Summary

  • The 1955 agreement, 1970 agreement and 1975 protocol allocated and refined uses of the Merck name between two corporate groups. They constrained those parties and supplied evidence in later proceedings, but they did not bind ICANN to divide one top-level label by territory or empower an objection expert to transfer an application, execute a Registry Agreement or change the root.
  • Five applications had to be kept procedurally separate: three exact-string .MERCK applications, Merck KGaA’s .EMERCK application and MSD Registry Holdings, Inc.’s .MERCKMSD application. The objection record, Reconsideration Request 14-9, the IRP declaration and the two CPE reports show distinct standards and remedies: no one route supplied a general merits appeal or an automatic award of the string.
  • The executable .MERCK outcome arose only after the exact-string contention was resolved, the two community applications were withdrawn, Merck Registry Holdings, Inc.’s standard application survived, ICANN signed the Registry Agreement on 10 September 2024 and IANA completed delegation processing in April 2026. Those records identify the external operator and sponsoring organisation; they do not disclose the consideration, releases, veto rights, deadlock rules or other private terms behind the withdrawals and the parties’ later cooperative description.

One label, three official records and three different powers

The most reliable entry point into the .MERCK case is not the first objection. It is the apparent disagreement among three later official records. ICANN’s application page for Merck Registry Holdings, Inc.’s standard application identifies application 1-1702-28003, records that it prevailed in contention, links a Registry Agreement and still displays the transitional label “In PDT”. The page itself warns that information for TLDs that have contracted with ICANN may no longer be current because the application portal is not maintained as the authoritative registry record after delegation.

ICANN’s separate .MERCK Registry Agreement page identifies Merck Registry Holdings, Inc. as the registry operator and gives 10 September 2024 as the agreement date. IANA’s delegation report dated 17 April 2026 then records the proposed manager as the same company, notes that the new-gTLD process had been completed, confirms that the sponsoring organisation matched ICANN’s contracted party, and records completion of the eligibility, contact and technical checks. The current IANA delegation record likewise names Merck Registry Holdings, Inc. as sponsoring organisation.

Those sources are not alternative versions of one status. They are records of different institutional acts. “Prevailed Contention” describes an application-programme result: one application remained able to proceed after rivalry for an identical string was resolved. A Registry Agreement is a bilateral legal instrument: it makes the named operator subject to enforceable obligations owed to ICANN. Delegation is a root-zone act: once the contractual and technical prerequisites have been satisfied, the sponsoring organisation and authoritative name-server data are entered into the global DNS.

An objection determination could stop an application from reaching the first of those later stages. It could not itself sign the contract or complete the root-zone change.

This separation is essential because application survival is often narrated as though it were ownership of a TLD. It is not. A successful applicant still needs contracting, pre-delegation processing and delegation. Conversely, a later contract or IANA report can supersede a stale application-page label for the particular question that the later record answers. Transparency means that all three records remain visible. Accountability requires identifying which institution had authority over which legal object and which later evidence proves implementation.

The same distinction explains why private territorial rights could not function as root-zone instructions. The DNS root requires a single, globally unique delegation for the ASCII label .MERCK. It has no native facility for returning one sponsoring organisation to a user in the United States or Canada and another sponsoring organisation to a user elsewhere. Websites beneath a TLD may use geolocation, access rules or content routing, but those are operational choices below the delegation layer. They do not divide the top-level label itself.

The Merck groups could participate throughout the programme by filing applications, submitting evidence, objecting, requesting reconsideration, commencing an Independent Review Process, seeking Community Priority Evaluation, negotiating and withdrawing. None of those participatory rights carried a vote over the result at another gate. WIPO and ICC experts decided objections within their mandates. ICANN accountability bodies applied their own jurisdictional tests. The evaluation provider scored CPE. ICANN controlled programme processing and contracting. IANA/PTI processed delegation after the required conditions were met.

The central governance problem was therefore not simply which party had the stronger historical claim. It was which institution could convert a claim into the next legally effective status.

The coexistence instruments governed parties, not the root

The conflict began long before the 2012 new-gTLD round. The two corporate groups share a history but became independent, and each developed legitimate commercial and trade-mark interests in the Merck name. The English Court of Appeal’s approved judgment describes a succession of arrangements intended to regulate that parallel legitimacy. The case file should not compress those arrangements into one timeless “coexistence agreement”, because their dates, functions and details differed.

The 1955 agreement addressed the use of Merck as a trade mark, corporate name, firm name or business name across the parties’ traditional territories. The court record describes a central territorial allocation: the US group’s position in the United States and Canada and the German group’s position outside those territories, together with rules for geographically qualified corporate uses. It also records the parties’ insistence on distinctions between use as a mark and use as a firm or corporate name. Those distinctions mattered to the contracting parties and later courts.

They were not classifications that the root could enforce merely by assigning a string.

The 1970 agreement restated and developed the settlement after corporate-name changes and further disputes. Its provisions continued to regulate the use of the Merck sign and the conditions under which qualified names could cross the territorial boundary. The agreement was forward-looking in the sense that it was intended to govern future conduct, but that did not turn it into a public allocation instrument. It remained a private contract whose interpretation and enforcement depended on an appropriate court, arbitral forum or other body with jurisdiction over the parties and the dispute.

The 1975 protocol addressed practical problems that had emerged under the 1970 arrangement. The court record identifies, among other matters, domestic stationery sent abroad, visiting cards used by travelling representatives and the prominence of geographic identifiers. These details reveal the environment for which the arrangements were designed. Location could often be inferred from the office, document, journey, market or publication in which the name appeared. The protocol refined party conduct; it did not assign any technical identifier.

Internet use placed pressure on that territorial architecture. A website can be reached from places that its operator did not treat as the primary market. A global corporate page can simultaneously address scientists, recruits, investors, suppliers and patients in several jurisdictions. Litigation therefore required courts to ask where online conduct was directed, what law governed, how prior findings affected later proceedings and what particular relief was available. The Singapore Court of Appeal’s 2021 judgment, for example, examined the effect of earlier English findings without treating them as a universal resolution of every issue under Singapore law.

A top-level domain sharpened the mismatch. A territorial court can interpret a contract, enjoin conduct, award relief between parties or recognise a foreign judgment. A new-gTLD objection expert received a narrower programme question. The expert could decide whether the objection met the applicable standard and, through that determination, whether the challenged application survived that gate. The expert could not write a territorial schedule into the root, create two delegations for one label, force a licence, transfer the application to the objector or make ICANN execute a contract.

This is why it is inaccurate to say that the old agreements had already decided who “owned” .MERCK. They allocated specified uses of a name and created rights and constraints between private parties. They could be evidence of existing legal rights, bona fide use, territorial expectations, intent and likely conflict. Yet the new-gTLD programme separated several questions: standing, timeliness, the merits of an objection, priority, contention resolution, contracting and delegation. The agreements could affect some of those questions. They did not erase the institutional boundaries among them.

Five applications, not one abstract contest over a name

The programme record contained five applications with different applicants, strings and procedural consequences:

Applied-for string Applicant and application Category Procedural position
.MERCK Merck KGaA, 1-980-7217 Community Exact-string contention; CPE did not prevail; later withdrawn
.MERCK Merck Registry Holdings, Inc., 1-1702-28003 Standard Exact-string contention; later prevailed; contracted and delegated
.MERCK Merck Registry Holdings, Inc., 1-1702-73085 Community Exact-string contention; CPE did not prevail; later withdrawn
.EMERCK Merck KGaA, 1-980-60636 Standard Separate string; objections failed; later delegated
.MERCKMSD MSD Registry Holdings, Inc., 1-1704-28482 Standard Separate string; objections failed; later contracted and delegated

The three .MERCK applications formed an exact-string contention set because only one identical top-level label could be delegated. The alternate strings did not automatically resolve that competition. .EMERCK gave Merck KGaA a separate application capable of continuing on its own record. .MERCKMSD did the same for MSD Registry Holdings, Inc. Neither alternative compelled its applicant to abandon .MERCK, and the existence of one did not award the other side the exact string.

The “community” designation also requires precision. It was an application category with commitments and a potential route to priority through Community Priority Evaluation. It was not a declaration of superior trade-mark title. Nor did it guarantee that the applicant constituted or represented a qualifying community for every other programme mechanism. Community-objection standing under Module 3 and CPE scoring under Module 4 used different tests, different decision-makers and different remedies. Treating the shared word “community” as one doctrine obscures the case.

The official objection-determination index records proceedings involving Legal Rights Objections, Community Objections and a String Confusion Objection. Each proceeding addressed only the application and objection before it. A decision concerning .EMERCK did not allocate .MERCK. A community-objection determination did not supply a CPE score. An LRO determination did not execute a Registry Agreement. The five-application map must remain intact for the later chronology to make sense.

Merck KGaA’s community .MERCK application also survived a Legal Rights Objection brought by Merck & Co., Inc.. The expert accepted an adequate basis for standing but rejected the objection, treating both sides as bona fide users in different territories and leaving later infringement or contract claims to ordinary legal processes. That result kept application 1-980-7217 in the exact-string contention set. It did not confer CPE priority or displace the two Merck Registry Holdings applications.

What the .EMERCK objection victories did—and did not—produce

The .EMERCK proceedings are a useful control because Merck KGaA’s application survived two different objections brought by the US group. In the WIPO Legal Rights Objection determination, the expert found an adequate basis for the objector’s standing but rejected the objection on the merits. The determination treated both groups as bona fide users of Merck-related marks in different territories. It did not accept the proposition that an applicant with legitimate rights in some countries must hold rights in every country before it may obtain a gTLD.

The expert also kept the remedy within the LRO proceeding. He declined to use the objection as a forum for a comprehensive interpretation of every coexistence obligation or every possible future use. The determination left ordinary legal proceedings available if operation of the TLD later infringed the objector’s rights. Rejection therefore meant that this objection did not eliminate .EMERCK. It did not declare all future uses lawful, rewrite the coexistence instruments or grant Merck KGaA a registry.

A separate String Confusion Objection determination also allowed the application to continue. That expert asked whether confusion between the applied-for string and the objector’s string was probable for the relevant internet user under the string-confusion standard. The initial “E”, visual and phonetic features, and the overall comparison were assessed for that limited purpose. Questions about trade-mark infringement or entitlement belonged to other processes. The objection was dismissed because the required probability of string confusion was not established.

Two failed objections removed two barriers; they did not complete the application. Contracting and delegation still had to follow. The later IANA record for .EMERCK identifies Merck KGaA as sponsoring organisation, gives 11 September 2014 as the registration date and links the delegation report. The legally significant chain is therefore objection survival followed by the later programme, contractual and root-zone steps. The existence of .EMERCK proves that those steps were completed. It does not prove a final interpretation of the coexistence instruments for every digital use.

This negative-gate logic recurs throughout the case. When an objection fails, the application is no longer stopped by that objection. The result is not a positive award of universal rights. Confusing those propositions inflates the authority of an expert determination and makes later contracting appear automatic when it was not.

WIPO declined to convert territorial legitimacy into a global veto

Merck KGaA brought Legal Rights Objections against all three US-group applications: the standard .MERCK application, the community .MERCK application and .MERCKMSD. The objections were rejected.

The expert did not deny that Merck KGaA possessed relevant rights or standing. The Applicant Guidebook required him to ask whether the potential use of the applied-for string would take unfair advantage of the distinctive character or reputation of the objector’s mark, unjustifiably impair it, or otherwise create an impermissible likelihood of confusion. For a trade-mark-based objection, the Guidebook supplied eight non-exclusive factors.

Those included the similarity of the string and mark, bona fide acquisition and use, the applicant’s intent, the applicant’s own rights, whether the applicant was commonly known by the sign and the likelihood of confusion as to source, sponsorship, affiliation or endorsement.

The determinations treated the two groups as bona fide rights holders whose strongest positions were territorially divided. The expert reasoned that legitimate rights in fewer than all countries did not, without more, make an application illegitimate. Otherwise, a trade-mark owner seeking a gTLD could effectively be required to possess registrations everywhere. He also declined to decide the whole contractual relationship. Future use that infringed rights could still be challenged in ordinary legal proceedings.

That result can be criticised. A globally visible top-level label may have practical effects outside an applicant’s traditional territory. A decision not to stop the application leaves those risks to later operating choices and litigation. But the institutional holding must remain bounded. The expert did not allocate .MERCK by country, order geo-targeting, impose a licence, award damages, transfer any application or direct ICANN to sign a contract. Rejection of the objections allowed the applications to continue. It did not immunise every later use.

The determinations’ references to the Uniform Domain Name Dispute Resolution Policy became part of the later accountability challenge. The LRO standard came from the Applicant Guidebook, not the UDRP, and the mechanisms address different disputes. The expert nevertheless referred to UDRP material as guidance when considering aspects of the comparison between sign and string. The question for reconsideration would not be whether that was the best possible trade-mark analysis. It would be whether the expert or ICANN staff had contradicted an established ICANN policy or process.

A more immediate problem was factual. The determinations included a statement attributing to the US-group applicants a commitment to use geo-targeting to prevent access from territories associated with Merck KGaA. Merck KGaA said that attribution was wrong: it had described its own geo-targeting practices, while the US group had not made the stated commitment. The expert issued an addendum acknowledging that the sentence had been included inadvertently, maintaining that he understood the parties’ actual positions and saying the sentence was immaterial to the conclusion.

The correction did not create an internal LRO appeal. That absence matters. A full merits-appeal system might have required a new panel to reconsider the eight factors after a demonstrated attribution error. A system concerned only with finality might have left the sentence untouched. ICANN’s structure occupied a narrower middle ground: reconsideration was available, but only if the requester could connect the challenged action to the criteria in the then-applicable Bylaws. The classification of the mistake—as a merits dispute or as a breach of established process—therefore controlled the available remedy.

ICC community objections turned first on standing

Merck KGaA also filed Community Objections against the two US-group .MERCK applications and the .MERCKMSD application. The ICC determination concerning the two .MERCK applications and the separate .MERCKMSD determination asked whether the objector was an established institution with an ongoing relationship to a clearly delineated community strongly associated with the applied-for string.

The expert accepted that Merck KGaA was an established institution and a substantial, recognised corporate group. The dispositive problem was the claimed Merck community. The evidence described subsidiaries and affiliated entities linked through ownership, control, branding and corporate governance. The expert concluded that the asserted boundaries were principally boundaries of ownership rather than membership or association of the kind required by the standing test. Merck KGaA therefore had not established the necessary relationship with a clearly delineated community distinct from the institution itself.

That standing conclusion prevented the objection from producing the requested result. The determinations contained further discussion, but none of it converted the proceeding into an allocation of the name or a CPE decision. The expert expressly kept Community Priority Evaluation outside the mission of the objection panel. The same corporate facts could be evaluated differently under CPE because the legal object and remedy were different.

The reverse-direction community objections failed at an earlier gate. Merck & Co., Inc. attempted to file community objections to Merck KGaA’s .MERCK and .EMERCK applications, but the filings reached the ICC ten and eleven minutes after the published deadline. The ICC rejected them as untimely. Merck & Co. sought rehearing after ICANN adopted a general resolution permitting dispute-resolution providers to use discretion over deadlines in appropriate circumstances.

Reconsideration Request 13-8 was filed by Merck KGaA, not by the late objector. Merck KGaA challenged the ICANN resolution that had supplied general discretion and argued, among other matters, that the decision-making record was incomplete and unfair. The Board Governance Committee recommended denial. It treated the resolution as a programme-wide grant of discretion rather than a command that the ICC reopen this particular filing, and found no proper basis for reconsideration under the then-Bylaws.

The record supports a limited conclusion. The two objections were not admitted because they were late; the general discretion resolution did not itself reverse that outcome; and Merck KGaA’s challenge to the resolution did not succeed. The record does not show what the ICC would have decided on the merits had either objection been accepted. Timeliness controlled access to the forum. It did not adjudicate the coexistence rights.

Reconsideration Request 14-9 could not become the missing LRO appeal

Merck KGaA’s Reconsideration Request 14-9 was the central test of the accountability architecture. The requested relief was direct: reject the three adverse WIPO determinations, convene a new panel and obtain new decisions applying the proper standards. Merck KGaA argued that the expert had improperly imported UDRP standards and had relied on the incorrect geo-targeting attribution.

The Board Governance Committee described a narrower jurisdiction. Under the then-applicable Bylaws, reconsideration of staff action or inaction could address contradiction of established ICANN policy or procedure. In the new-gTLD programme, a third-party expert determination could be challenged through reconsideration if the panel had failed to follow established policy or process, or if ICANN staff had failed to follow policy or process when accepting the determination. The BGC expressly said it would not conduct substantive review of the expert’s conclusion on the LRO merits.

On the UDRP issue, the BGC found that the expert had addressed the Guidebook’s eight non-exclusive factors and had used UDRP material as context rather than replacing the LRO standard. The dispute-resolution procedure allowed an expert to refer to rules or principles considered applicable. Merck KGaA’s disagreement with that reasoning therefore did not identify an established policy or process that had been violated.

On geo-targeting, the BGC accepted that the challenged statement was wrong. It relied, however, on the expert’s addendum: the expert acknowledged the inadvertent inclusion, said he had understood which party had described geo-targeting and maintained that the sentence was immaterial. Merck KGaA did not identify a Guidebook rule requiring the expert to reopen the record, reconvene the proceeding or repeat the complete factor analysis after the addendum. The BGC treated the remaining dispute as one about the expert’s assessment of materiality.

The request was denied. Because the BGC treated the matter as a challenge to staff action or inaction, its determination was final under the then-applicable reconsideration provisions without further Board consideration. The effect was practical and narrow. Reconsideration did not erase the admitted factual error, but it also did not supply the replacement panel Merck KGaA had requested. The original determinations, read with the addendum, remained in force, and the three applications remained alive.

This outcome illustrates the difference between review access and an enforceable remedy. Merck KGaA obtained a filed request, a published analysis and reasons. It did not obtain a fresh LRO merits decision. Limited review protected programme finality and the assigned jurisdiction of the first-instance expert. It also meant that a factual error could remain non-dispositive because the original decision-maker characterised it as immaterial and the requester could not connect the correction process to a breached rule.

The legitimacy question is therefore more difficult than whether “review was available”. Review can expose an error without changing status. It can test procedural compliance without supplying the remedy requested. In a sequential system, the important inquiry is what the review body can order and which later institution must implement that order. Reconsideration 14-9 had no jurisdiction to award the string, sign a contract or alter the root, and the BGC found no trigger for remitting the objections to a new expert.

The IRP reviewed ICANN’s conduct, not the trade-mark dispute anew

Merck KGaA then commenced an Independent Review Process against ICANN. The final declaration considered whether ICANN’s conduct, including the BGC’s treatment of Reconsideration Request 14-9, complied with the applicable Articles, Bylaws and programme commitments. It did not give the panel unrestricted authority to decide which group should receive .MERCK.

The panel rejected an attempt to use the IRP as the merits appeal absent from the LRO procedure. Its task was to assess the relevant Board conduct against the governing instruments and the applicable standard of review. It was not empowered to second-guess the expert’s substantive conclusion merely because it might have weighed the trade-mark evidence differently. The declaration examined whether the BGC had exercised due diligence and care with an adequate factual record, not whether the panel itself would have reached the same LRO result.

Merck KGaA also argued that ICANN had treated comparable cases differently and had an improper financial interest connected to a possible auction. The panel rejected those claims. It found that different cases could justify different choices among the options available to ICANN and that difference alone did not establish prohibited discrimination. It also regarded the auction-interest allegation as speculative and unsupported on the record before it.

The final declaration’s operative section stated that Merck KGaA had not succeeded, identified ICANN as the prevailing party and required Merck KGaA to pay ICANN US$48,588.54 in costs. That was a cost allocation, not damages for trade-mark infringement and not a substantive award concerning the Merck name. On 3 February 2016, the ICANN Board accepted the declaration’s findings.

There was no remand for a new WIPO panel, no reversal of the LRO determinations and no root-zone consequence. Merck KGaA obtained access to an accountability forum and a reasoned declaration. It did not obtain the application-status change it sought. The only directly quantified order in the declaration ran against Merck KGaA for ICANN’s costs.

The distinction does not mean the IRP was meaningless. It tested ICANN’s fidelity to its governing instruments and produced a public explanation of the limits of that review. But institutional scrutiny and merits correction are not interchangeable. Where the applicable process omits a substantive appeal, an accountability mechanism cannot be assumed to create one simply because the underlying decision is consequential.

Community Priority Evaluation was a separate route to priority

After the objections and the IRP, the exact .MERCK contention set still contained three applications. Each community applicant could seek priority through Community Priority Evaluation. Under the Guidebook, an application needed at least 14 of 16 points across community establishment, nexus between the string and community, registration policies and community endorsement. CPE was not a rehearing of the WIPO or ICC proceedings. It asked whether a community application met the separate priority criteria.

The CPE report for Merck KGaA’s community application, dated 10 August 2016, awarded 11 points. It gave full credit for community establishment and registration policies, three points for endorsement and zero for nexus. The evaluator reasoned that the string .MERCK also identified the substantial US-based Merck group outside the community defined by the application. The string therefore substantially over-reached that community. The report said the application did not prevail and could still resolve contention through the other methods in Module 4.

The CPE report for Merck Registry Holdings, Inc.’s community application reached a related but not identical outcome. It awarded 9 points and also gave zero for nexus because the string identified Merck KGaA outside the applicant’s defined community. It made additional deductions under registration policies. Neither score was a holding that the groups possessed equal trade-mark rights everywhere. The shared result was narrower: neither applicant could define its own corporate community and also show that the globally shared string uniquely or sufficiently identified that community under the CPE criteria.

The symmetry exposes the structural problem. Each group could describe a coherent corporate network. Yet .MERCK was also the name by which the other group was known in its principal territory. A community definition narrow enough to be administratively clear did not make the string exclusive to that community. CPE therefore did not translate territorial coexistence into priority.

Merck KGaA challenged its report in Reconsideration Request 16-12. While that request was pending, ICANN undertook a broader review of the CPE process. On 27 January 2019, the Board denied Request 16-12. It found no established policy or procedure requiring the evaluator to award nexus points because of geo-targeting, coexistence litigation or the applicant’s territorial rights. It also noted that the same over-reach analysis had been applied to both Merck community applications.

The consequence was not an automatic victory for the standard application. The special priority route had failed for both community applicants. Ordinary contention remained. It still had to be resolved through withdrawal, private agreement or the auction of last resort. Losing CPE returned an application to the contention problem; it did not select the eventual operator.

The auction backstop preserved ICANN’s control of the timetable

By 2019, applicants were reporting efforts to resolve the .MERCK contention privately, while ICANN maintained the auction of last resort as the programme backstop. Merck KGaA and Merck Registry Holdings, Inc. jointly sought a second postponement of the scheduled auction. After staff refused, they filed Reconsideration Request 19-4, arguing that the decision had not properly accounted for the litigation history and settlement progress and was inconsistent with the programme’s preference for private resolution.

On 26 January 2020, the Board denied Request 19-4. It accepted the recommendation that staff had considered the relevant material and had not violated established policy by refusing a second postponement. At the same time, the Board directed the organisation to obtain an update and allowed limited discretion if the applicants jointly demonstrated that a private resolution was very close. Otherwise, processing, including auction scheduling, was to continue.

That disposition preserved two principles. Applicants could negotiate and present evidence of progress. They could not acquire indefinite control of the programme calendar merely because both preferred more time. ICANN’s preference for private resolution did not create an unlimited entitlement to extensions. Nor did ICANN’s residual ability to grant a short accommodation transfer scheduling power to the parties.

The later public application pages show Merck KGaA’s community application and Merck Registry Holdings, Inc.’s community application as withdrawn. They show the standard Merck Registry Holdings application as having prevailed in contention. These records are sufficient to establish the programme outcome. They do not disclose the withdrawal instruments, the dates on which private obligations became effective, the consideration exchanged or the reasons each applicant agreed to withdraw.

The evidence therefore does not require a narrative in which an auction selected the operator. It supports a more cautious account: the auction remained a credible backstop; the two community applications were withdrawn; the standard application survived. Whether particular private terms caused each withdrawal, and what concessions accompanied them, remains outside the public record reviewed here.

Contract and delegation converted survival into an executable namespace

The alternate strings reached later stages while .MERCK remained contested. The .EMERCK IANA record identifies Merck KGaA as sponsoring organisation and records registration in September 2014. ICANN’s .MERCKMSD Registry Agreement page names MSD Registry Holdings, Inc. as operator under an agreement dated 14 July 2016. The .MERCKMSD IANA record records registration in June 2017. These are separate outcomes for separate strings, not partial awards of .MERCK.

For the exact string, the public chain is now clear. The two community application pages show withdrawal. The standard application page records that Merck Registry Holdings, Inc. prevailed in contention. ICANN’s Registry Agreement record names that company as operator and dates the agreement to 10 September 2024. IANA’s 17 April 2026 report names the same company as proposed manager, confirms the match with ICANN’s contracted party and records completion of the required checks. The root database now lists Merck Registry Holdings, Inc. as sponsoring organisation.

These records establish external legal and operational authority. The contracted operator owes the Registry Agreement obligations to ICANN. The sponsoring organisation is the entity recorded for the root-zone delegation. Neither status should be replaced by an inference from corporate publicity.

The operator’s public “About” page describes a cooperative arrangement reached in 2024 and says that the namespace operates under joint control through MM Domain Holdco Ltd., characterised as a 50/50 joint venture between the two Merck groups. That is relevant evidence of how the operator publicly presents the arrangement, but it is not the underlying settlement or shareholder agreement. The UK Companies House entry confirms that MM Domain Holdco Limited is an active company incorporated on 25 November 2020. Merck KGaA’s 2023 list of shareholdings records a 50 per cent interest and the annual report treats the company as equity-accounted.

Together, those sources support an attributed statement that a corporate vehicle exists and that Merck KGaA reported a half interest. They do not establish the content of the private governance instrument. The public record does not disclose consideration, releases, board composition, reserved matters, veto rights, deadlock procedures, policy-allocation rules, emergency authority or the circumstances of every withdrawal. It also does not show that MM Domain Holdco Limited replaced Merck Registry Holdings, Inc. as ICANN’s contract counterparty or IANA’s sponsoring organisation.

That distinction matters because “joint control” can describe several different arrangements. It may mean equal share ownership, unanimity for specified decisions, consultation rights, board parity or a veto over reserved matters. It does not necessarily mean that either shareholder can bind the registry operator, instruct its technical provider or answer an ICANN compliance notice. Without the operative documents, those questions remain open.

The interval between contract execution in September 2024 and the delegation report in April 2026 is similarly bounded. The IANA report proves that the new-gTLD process and required checks were completed by the time of delegation. Public sources reviewed for this article do not identify a single cause for the length of the interval. Duration alone does not prove non-compliance, regulatory delay or internal deadlock.

The final namespace was therefore not created by one decisive adjudication. The coexistence instruments constrained the parties. WIPO and ICC experts controlled objections under assigned standards. Reconsideration and IRP bodies controlled bounded accountability questions. The CPE evaluator controlled priority scoring. Applicants controlled whether to settle or withdraw, subject to programme rules. ICANN controlled processing and the decision to enter a Registry Agreement. IANA/PTI controlled delegation processing after the prerequisites were satisfied.

Each gate was consequential, but no earlier institution possessed the powers of the later one.

Two counterfactuals make the remedy boundary visible

The first counterfactual is a territorially programmable root. Suppose the root could delegate .MERCK to Merck Registry Holdings, Inc. for users in the United States and Canada and to Merck KGaA elsewhere. The coexistence instruments might then have supplied a starting allocation map, subject to difficult interpretation. Travellers, virtual private networks, multinational organisations, conflicting geolocation data and cross-border content would still create disputes. Yet the delegation layer could at least attempt to mirror territory.

The actual root cannot do that. It records one globally unique label and one sponsoring organisation. Geo-targeting below the root can regulate access or presentation, but it does not create two top-level delegations. Because the technical object was indivisible, the programme ultimately required one application to survive for .MERCK.

The second counterfactual is a substantive appeal from LRO determinations. Under that design, a demonstrated factual misattribution might have triggered a rehearing. A reviewing body could have reweighed the Guidebook factors, the UDRP references, the coexistence evidence and the stated materiality of the error. ICANN’s reconsideration and IRP mechanisms did not provide that jurisdiction in this case. Merck KGaA had to establish the relevant policy, process, Articles or Bylaws violation. The BGC and IRP panel found no basis for the remedy requested.

Those counterfactuals prevent two overstatements. The private agreements were not irrelevant; they shaped rights, evidence, litigation and settlement incentives. But they did not allocate the root. The accountability proceedings were not empty; they supplied scrutiny and reasons. But they did not constitute a general merits appeal or produce a new LRO panel.

The case therefore ends with a sequential rather than heroic explanation. Merck Registry Holdings, Inc. did not become operator because one forum declared it the universal owner of the name. Its standard application remained after objections, accountability proceedings, failed CPE attempts and withdrawals; ICANN then entered a contract with it; and IANA completed the root-zone process. In a system of non-substitutable gates, a persuasive claim matters only when presented to an institution that has both jurisdiction to decide it and authority to produce the next executable status.