural requirements under UDRP were met and there is no other reason why it would be inappropriate to provide a decision.