Summary

  • Nu Dotco LLC won ICANN’s July 2016 auction of last resort for .WEB with a US$135 million bid financed by Verisign under an agreement contemplating a later assignment of the registry agreement, subject to ICANN’s consent. The auction removed the six other .WEB applications from the path to contracting. It did not decide whether the financing-and-assignment structure complied with every part of the 2012 Applicant Guidebook, nor did it create an automatic right to a registry contract or delegation.
  • The first Independent Review Process found that ICANN had violated its Articles and Bylaws by failing to decide the complaint about that agreement before moving towards contracting. The panel did not find the agreement unlawful, disqualify Nu Dotco, award .WEB to Afilias or set a replacement auction price. It made constitutional declarations, protected the status quo, recommended a first-instance Board decision and imposed specified cost reimbursements. Afilias remained the prevailing party on those core declarations, although the panel’s 21 December 2021 ruling denied its post-declaration application as frivolous and ordered it to reimburse ICANN US$236,884.39 in legal fees.
  • In April 2023, following a Board Accountability Mechanisms Committee review, the ICANN Board found that Nu Dotco had not breached the Guidebook or Auction Rules and directed processing to continue. Altanovo, the renamed Afilias applicant, challenged that determination in a second IRP but withdrew the case on 22 July 2026. No public merits declaration reviewed the Board’s interpretation.
  • The IANA delegation record was updated on 23 July 2026 and names VeriSign, Inc. as .WEB’s sponsoring organisation. That record establishes the operational endpoint. It does not disclose the confidential resolution of the disputes or every contractual act by which Verisign became the registry operator.

The root-zone line that the auction could not write

On 23 July 2026, the IANA delegation record for .WEB displayed the result that the US$135 million auction had failed to produce for almost ten years. It identified .WEB as a generic top-level domain, named VeriSign, Inc. as the sponsoring organisation, listed the registry name servers and Registration Data Access Protocol service, and recorded that the entry had been updated that day.

The page is an operational record, not an adjudication. It does not recount the auction, interpret the Applicant Guidebook, describe the Domain Acquisition Agreement between Nu Dotco and Verisign, or explain why a pending accountability case ended. Its narrowness is precisely what makes it useful. The public root record answers a final technical question—who is recorded as the sponsor of the delegated top-level domain—without pretending to answer the legal and institutional questions that delayed that result.

The .WEB controversy was repeatedly described as a dispute over who had “won” the string. That shorthand collapsed several decisions into one. An auction could determine which application survived contention. ICANN still had to decide whether the applicant and its application complied with programme rules, complete contracting, address any proposed assignment and satisfy the conditions for delegation. Root-zone implementation then had to make the authoritative DNS change. These powers were connected, but they were not interchangeable.

Verisign’s 22 July 2026 delegation announcement added commercial context while preserving an important evidentiary limit. The company said .WEB had been delegated into the global DNS root with Verisign as the designated registry operator, that prior disputes had been resolved and that the details were confidential. It also said the top-level domain would be governed by ICANN’s Base Registry Agreement and that registrations would be offered through channel partners later in 2026. Delegation therefore did not mean that public registrations had already opened, and the announcement did not reveal the bargain or contractual sequence that placed Verisign, rather than Nu Dotco, at the endpoint.

The timing permits chronology, not certainty about causation. ICANN’s docket for Altanovo’s second .WEB IRP records that the claimant withdrew its request for independent review on 22 July 2026. Verisign announced delegation that day; IANA recorded an update the next day. The public sources do not establish that any party paid a particular settlement amount, admitted wrongdoing, abandoned a legal position in exchange for a specified concession or accepted a particular contractual structure.

Withdrawal ended the pending review. It did not produce a panel decision approving the ICANN Board’s reasoning. Delegation made the allocation operational. It did not retrospectively convert every disputed step into an adjudicated one. The IANA record authoritatively records the sponsoring organisation in the root database; it does not disclose the pathway by which the parties and ICANN reached that result. The difference between a public endpoint and an undisclosed route is central to the case.

What the US$135 million auction actually selected

Seven entities applied to operate .WEB in the 2012 new-gTLD round. Because the applicants did not privately resolve their contention, the string proceeded to ICANN’s mechanism of last resort. The official auction announcement records that Power Auctions LLC conducted the auction on 27 July 2016 and that Nu Dotco LLC prevailed at US$135 million. ICANN’s later Board records describe the event as taking place over 27 and 28 July and identify Afilias Domains No. 3 Limited, whose parent at the time was Afilias, Inc. and which was later renamed Altanovo Domains Limited, as the second-highest .WEB bidder.

The auction sat within an indirect contention set that also included one application for .WEBS. That application survived separately at a nominal price. For .WEB, however, Nu Dotco’s bid removed six directly competing .WEB applications from the route to contracting. The amount was exceptional, but the institutional difficulty arose from the source and purpose of the money.

Verisign was not a named applicant for .WEB. It had entered into a Domain Acquisition Agreement with Nu Dotco on 25 August 2015 and a Confirmation of Understanding on 26 July 2016, one day before the auction began. The first IRP’s corrected final declaration and the ICANN Board’s April 2023 resolution and rationale establish the public outline: Verisign provided the funds used for Nu Dotco’s bid in exchange for, among other things, a contemplated future assignment of the .WEB registry agreement, subject to ICANN’s consent.

That description contains several distinct legal positions. Nu Dotco was the applicant admitted to the auction. Verisign supplied the financial capacity behind the winning bid. Nu Dotco and Verisign maintained that the object of any later transfer would be a registry agreement that did not yet exist, not the pending application. ICANN’s approval remained a condition of a future assignment. Whether these formal distinctions reflected the practical allocation of authority under the agreement was the contested question; it could not be answered merely by repeating the labels chosen by the contracting parties.

Afilias argued that Nu Dotco had become a vehicle through which Verisign acquired the practical benefits and control of the application; that the agreement transferred application-related rights or made Nu Dotco’s public application misleading; and that the auction was no longer the contest among disclosed applicants that the programme appeared to present. Nu Dotco and Verisign argued that Nu Dotco remained the applicant, retained programme responsibilities, submitted the bids and had agreed only to seek a later transfer of a registry agreement after contracting and with ICANN’s permission.

The parties also disputed whether the agreement altered information that Nu Dotco had been required to update before the auction.

The public record cannot independently resolve every detail because substantial parts of the DAA remain redacted. The 2023 Board rationale acknowledged that the agreement gave Verisign various rights concerning how Nu Dotco would proceed, while withholding important particulars as confidential. The first IRP declaration is also heavily redacted. These documents are sufficient to identify the institutional question, the competing interpretations, the decision-maker and the remedies sought. They are not sufficient to reconstruct every veto, incentive, obligation or commercial term.

The auction could not cure that evidentiary problem. Its authorised provider could administer the bidding rules, receive bids, identify the prevailing participant and determine the amount payable. It was not the body charged with deciding whether a separate financing contract changed the applicant’s ownership, control or rights under the Guidebook. Completion of the winning-payment obligation could establish compliance with the auction’s payment terms. It could not certify compliance with every non-auction rule or validate the DAA.

This is the first institutional answer to the central question. No single actor controlled the whole allocation. The auction resolved contention under its own rules. ICANN organisation and ultimately the Board had responsibility for applying programme rules to the application and the funding arrangement. An IRP panel could test whether ICANN had acted consistently with its Articles and Bylaws. Contracting and assignment rules governed the route from applicant to operator. The IANA functions supplied the public evidence of root-zone implementation.

The delay reflected both that divided authority and ICANN’s initial failure to use the first-instance power that belonged to it.

Four institutional objects, not one transferable prize

The dispute becomes more intelligible when four objects are kept separate.

The first was the application. Nu Dotco had submitted a proposal to operate .WEB. Under the 2012 Applicant Guidebook, the application carried representations, continuing information duties, evaluation requirements and restrictions on transferring application-related rights or obligations. It was not itself a delegated top-level domain, and it was not an executed registry agreement.

The second object was the auction position. Once the contention set reached the auction of last resort, the participating applicants bid for the right to have an application survive contention and continue through the programme. Nu Dotco’s US$135 million result was decisive among the seven .WEB applicants. It was not conclusive as to every remaining condition. The 2012 programme was organised in stages even though its rulebook did not expressly address the DAA structure.

The 2026 Round contention rules now state the sequence in direct, high-level terms: after an auction, one application is eligible to proceed towards delegation, pending applicant and application evaluation and successful execution of a contract.

The third object was the registry agreement. A successful applicant that cleared the required processes could enter a contract with ICANN to operate the gTLD. A registry agreement could later be assigned only through the applicable contractual procedure and with ICANN’s approval. Nu Dotco and Verisign built their defence around the temporal difference between an existing application and a future contract. They maintained that the DAA did not transfer the application, but contemplated a conditional assignment after Nu Dotco had first obtained a registry agreement.

The fourth object was delegation in the DNS root. Execution of a registry agreement was not identical to delegation. The operator and its registry arrangements still had to satisfy the applicable pre-delegation requirements, and the authoritative root-zone data had to change. The IANA record is therefore not a ceremonial confirmation of an auction result. It is evidence that the institutional chain reached the operational DNS, and it identifies the sponsoring organisation at that endpoint.

Different actors controlled different objects. Nu Dotco controlled its application subject to ICANN’s rules. The auction mechanism resolved contention among eligible applications. ICANN had to decide whether the applicant and agreement complied with the Guidebook and whether contracting should proceed. The Board retained the published decision-making role after the first IRP. ICANN also held consent authority over a later assignment of a registry agreement. The IRP panel could review ICANN’s constitutional compliance and issue the forms of relief available under the Bylaws and procedures.

The IANA record reflected implementation; it did not adjudicate the DAA.

Confusion entered whenever one stage was treated as proof of another. Saying that Nu Dotco won the auction did not answer whether Verisign had acquired prohibited application rights. Saying that Verisign funded the bid did not by itself prove that it legally controlled Nu Dotco’s application. Saying that a future assignment required ICANN’s consent did not answer whether the agreement had already shifted enough decision-making authority to trigger a disclosure or transfer rule. Saying that the first IRP found ICANN in breach did not mean the panel held the DAA unlawful.

Saying that the Board found no rule violation did not mean a later panel endorsed its conclusion. Saying that Verisign appears in the root record does not reveal the confidential terms that preceded delegation.

The case operated as an allocation constitution assembled from several instruments rather than a single decision. Its weakness was not merely textual ambiguity. Ambiguity interacted with divided authority. ICANN staff did not make a first-instance compliance determination; the Board deferred; operational processing later resumed; and the reviewer was told it should respect ICANN’s priority to decide the programme merits. That produced a circular gap. The institution with programme authority had not ruled, while the reviewing body was urged not to rule first.

The 2012 rules did not name the funding arrangement

The textual dispute began with two parts of the 2012 Guidebook. Its application terms said that an applicant would acquire rights in connection with a gTLD only upon entering a registry agreement and that the applicant could not resell, assign or transfer its rights or obligations in connection with the application. Elsewhere, the Guidebook required prompt notification when submitted information became untrue or inaccurate, including changes in financial position and ownership or control. A failure to report a change that rendered an application false or misleading could lead to denial.

Those provisions supported competing constructions because none expressly described a non-applicant financing an auction bid under a pre-existing agreement for a later registry transfer. Afilias advanced a functional reading. On that account, the decisive issue was not whether the DAA used the formal language of an application assignment, but whether the funding, contractual rights and promised future benefit transferred the practical substance of Nu Dotco’s position to Verisign.

If Verisign could direct bidding, constrain application decisions or determine the route to operation, the arrangement might amount to a transfer of application-related rights or a change that made Nu Dotco’s disclosures misleading even while formal title remained unchanged.

Nu Dotco and Verisign advanced a formal and temporal reading. Nu Dotco remained the named applicant, retained the obligations imposed on applicants, submitted the bids and paid the winning amount with funds supplied under the agreement. The contemplated transfer concerned a registry agreement that could arise only after contention resolution, evaluation and contracting. Because no registry agreement existed at the auction stage and a later assignment required ICANN’s approval, they argued that no present transfer of the application had occurred.

The Board ultimately accepted that distinction. In April 2023 it found that Nu Dotco remained the applicant and bidder, had not sold or transferred the application, and had agreed only to a possible future assignment of a registry agreement. Nu Dotco would first have to obtain that agreement and then ask ICANN to approve an assignment to Verisign. The Board also concluded that Nu Dotco’s bids were legitimate under the Auction Rules and that the applicant could satisfy the winning payment.

The rationale did not claim that the old rulebook supplied an explicit answer. The Board and the Board Accountability Mechanisms Committee described the DAA as falling into a grey area that the Guidebook and Auction Rules did not specifically address. Both sides, the Board said, had plausible arguments, but neither position fitted the arrangement exactly. That acknowledgment matters. The Board did not simply choose between two self-evident applications of a clear clause. It exercised interpretive discretion over a structure the programme had not anticipated in express terms.

The Board also directed ICANN to consider DAA-like agreements, proposed registry-agreement assignments and pre-auction communications when writing the rules for the next round. That prospective instruction was not a finding that Nu Dotco should have lost under the 2012 rules. It was not a general prohibition on third-party finance or future assignments. It was recognition that the earlier framework had failed to attach a sufficiently explicit evidence duty and decision point to a high-value funding-and-transfer arrangement.

A separate competition investigation illustrates why institutional boundaries mattered. The Board’s public record states that a United States Department of Justice inquiry closed in early 2018 without action. That did not determine whether Nu Dotco had complied with ICANN’s programme rules. The antitrust authority and ICANN were applying different instruments, using different procedures and possessing different remedies. Non-enforcement by one institution could not substitute for a reasoned decision by the other.

The rule-design problem was more precise than a general demand for transparency. A functional rule had to identify which relationships required disclosure, when disclosure was due, who would receive it and what consequence followed. Was material third-party auction funding enough? Did the duty arise only when the funder obtained decision rights? Did a pre-arranged future transfer always require notice? Was control defined by ownership, veto power, the ability to set bids, economic benefit or a combination? Could commercial detail remain confidential while the identity and institutional role of the funder were disclosed?

Would a breach trigger correction, re-evaluation, auction exclusion or disqualification?

Because the 2012 text did not answer these questions directly, ICANN interpreted the arrangement after the auction rather than applying a published test before the US$135 million result became the basis for progression. That remained a legitimacy problem even if the Board’s eventual reading was available under the text. Scarce DNS resources are allocated more credibly when the control test is known before bidding, competing applicants can see that the same test applies to all of them, and the responsible body must give reasons while review can still prevent implementation.

The process failed before the merits were decided

The first IRP’s central finding was procedural and constitutional. It did not hold that the DAA violated the Guidebook. It held that ICANN had failed to decide the complaint at the proper institutional level and had nevertheless moved towards contracting. That distinction defines both the breach and the limits of the remedy.

After Verisign’s role became public, ICANN sent a questionnaire to the .WEB applicants on 16 September 2016. The corrected IRP declaration later identified defects in the preparation and issuance of that questionnaire and in ICANN’s failure to tell Afilias about a Board decision made on 3 November 2016. The Board had chosen to defer consideration while another .WEB accountability process remained pending. Afilias was not informed. The panel concluded that these acts and omissions breached ICANN’s commitments to openness, transparency and fair procedure.

Deferral did not resolve the DAA. The sharper failure came in June 2018. The panel’s chronology records that ICANN took the .WEB contention set off hold on 6 June. Staff approved a draft registry agreement on 12 June and sent it to Nu Dotco on 14 June. Nu Dotco signed and returned it, and staff approved execution on ICANN’s behalf. Before ICANN executed the agreement, Afilias invoked the Cooperative Engagement Process on 18 June. ICANN put the contention set back on hold two days later, and the unexecuted agreement then in process was voided. When the cooperative-engagement stage closed, Afilias filed the IRP on 14 November 2018.

The panel treated the movement towards contracting as inconsistent with ICANN’s responsibilities because the organisation had still not pronounced on the central allegations. Staff had not made a first-instance compliance determination. The Board had neither stopped staff from progressing the application nor decided the complaint after the other accountability proceedings ended. Yet in the IRP, ICANN argued that the panel should respect the Board’s expertise and discretion on programme merits.

The position was institutionally untenable: ICANN invoked its priority to decide while allowing implementation machinery to advance without that decision.

The breach was framed against ICANN’s constitutional documents. The panel found that documented policies had not been applied objectively and fairly and that required standards of openness, transparency and fairness had not been met. Its criticism was not that ICANN had reached the wrong substantive interpretation. It was that ICANN had failed to reach one before taking steps towards a registry agreement that could have made the dispute much harder to remedy.

That explains how Afilias could prevail on liability without receiving .WEB. A process violation can be complete while the lawful substantive outcome remains open. A reviewing body may require an institution to make the decision it avoided, use a proper procedure and give reasons without substituting its own preferred allocation. The first .WEB panel adopted that remedial boundary. ICANN had to confront the complaint, but the panel would not become the New gTLD Program administrator in the first instance.

A related court case increased the practical importance of the internal accountability route. Ruby Glen, another .WEB applicant, sued ICANN. In October 2018, the United States Court of Appeals for the Ninth Circuit affirmed dismissal because the Applicant Guidebook’s covenant not to sue barred the claims. The court reasoned that Ruby Glen still had recourse through the Independent Review Process, which it treated as an alternative dispute-resolution mechanism. The court did not decide whether the DAA complied with the Guidebook. Its judgment nevertheless showed why the quality and remedial reach of the IRP mattered when applicants’ access to ordinary litigation was contractually restricted.

The first IRP therefore tested more than a private agreement. It tested whether ICANN could preserve first-instance discretion by postponing a decision while allowing an application to advance. The panel’s answer was no. Deference presupposes an exercise of authority. Institutional expertise does not justify institutional silence.

What the first IRP could bind—and what it refused to allocate

The panel issued its Final Decision on 20 May 2021, and a corrected version followed on 15 July. The declaration reached final status on 21 December 2021, when the panel denied Afilias’s separate Article 33 application for an additional decision and interpretation. ICANN’s Board records generally call the underlying instrument the Final Declaration.

Afilias sought relief that would have transformed the panel from reviewer into allocator. It asked for Nu Dotco’s application to be rejected, its auction bids to be disqualified, Nu Dotco to be deemed ineligible for a registry agreement, .WEB to be offered to Afilias as the next-highest bidder and Afilias’s price to be set at US$71.9 million. Those requests treated the alleged breach as capable of producing a substitute allocation through the IRP.

The panel refused. It held that ICANN possessed the relevant knowledge, expertise and programme responsibility to decide in the first instance whether the DAA complied with the rules and, if not, what consequences should follow. Because ICANN had not made that determination, the requested allocation remedies were premature. The panel did not direct ICANN to contract with Afilias, set a replacement price or itself invalidate the auction result.

The refusal did not make the IRP inconsequential. The panel declared that ICANN had violated its Articles and Bylaws. It directed ICANN to stay action that would further delegation until ICANN had considered the Final Decision. It separately recommended that the hold continue until the Board had considered the panel’s opinion, ruled on the DAA’s compliance and, if it found a violation, determined whether Nu Dotco’s application or bids should be rejected or disqualified. The distinction between direction and recommendation matters.

The first protected an immediate opportunity for institutional consideration; the second described the Board-level merits decision the panel believed should precede further delegation.

The panel also imposed monetary consequences, but not all in one direction. In the Final Declaration it ordered ICANN to reimburse US$450,000 in legal fees associated with Afilias’s emergency interim-relief proceedings and US$479,458.27 representing Afilias’s share of IRP costs. In that 21 December 2021 ruling, however, the panel denied Afilias’s application in full, found it frivolous and ordered Afilias to reimburse ICANN US$236,884.39 in legal fees incurred in responding to it. The panel left ICANN to bear US$140,335.30 in panel fees and expenses for that application, taking account of the IRP as a whole. The Board’s January 2022 resolutions preserved the distinction: Afilias remained the prevailing party on the core constitutional declarations, while ICANN prevailed on the later Article 33 application. The Board directed the reimbursements owed to Afilias and asked the Board Accountability Mechanisms Committee to examine the panel’s recommendation before any further processing of the .WEB applications.

These remedies expose the IRP’s constitutional role. The panel could determine whether ICANN action or inaction violated the Articles or Bylaws, preserve a meaningful opportunity for review and allocate specified costs. Its declarations and interim protection made continued processing contingent on the institution confronting a decision it had avoided. It was not automatically authorised to substitute its own programme decision whenever a constitutional violation was found.

That boundary did not prevent the IRP from affecting allocation. A declaration can invalidate the procedural basis on which an application advances. A stay can prevent delegation while the responsible body reconsiders. A finding can force a decision under the proper standard and create a record for later review. In a system where a root-zone change creates technical and commercial reliance, preventing premature implementation may be the most consequential remedy available. The panel’s power operated through sequence: it constrained the allocator before the operational endpoint, although it did not award the string itself.

The model contains a genuine tension. A reviewer unable to grant the claimant’s desired resource may appear weak after years of delay and expense. A reviewer that reallocates a top-level domain without the first-instance record, programme responsibility or participation of all affected interests may become an unaccountable allocator. The panel chose procedural correction rather than substantive substitution. That choice returned the burden to the Board. Once the declaration was final, ICANN could no longer rely on staff progression, indefinite deferral or abstract claims of expertise. It had to make and publish the missing decision.

The Board’s pause and its April 2023 interpretation

The Board’s response unfolded in stages. On 16 January 2022, it acknowledged the panel’s findings, accepted the specified cost consequences and referred the non-binding recommendation to the BAMC. The Board did not yet decide the DAA’s compliance, but it preserved the opportunity for a first-instance review before further processing.

On 10 March 2022, the Board made the control point explicit. Its resolution asked the BAMC to evaluate the allegations about the DAA and separate allegations about Afilias’s conduct during the auction blackout period. It requested recommendations on whether either side’s conduct warranted disqualification or another consequence and directed ICANN organisation to continue refraining from contracting for or delegating .WEB until ICANN had made its determination.

The post-IRP process included written submissions and supporting material from Altanovo, Nu Dotco and Verisign in July and August 2022. That participation did not give the parties decision authority. They could present evidence, contest interpretations and defend their interests. The BAMC could assess the record and recommend an outcome. The Board retained the institutional power to adopt the published determination and direct what ICANN organisation should do next.

On 30 April 2023, the Board made that determination. It found that Nu Dotco had not violated the Guidebook or Auction Rules by entering the DAA or participating in the auction. It directed the interim president and chief executive, or her designee, to continue processing Nu Dotco’s .WEB application. It declined to make a final determination at that time on whether Altanovo had breached the blackout period because the decision to progress Nu Dotco’s application made that issue unnecessary to the outcome then before the Board.

The reasoning rested on the separation between the application and a future registry agreement. Nu Dotco remained the applicant. In the Board’s view, it had not sold or transferred the application and had retained its programme responsibilities. The DAA contemplated a possible future assignment of a registry agreement. Nu Dotco would first have to obtain that contract; only then could it ask ICANN to approve an assignment to Verisign. The Board also found that Nu Dotco remained the bidder, made legitimate bids and could fulfil the payment obligation.

This was the first published programme-compliance decision that the earlier IRP had identified as missing. It was not an IRP holding and did not derive its authority from the panel’s view of the DAA. The panel had required ICANN to exercise first-instance responsibility; the Board then supplied its own interpretation of the 2012 rules. That distinction determined the next review route. A challenger could contest the Board’s reasoning, treatment of evidence, consistency and Bylaws compliance in a new accountability case. It could not accurately maintain that the first panel had already held the DAA prohibited.

The Board’s language also disclosed the uncertainty of its task. It said no provision directly addressed an arrangement like the DAA and described the agreement as a grey area. It acknowledged that Verisign possessed contractual rights concerning how Nu Dotco would proceed while leaving the details redacted. The Board nevertheless concluded that those rights did not amount to a transfer of the application or make Nu Dotco’s application false. The decision was therefore an interpretation of incomplete and partly confidential evidence under an old rulebook, not the mechanical application of a clause drafted for this structure.

Confidentiality imposed a continuing legitimacy cost. Parties may have proper reasons to protect commercially sensitive terms, and ICANN may need confidential access to investigate them. But competing applicants and the wider community must be able to test whether a published rationale fits the evidence. Here, the public could see the legal framework and the Board’s conclusion, but not every contractual constraint on which the conclusion depended. The responsible response is not to invent the missing text. It is to distinguish a reasoned public decision from a fully auditable one.

The Board also looked forward. It directed ICANN to consider arrangements like the DAA, proposed assignments and pre-auction communications when drafting the next Guidebook and auction rules. That prospective response recognised a governance deficit without reversing the 2012-round result. In effect, the Board held that the old rules did not prohibit Nu Dotco’s arrangement as the Board understood it, while accepting that future applicants should receive a clearer account of disclosure duties, control tests and permitted conduct.

A second review that ended without a merits declaration

Altanovo filed a second request for independent review on 14 July 2023. This time, the object of review was no longer ICANN’s failure to decide. The Board had issued a decision. The new case could test whether the April 2023 determination treated the evidence fairly, interpreted the Guidebook consistently, respected required procedures and complied with ICANN’s Articles and Bylaws.

The official docket shows a substantial proceeding rather than a placeholder filing. It includes ICANN’s response, requests by Nu Dotco and Verisign to participate as amici, arguments over the validity and scope of IRP Rule 7, expert reports, extensive exhibits, pre-hearing briefs and a series of procedural orders. Filings continued through November 2025. The sequence intended by the first panel had therefore materialised: first a Board programme decision, then an independent challenge to that decision.

It did not culminate in a public adjudication. The docket is titled “Altanovo Domains Limited (.WEB) - Withdrawn” and states that the claimant withdrew its request on 22 July 2026. No final merits declaration appears. Verisign announced on the same date that the disputes had been resolved and that the details were confidential.

Withdrawal has a precise evidentiary effect. It terminated the proceeding without a panel ruling that the Board was right or wrong. It did not establish that Altanovo’s claims lacked merit. It did not establish that ICANN, Nu Dotco or Verisign conceded them. The April 2023 resolution remained the operative published interpretation because no later merits declaration displaced it, not because the second panel independently endorsed it.

The statement that disputes had been resolved should likewise be treated as evidence of closure, not an account of its terms. The public can observe the withdrawal, announcement and delegation. It cannot see the confidential arrangements behind them. Institutional analysis should not fill that gap with a presumed settlement price, concession, admission or motive.

The second IRP still mattered. It kept the Board’s decision under an active accountability challenge, generated a detailed public docket and exposed competing theories about evidence, procedure and the scope of review. Its withdrawal nevertheless left the question it was positioned to answer—whether the 2023 Board decision itself complied with ICANN’s constitutional obligations—without an independent public resolution.

From confidential closure to public delegation

The final public steps occurred in quick succession. On 22 July 2026, Altanovo withdrew the second IRP and Verisign announced that .WEB had been delegated. On 23 July, the IANA record listed VeriSign, Inc. as sponsoring organisation. Those documents establish that the string entered the root, identify the registry sponsor and record Verisign’s stated intention to launch registrations later in 2026.

They do not establish every intervening contractual act. The Board’s 2023 rationale described a possible route in which Nu Dotco would obtain a registry agreement and then request ICANN’s approval for assignment. The 2026 endpoint is consistent with Verisign ultimately becoming the operator, but the cited public documents do not disclose whether that exact sequence occurred, whether a differently structured route was used or what obligations accompanied the confidential resolution.

That gap does not erase the allocation of formal power. A private agreement could distribute economic interests and impose duties between Nu Dotco and Verisign, but it could not insert .WEB into the root. Withdrawal could terminate the pending IRP, but it could not itself create a registry contract. Any route to delegation still required ICANN to complete the applicable application and contracting decisions and approve any change for which its consent was required. The IANA functions and root-zone processes then had to implement delegation. The final result depended on public institutional acts as well as private bargains.

Delegation changed the practical setting for any later remedy. Before the root-zone change, a hold could preserve the contested resource while ICANN reconsidered. After delegation, a remedy would encounter an operating registry, technical dependencies, registrar preparations and prospective registrants. That does not make post-delegation review legally impossible, but it raises continuity costs and makes reasoned decision-making before delegation more important. The first panel’s interim protection mattered because it operated before those dependencies hardened.

The .WEB case therefore shows how technical implementation can carry constitutional weight without becoming constitutional adjudication. IANA did not decide whether the DAA complied with the Guidebook. Yet the root-zone record made the allocation real in a way that an auction result, a Board resolution or a withdrawn pleading could not. Implementation was the final link in a legal and institutional chain, and its completion changed the remedial landscape.

The counterfactual rulebook

The useful counterfactual is not a blanket prohibition on third-party finance or future registry transfers. It is a rulebook that identifies their governance risks before an auction and assigns a decision stage to them. Had the 2012 Guidebook expressly required disclosure of material bid funding, contractual decision rights and pre-arranged post-auction transfers, ICANN could have assessed the DAA against a published test before accepting the US$135 million result as the basis for further processing.

The first element would have been a defined funding disclosure. An applicant should identify a non-applicant that supplies, guarantees or reimburses a material part of an auction bid. Disclosure would not make the funder the controller automatically. It would create the evidence needed to distinguish ordinary arm’s-length credit from an arrangement in which the funder can set bids, veto application decisions, direct contracting or receive the registry as the expected return.

The second element would have been a beneficial-control test. The rule should ask who can exercise direct or indirect decision-making over the application, auction strategy, contracting, registry policies or a relevant affiliate, whether through ownership, contractual rights or another mechanism. That approach moves beyond formal title without reducing control to an assumption that the economically strongest participant must be the “real” applicant. It requires specified evidence of power.

The third element would have been a duty to disclose a pre-arranged transfer. If an applicant had agreed before contention resolution to seek assignment of a future registry agreement, ICANN should have been told the intended transferee, the broad commercial purpose and any decision rights affecting the application. Detailed price and trade-secret terms might properly remain confidential, but facts relevant to equal treatment should not be invisible to ICANN until the eve of, or after, the auction.

The fourth element would have been a mandatory pre-auction determination. Disclosure without a decision point can reproduce the same delay. A designated ICANN body should have decided whether the arrangement changed control, rendered the application inaccurate, undermined the applicant’s bona fide intention to operate or otherwise affected auction eligibility. It should have published reasons sufficient for competing applicants to understand the test while protecting genuinely confidential details. An accountability route should have been available before the auction or, at minimum, before contracting and delegation.

The fifth element would have been a staged remedy. Not every omission should produce the same consequence. ICANN could require correction, publication or re-evaluation where a defect was curable; pause auction eligibility while control facts were disputed; disqualify an applicant where material control was deliberately concealed or the applicant lacked a genuine operational role; and review any later assignment separately. The rule should identify the decision-maker, evidentiary standard, timetable and what remains on hold during review.

The 2026 Round Applicant Guidebook addresses parts of this counterfactual. Its application question on Ultimate Control requires an applicant to disclose persons or entities, including those providing financing, that exercise or can exercise direct or indirect decision-making or management over the application or related operations through ownership, contractual rights or another means. The question is not a demand to identify every lender. It is a control-linked disclosure that reaches beyond the formal name of the applying entity.

The Application Change Request rules require an applicant to update information that becomes untrue, inaccurate or outdated promptly and, in any event, within seven days of becoming aware of the relevant circumstance. Material changes to public portions of an application can be published for a 30-day comment period, and a failure to notify ICANN of changes that make information false or misleading may prevent an application from proceeding. These provisions attach timing and process to the information duty more explicitly than the 2012 controversy did in practice.

The 2026 Guidebook’s Good Faith Intent provision requires an applicant to submit each application with a bona fide intention to operate the applied-for gTLD. Separately, Module 5’s contention rules prohibit specified communications and arrangements among applicants for the same string, including certain exchanges of value for withdrawal and negotiations over post-auction transfers. They require applicants to certify compliance before signing a registry agreement or withdrawing and reserve remedies including disqualification, fee forfeiture, financial penalties and legal action. Module 5 also makes clear that an auction leaves one application eligible to continue, still subject to evaluation and successful contracting.

These are material improvements, but they do not erase every issue exposed by .WEB. The Ultimate Control question reaches a financing provider only when the provider exercises or can exercise relevant decision-making or management; parties may still disagree about whether contractual protections cross that line. The private-contention rules concentrate on communications and arrangements among applicants for the same string, while the .WEB DAA involved a non-applicant funder. A future assignment can still raise questions about how long an applicant must genuinely intend to operate and when a planned transition becomes inconsistent with that intention. As of 31 July 2026, however, Module 5 still described its auction section as a high-level overview and said that detailed auction rules, procedures and a schedule would be developed with the auction provider and made available no later than 60 days before the first auction. ICANN’s auction-provider procurement asked a provider to design, administer and operate the auctions and placed final evaluation, contracting and award in the third quarter of 2026. The current Guidebook therefore supplies authoritative principles and eligibility rules, but the later detailed auction instrument remains forthcoming. Future payment, bidder-qualification and operational procedures cannot yet be assessed as though that instrument had already been published.

The deeper lesson is not that drafting can eliminate ambiguity. It is that foreseeable ambiguity should be connected to an evidence duty, a named decision-maker and a stage at which review remains effective. A question asking who has ultimate control compels an applicant to identify relevant actors. ICANN must assess the answer. Material changes can enter a public process. Competing applicants can challenge a reasoned determination. A reviewer can then test an actual decision instead of confronting years of deferral after a US$135 million auction.

Had that structure existed in 2016, several outcomes were possible. ICANN might have found the DAA permissible and admitted Nu Dotco to the auction with the relationship disclosed. It might have required amendments or safeguards. It might have found that Verisign’s rights crossed a control threshold and required a change in the application or excluded Nu Dotco. Whatever the result, the principal institutional question would have been answered before the auction price became commercially and politically difficult to unwind. The counterfactual does not guarantee a different winner.

It relocates interpretation to the stage where equal treatment and effective review are most credible.

Allocation authority after ten years

The answer to the .WEB question is layered. Nu Dotco’s US$135 million bid gave its application the winning position in the 2016 auction of last resort. The auction did not possess authority to make the definitive programme-compliance decision about the DAA. ICANN did. After the first IRP exposed ICANN’s failure to exercise that responsibility before advancing towards contracting, the Board, supported by the BAMC review, issued the April 2023 determination that Nu Dotco had not violated the Guidebook or Auction Rules.

The first IRP’s authority was significant but bounded. It could declare that ICANN action and inaction violated the Articles and Bylaws, preserve the status quo until ICANN considered the declaration, recommend a first-instance Board determination and award specified costs. It refused to disqualify Nu Dotco itself, transfer .WEB to Afilias or set a replacement auction price. Review constrained the allocator; it did not replace the allocator.

The second IRP could have supplied independent review of the Board’s merits decision. Its withdrawal in July 2026 meant that no public final declaration performed that function. The Board’s reasoning remained operative, but it did not acquire an independent merits endorsement from the withdrawn case. Confidential closure ended the dispute without answering every public evidentiary question.

ICANN’s contracting and consent powers then connected the application outcome to the registry operator. The IANA record supplied the public evidence of the final operational act. Naming Verisign as sponsoring organisation is an endpoint that neither the auction nor the IRP could produce alone. The auction selected; the Board interpreted; the IRP reviewed constitutional compliance; contracting translated an application into operator rights; and root-zone implementation made the top-level domain exist in the DNS.

The remaining uncertainty is bounded and consequential. Public sources do not disclose the terms on which the second IRP was withdrawn or the exact contractual sequence by which Verisign became the sponsoring organisation. A sound account cannot replace that missing evidence with a presumed payment, concession or admission. The record is nevertheless sufficient to describe the institutional structure: auction selection, delayed first-instance interpretation, binding constitutional declarations and a bounded stay, an unadjudicated final challenge and operational delegation.

After almost ten years, .WEB did not reveal one sovereign inside ICANN’s system. It revealed a chain of specialised powers, each capable of stopping or advancing the string and none sufficient by itself. The failure occurred when the institution with first-instance authority delayed its decision while implementation moved. The corrective came when independent review made further processing contingent on ICANN considering the declaration and recommended that the Board exercise its first-instance authority without the panel dictating the allocation. Finality arrived only when the root-zone record changed.