Introduction
The Delhi High Court’s handling of the running dispute between Ashiana Ispat Limited (AIL) and Kamdhenu Limited (KL) over the mark ‘AL KAMDHENU GOLD’ has produced two distinct but related judicial moments: a September 2025 Division Bench ruling on the Trade Marks Registry’s examination failures, and an April 2026 interim injunction resolving (provisionally) who actually owns the disputed mark. Read together, they offer a useful case study in a question that has nagged Indian trademark practice for years; does the Registry’s examination process provide the safeguard the Trade Marks Act, 1999 promises, or has it become a formality that courts must repeatedly step in to correct?
This article argues that the Delhi High Court’s decision in Ashiana Ispat Ltd v. Kamdhenu Ltd (2025 DHC 7801-DB) is an important and welcome step in strengthening judicial oversight of the trademark registration process. At the same time, the judgment highlights a broader institutional problem. Over time, the Registry’s examination process has come to be treated as a largely administrative exercise, even though the Trade Marks Act, 1999 and now the Delhi High Court has made it clear that examiners are expected to perform a quasi-judicial function that requires independent reasoning and careful application of the law. Judicial correction, however well-reasoned, is not a substitute for systemic reform, and this case exposes exactly why.
A History of Judicial Correction: Precedents Predating Ashiana Ispat
The examination failure at the heart of this dispute did not arise in a vacuum, and it is worth situating it against a pattern of prior Delhi High Court interventions into the Registry’s internal processes, a pattern that suggests the Kamdhenu litigation is less an aberration than a continuation of a long-running institutional problem. In Asianet Star Communications Pvt. Ltd. v. Registrar of Trademarks (W.P.(C) 11284/2019), the Court found that a lack of coordination between the Registry’s own departments, those handling change-of-address requests, renewals, and post-registration matters had caused a mark to be abandoned for non-renewal after a renewal notice was sent to an outdated address; the Court went beyond the individual petitioner’s relief to direct a broader review of the Registry’s processing procedures, from filing through to renewal. In Pawandeep Singh v. Registrar of Trademarks WP(C) – IPD 7/2022 (SCC OnLine), Justice Prathiba M. Singh noted that the Registry’s virtual platform was severely outdated, allowing only three people to join a call at a time. She held that orders passed by the Registrar deal with precious business rights, and refusing a mark without a fair hearing completely offends natural justice. In Shell Brands International AG v. The Registrar of Trade Marks 2023: DHC:2955, the Delhi High Court set aside an order of the Trade Marks Registry refusing registration, holding the decision to be unreasoned and vitiated by procedural impropriety. Justice Amit Bansal observed that under Rule 27 of the Trade Marks Rules, 2002, which governed the application at the material time, the submission of a user affidavit was not a mandatory, automatic requirement upon initial filing, but rather a discretionary power vested in the Registrar. Consequently, the Court held that the Registry could not summarily reject an application for the absence of a user affidavit without first explicitly directing the applicant to furnish one, ruling that such a practice constitutes a clear breach of natural justice and fair play. Furthermore, the judgment addressed broader systemic irregularities within the Registry, strongly deprecating the issuance of cryptic, non-speaking orders, the introduction of uncommunicated grounds of refusal, and the imposition of evidentiary requirements at the eleventh hour without affording the applicant an opportunity to be heard. The Court underscored that administrative decisions must be fully reasoned and transparently anchored in both law and fact. Finally, the judgment implicitly contextualized the statutory evolution of the provision, noting that while the 2002 Rules employed the permissive “may” to render the filing discretionary, the subsequent Trade Marks Rules, 2017 relocated this provision to Rule 25 and substituted the term with “shall” for prospective applications.
Read together, these cases establish that the specific failure in the Kamdhenu dispute, i.e marks not properly surfaced and cited during examination, is not a one-off lapse but a recurring institutional weakness that the judiciary has been correcting, case by case, for years without the underlying process itself being reformed.
Factual Background
The trademark dispute between KL and AIL stems from a 1997 licensing arrangement following KL’s 1994 adoption of the “KAMDHENU” mark. A 2002 agreement allowed AIL to use and seek registration for the composite mark “AL KAMDHENU GOLD,” explicitly tying AIL’s absolute ownership to successful registration while remaining dependent on KL’s proprietary rights in the interim. However, AIL’s initial 2002 application was abandoned in 2008 due to an alleged attorney oversight, sparking a long standing dispute over AIL’s claims of continuous commercial usage. Following the realization of the prior abandonment, AIL pursued fresh refilings in 2024, the same year KL officially terminated the underlying License User Agreement on September 19, 2024, over delayed royalties and unfulfilled orders.
Over time, AIL filed five separate applications for the composite mark, which the Trade Marks Registry published or examined without citing KL’s prior registered marks as conflicts. Rather than filing standard statutory oppositions, KL successfully bypassed the process using a writ petition to protect the integrity of the public register.
A Single Judge agreed. The Registry, for its part, effectively conceded the error: it retracted an acceptance order, opened disciplinary proceedings against the examiners involved, and the Single Judge directed a fresh, de novo examination of all five applications without formally issuing notice to AIL or allowing it to file a counter affidavit. AIL appealed that procedural omission to the Division Bench, setting up the 2025 ruling this article examines.
Failure to Cite Prior Marks
At the root of this entire dispute lies a failure so basic that it deserves to be named as its own pathology rather than folded into a general critique of Registry inefficiency: what might be called the “blind search” syndrome, an examination process that formally runs a conflict search but functionally fails to see what is directly in front of it.
The mechanics of the failure bear restating plainly. Section 11 of the Trade Marks Act, 1999 obliges the Registrar to refuse registration of a mark that is identical or deceptively similar to an earlier mark, in respect of similar goods or services, where confusion is likely. To operationalise this, the Examination Report is supposed to be preceded by a search of the register that surfaces every conflicting prior mark, so both the examiner and the applicant know exactly what objections exist before the application is advertised. In the AIL applications, that search reportedly failed to cite a substantial number of KL’s own registered marks, by some accounts as many as sixteen despite the fact that AIL’s applied-for mark, ‘AL KAMDHENU GOLD’, shares its entire dominant and most distinctive element, the coined word ‘KAMDHENU’, with KL’s registrations. This is not a case of two marks sharing a common descriptive term, or a phonetic overlap requiring nuanced comparison. It is, on its face, about as straightforward a conflict as trademark examination ever produces.
Three features make this “blind search” pattern particularly troubling, rather than merely regrettable.
First, it is not isolated to a single application. The omission recurred across multiple filings by the same applicant, for the same core mark, over an extended period. A one-off miss might be explained by examiner oversight on a busy day; a repeated miss across several applications points instead to a structural gap, either in how the Registry’s search database is indexed and queried, or in the practices examiners are trained to follow when a mark incorporates another proprietor’s coined term as its dominant feature.
Second, the error was not self-correcting. The applications were not merely examined imperfectly and then caught at the opposition stage, as the statute’s layered safeguards are designed to allow. Several were advertised, and one was accepted outright, meaning the “blind spot” persisted through multiple stages of internal review before an external party (KL, via writ petition) was forced to catch what the Registry’s own process should have caught. That the Registry, once alerted, retracted the acceptance and opened disciplinary proceedings against the examiners involved is telling; it amounts to an institutional admission that this was a process failure, not a defensible judgment call.
Third, and most significantly for the “syndrome” framing, the burden of catching the blind spot fell entirely on the prior proprietor. Kamdhenu was left to police its own register entries, monitoring journal publications, cross-checking new filings, and ultimately resorting to writ jurisdiction, rather than being able to rely on the examination process to do the work the statute assigns to it. A “blind search” does not merely produce an isolated error; it quietly transfers the cost of vigilance from the institution designed to bear it (the Registry) onto private parties, who must now duplicate, at their own expense, a function they are statutorily entitled to expect the state to perform.
Framed this way, the Division Bench’s insistence that the Registrar is “duty-bound” to cite all prior conflicting marks in the FER is not simply doctrinal housekeeping. It is a direct response to a demonstrated pattern of the search function failing at exactly the point where it matters most, cases involving a shared, highly distinctive coined term, which ought to be the easiest conflicts for any properly functioning search tool to flag, automated or otherwise.
The Division Bench’s Reasoning
Justices C. Hari Shankar and Om Prakash Shukla were asked to resolve several interlinked questions: whether a writ petition under Article 226 was maintainable at all given the existence of an alternate statutory remedy (opposition under Section 21); whether, even if maintainable, the Single Judge ought to have entertained it; whether natural justice had been violated by remanding the applications without hearing AIL; and, more broadly, how far courts should intervene in the Registry’s internal procedures before the opposition stage is even reached.
On maintainability, the Bench held to well-established Supreme Court principle; the existence of an alternate remedy is a rule of discretion, not an absolute jurisdictional bar. Courts retain the power to entertain a writ in exceptional circumstances, breach of natural justice, want of jurisdiction, or violation of fundamental rights being the recognised categories. This is not novel doctrine, but its application here is significant: it treats a defect in the Registry’s search-and-cite function as potentially rising to the level of a natural justice violation, not merely a correctable clerical slip to be sorted out later in opposition proceedings.
Substantively, the Bench affirmed something that ought to be uncontroversial but evidently required restating: the Registrar is duty bound to cite all prior, confusingly similar marks in the First Examination Report, and proprietors of existing registrations have a right to see their marks properly considered before a competing application is advertised. Framed this way, the Registry’s examination is not a courtesy extended to applicants; it is a check exercised on behalf of the entire register’s integrity, and existing rights-holders are entitled to rely on it being done properly.
A Quasi-Judicial Function, Treated Administratively
The most conceptually important move in the judgment and the one that deserves the most critical scrutiny is the characterisation of the Registry’s examination function as quasi-judicial rather than purely administrative. This distinction is not academic. An administrative act attracts a lower standard of procedural rigor and a correspondingly narrower scope for judicial review; a quasi-judicial one demands reasoned decision-making, adherence to natural justice, and a higher tolerance for writ intervention precisely because errors touch on vested rights.
The problem this case surfaces is that the Registry, as an institution, does not appear to operate as though it understands itself in quasi-judicial terms. As the “blind search” pattern discussed above illustrates, the omission was not a subtle, borderline case of trademark similarity that a reasonable examiner might miss; it was the kind of failure that suggests inadequate search methodology, database gaps, or an examiner workload that leaves too little time for the careful cross-referencing the statute contemplates.
Herein lies the critical tension: the judiciary can declare the function quasi-judicial and insist on procedural rigor case by case, but it cannot, through adjudication alone, fix an under-resourced or poorly systematised examination apparatus. The Division Bench’s ruling corrects the specific injustice before it; it does not and structurally cannot mandate the IT infrastructure, staffing levels, or training protocols that would prevent recurrence. That gap between judicial correction and institutional reform is where the real accountability deficit lies.
The Procedural Fairness Problem Cuts Both Ways
A point too easily lost in commentary sympathetic to Kamdhenu’s position is that AIL’s own procedural grievance was not trivial. The Single Judge’s order for de novo examination was passed without notice to AIL and without allowing it to file a counter-affidavit, a due process shortfall of exactly the kind the Registry itself was accused of. This is a useful corrective to any narrative that frames the case as simply “Registry bad, court good.” Natural justice is not a one-way ratchet that only protects incumbent rights-holders against a negligent Registry; it equally protects an applicant from having its pending rights unwound without a hearing, however sympathetic the underlying complaint against the Registry might be.
The Division Bench’s willingness to take this seriously, treating it as a live ground of appeal, rather than a technicality to be waved through in the interest of correcting the Registry’s error, reflects well on the court’s even handedness. But it also illustrates a second, less remarked upon institutional failure when the Registry errs, the remedy for that error can itself generate fresh procedural harm to third parties who relied, in good faith, on the Registry’s original (flawed) processing of their applications. A better-functioning Registry would prevent this two-tier harm entirely, rather than leaving courts to referee a correction process that inevitably creates new casualties.
Comparative Context and the International Benchmark Problem
Commentary surrounding the ruling has pointed to the European Union Intellectual Property Office, the United States Patent and Trademark Office, and the UK Intellectual Property Office as offices with more developed conflict-search tooling and applicant-communication standards. This comparison is fair as far as it goes, but it should not be deployed uncritically. India’s Trade Marks Registry operates at a dramatically different volume and resourcing scale, and importing wholesale procedural architecture from better funded offices is not a costless fix. The more useful lesson from those comparators is not their specific software or staffing ratios, but their underlying institutional posture: examination is treated as the primary integrity check on the register, with opposition proceedings functioning as a secondary safety net, not, as appears to be the practical reality in India, the other way around.
If opposition proceedings under Section 21 are, in practice, doing the heavy lifting that examination reports were meant to do, then the four-month opposition window becomes the de facto first line of defence for existing proprietors, who must now actively monitor the trademark journal rather than trust that the Registry will flag conflicts on their behalf. That is a meaningfully worse system than the one the statute appears to contemplate, and it disproportionately burdens smaller proprietors who lack the resources for continuous journal-watching, while larger, litigation-capable entities like Kamdhenu can absorb the cost of writ petitions when the Registry fails them.
The Aftermath: What the Merits Ruling Reveals
The subsequent April 2026 ruling on the underlying ownership dispute, where the court found the 2002 agreement to be a licence rather than an assignment, and granted Kamdhenu a comprehensive interim injunction against AIL, is worth reading back into the Registry accountability question. It underscores just how much was riding on the Registry getting its examination right the first time. Had the initial search and cite process functioned properly, Kamdhenu’s prior rights would have been flagged automatically, opposition would likely have followed as a matter of course, and the parties’ underlying contractual dispute over licence versus assignment might have been resolved through the ordinary opposition process years earlier rather than culminating, years later, in an injunction that has forced AIL to suspend commercial operations entirely, with knock on consequences serious enough to generate a qualified audit opinion and going-concern doubts in its financial statements.
This is the strongest practical argument for taking Registry accountability seriously, examination failures are not abstract administrative lapses. They generate years of downstream litigation, commercial uncertainty for both sides, and as here genuinely severe financial consequences for a company whose operations ultimately depended on a registration process that malfunctioned at the outset.
Conclusion
Ashiana Ispat v. Kamdhenu is a doctrinally sound and, in many respects, admirable judgment. It correctly reasserts that the Registry’s examination function is quasi-judicial, that existing proprietors have an enforceable right to have their marks properly considered, and that writ jurisdiction remains available in appropriate cases despite the existence of statutory alternate remedies. It also, commendably, protects the procedural rights of the party accused of benefiting from the Registry’s error, refusing to let institutional failure justify shortcuts in natural justice for anyone.
But the case should not be read as evidence that the system is working. It is better read as evidence that the judiciary is, once again, doing the Registry’s job for it after the fact correcting individual injustices without addressing the structural conditions that produced them. Until the Registry’s internal search methodology, examiner training, and resourcing are reformed as a matter of institutional policy rather than case by case judicial admonition, disputes of this kind will keep recurring, and the register’s integrity will continue to depend less on the examination process the statute prescribes and more on which proprietors have the resources and appetite to litigate their way to protection. That is not accountability. It is a workaround.
