[Saksham Agrawal is a third-year B.A., LLB, Hons., student at National Law School of India University, Bangalore]
In Engineering Projects India Ltd v. MSA Global LLC (25 July 2025), the Delhi High Court issued an anti-arbitration injunction restraining the continuation of an ICC arbitration seated in Singapore. The judgment purports to protect the integrity of the arbitral process in the face of alleged procedural misconduct, specifically, an arbitrator’s failure to disclose relevant prior dealings with the appointing party. The arbitration, the Court held, had taken on an “oppressive” character, and to allow it to proceed would be unconscionable and contrary to Indian public policy (paragraph 104). The tribunal, the Court feared, was being tactically manipulated to foreclose a meaningful adjudication of challenge proceedings that were still underway before the ICC and Singaporean courts (paragraph 120).
Although the facts raise understandable concerns about party autonomy, the judgment itself is an unsustainable departure from the core principles of international arbitration law. The Court’s assumption of supervisory jurisdiction over a foreign-seated arbitration, in which the institutional mechanism and the seat court were both actively seized, represents a direct challenge to the architecture of arbitral territoriality. It disregards the structure of the New York Convention, the legislative scheme of the Arbitration and Conciliation Act, (particularly post-Bharat Aluminium Co. v. Kaiser Aluminium Technical Services Inc. or BALCO, as it is popularly known), and the allocation of judicial responsibility in the UNCITRAL Model Law. More broadly, it reveals an anxiety about institutional accountability in arbitration that, while normatively significant, cannot be addressed through jurisdictional overreach.
This article attempts to evaluate the Court’s unprecedented intervention, contending that it marks an unwarranted assertion of jurisdiction over a foreign-seated arbitration. First, it examines the judgment’s deviation from the territorial framework governing arbitral supervision. Second, it considers whether concerns around arbitrator disclosure and procedural fairness can justify such judicial intervention.
Territoriality and the Architecture of Arbitral Supervision
The international legal framework governing arbitration is structured around the principle of territoriality, which designates the courts of the seat of arbitration as the exclusive locus of procedural supervision. This principle was given formal expression in Article 5 of the UNCITRAL Model Law, which states that “no court shall intervene except where so provided in this Law.” Article 6 then assigns to the courts of the seat the authority to rule on matters such as interim measures, challenges to arbitrators, and applications to set aside awards. These provisions are not procedural niceties: they comprise the fundamental ordering of jurisdiction in international arbitration. Their purpose is to ensure that arbitral proceedings are insulated from parallel interference by multiple legal systems, and that all supervisory review is channelled through a single forum.
The Indian legal system has long incorporated this model. The Supreme Court’s decision in BALCO marked a turning point, holding that the jurisdiction of Indian courts under Part I of the Act does not extend to arbitrations seated outside India. In rejecting earlier authorities such as Bhatia International v. Bulk Trading SA, the Court in BALCO adopted a strict territorialist position, affirming that the seat of arbitration necessarily determines the law that governs arbitration (BALCO, paragraph 121). Following BALCO, Indian courts have consistently declined to interfere with the conduct of foreign-seated arbitrations, recognising that procedural integrity is a matter for the seat court and the applicable institutional rules (see here, here, and here).
The Court’s decision in EPI v. MSA Global, however, disregards this settled position. Despite acknowledging that Singapore was the designated seat, and that the ICC was the administering institution, the Court assumed jurisdiction not under the Act but under its inherent powers via Sections 9 and 151 of the Code of Civil Procedure (paragraph 63). The Court asserted that although interim relief is barred, civil courts retain limited authority to grant anti-arbitration injunctions in exceptional cases of oppression or abuse (paragraphs 55–64). It found that the arbitration was proceeding in a manner that would cause grave injustice to the plaintiff due to the continued participation of an arbitrator who had failed to disclose a prior connection with the claimant’s controller (paragraph 121). On this basis, the Court issued an anti-arbitration injunction, restraining the continuation of proceedings in Singapore (paragraph 123).
This sidestep of BALCO is untenable. First, the Act is a self-contained code that expressly limits judicial intervention to what it permits (Section 5). BALCO emphasised that any contrary reading would unravel the coherence of the seat-based model, which anchors arbitral autonomy in the law and courts of the seat. Second, by granting an anti-arbitration injunction in the absence of any challenge to the validity of the arbitration agreement or the tribunal’s jurisdiction, matters being simultaneously reviewed by the Singapore courts, the Court effectively arrogated to itself a supervisory function it does not possess under Indian law. The judgment offers no cogent justification for this deviation, nor does it engage meaningfully with the core holding in BALCO. It thus sits in open contradiction to binding precedent and weakens India’s commitment to a seat-based arbitration regime.
Very recently, in Star Hydro Power Ltd v National Transmission and Despatch Company Ltd, the English Court of Appeal reasoned that to allow such pre-emptive actions would violate the New York Convention’s allocation of authority under Article V, which reserves procedural supervision to the courts of the seat (paragraph 61). The Court’s judgment in EPI v. MSA Global thus places Indian law at odds not only with its own precedents, but also with the jurisprudence of leading arbitral jurisdictions.
Disclosure, Party Autonomy, and the Problem of the Remedial Vacuum
While the jurisdictional defect in the Court’s reasoning is plain, it would be reductive to treat EPI v. MSA Global as merely an error of supervisory allocation. The judgment responds, albeit in a misdirected fashion, to a deeper tension within arbitral procedure: the gap between disclosure obligations and effective remedial mechanisms. The arbitrator in question, Mr Yeap, failed to disclose his prior involvement in arbitral proceedings where MSA’s principal, Mr Atwal, had also been a party (paragraphs 4-9). The ICC Court, in applying Article 11(2) of the ICC Rules, acknowledged that this failure was “regrettable,” but nonetheless declined to remove Mr Yeap, concluding that there were no “reasonable doubts” as to his independence (paragraph 10).
The Court found this institutional response inadequate. In its view, Mr Yeap’s non-disclosure undermined EPI’s ability to make an informed choice regarding whether to challenge the arbitrator, a procedural entitlement rooted in party autonomy (paragraph 85). What troubled the Court was not the absence of demonstrable bias, but the procedural asymmetry created by the non-disclosure. It focused on the potential perception of bias “from the standpoint of the parties” as under the ICC Rules (paragraph 80). To allow the arbitration to proceed under these conditions was, in the Court’s words, “oppressive” and “unreasonably harsh” (paragraph 90).
This concern is not unique to Indian jurisprudence. Recent investment arbitration cases have similarly recognised that the failure to disclose relevant material, even in the absence of actual bias, can infringe the procedural rights of the non-disclosing party. In Rockhopper Italia S.P.A. v. Italian Republic, the ad hoc annulment committee annulled an award on the ground that an arbitrator had failed to disclose his prior criminal convictions. The committee held that the failure affected the integrity of the process and the constitution of the arbitral tribunal by denying the respondent the opportunity to evaluate the information and challenge the appointment, even if there was no evidence of bias (paragraphs 411-416).
Moreover, under the ICC’s own procedural regime, the standard for disclosure under Article 11(2) is distinct from the standard for disqualification. Whereas the latter turns on “justifiable doubts” as to impartiality, the former requires disclosure of any information that “might” give rise to such doubts “in the eyes of the parties.” The point of the disclosure rule is prophylactic: it protects not only against actual bias but also against the loss of procedural legitimacy that arises when information is asymmetrically held. Non-disclosure deprives the party of its agency to assess neutrality, and thereby impairs the consensual basis of arbitration.
This conceptual distinction between procedural injury and actual bias was insufficiently appreciated by the ICC (paragraph 10). It is against this backdrop that the Court’s frustration can be understood. What it confronted was a legal system in which the breach of a disclosure obligation is formally acknowledged, but does not, on its own, give rise to a remedy. The party’s loss, i.e., the deprivation of a meaningful choice, is neither quantified nor redressed.
The question, then, is not whether the Court identified a real problem. It did. The problem lies in the remedy it chose. International arbitration has long recognised that not every procedural harm is justiciable in real time. The system allows for institutional error and relies on deferred review either at the seat (for setting aside) or in the enforcing court (Article V of the New York Convention). EPI v. MSA Global refuses to accept this structural constraint. By intervening in the midst of ongoing proceedings, it collapses the temporal and jurisdictional architecture of arbitration and asserts a supervisory prerogative that Indian courts have long disavowed.
To allow any national court to substitute its own view of fairness for that of the seat court is to invite fragmentation. Parties would be incentivised to initiate satellite litigation in their home courts, relying on elastic concepts of “public policy” or “oppression” to resist adverse procedural outcomes. The risk is not hypothetical. In EPI v. MSA Global, the Singapore High Court had already been seized of the challenge to Mr Yeap’s appointment (paragraph 29). The Court was not supplementing a gap in review; it was pre-empting a decision by the only court with jurisdiction to grant relief. The result is a duplication of proceedings, forum shopping, and the very inefficiency arbitration is designed to avoid.
More troubling still is the possibility that such interventions could undermine the enforceability of awards. If tribunals are restrained mid-proceeding by courts lacking supervisory jurisdiction, the final award may be indefinitely delayed or never rendered. In Yukos Capital SARL v OJSC Rosneft Oil Company, the English High Court held that a set-aside decision by the seat court could be disregarded at the enforcement stage where it violated natural justice. But that case involved an enforcement court looking at whether the award fails one of the New York Convention grounds. EPI v. MSA Global involves a non-seat court interfering with the arbitration while it is proceeding. To transpose Yukos Capital’s reasoning to the context of anti-arbitration injunctions would convert a narrow exception into a licence for judicial overreach.
To be clear, this is not an argument for institutional deference at all costs. Arbitration institutions must be held accountable for procedural failures, particularly where disclosure obligations are breached. But that accountability must operate within the limits of the system’s architecture. The appropriate remedy is either review by the seat court, or, if that court’s decision is so egregiously flawed as to offend fundamental fairness, a refusal to enforce the award in a jurisdiction with a public policy exception. What cannot be permitted is a system in which multiple national courts act simultaneously and inconsistently, each asserting its own procedural vision. The judgment in EPI v. MSA Global invites just that.
Conclusion
The Court’s intervention, though driven by concerns over procedural fairness and party autonomy, unsettles foundational principle of seat-based jurisdiction. By asserting supervisory authority over a foreign-seated arbitration, it risks undermining coherence in arbitral governance and casts doubt on India’s commitment to a rules-based arbitration framework. Courts, for their part, must resist the temptation to fill perceived gaps through jurisdictional overreach. Instead, they should defer to the seat’s legal order, intervening only at the enforcement stage and only when public policy is clearly violated. Rebalancing in this way preserves both party autonomy and systemic legitimacy.
– Saksham Agrawal