For the first time, the DIFC Court of Appeal has set aside substantive parts of a DIFC-seated DIAC award under Article 41 of the Arbitration Law — and set aside the first-instance judgment that upheld it for reasons too thin to review. Oheo Bank v Parker [2025] DIFC CA 006 is now the leading DIFC authority on the fair-hearing and scope-of-submission grounds.
The DIFC's supervisory jurisdiction over arbitration is built on a familiar promise — maximum support, minimum interference. In Oheo Bank v Parker [2025] DIFC CA 006 (24 April 2026), the Court of Appeal marked the outer edge of that promise. It is the first known occasion on which the DIFC Court of Appeal has set aside part of a DIFC-seated award under the DIFC Arbitration Law (DIFC Law No. 1 of 2008), and it did so while also setting aside — in full — the Court of First Instance judgment that had upheld the award, on the separate footing that the reasons given were too sparse to be reviewed.
The dispute in one paragraph
A customer, Parker, sued its bank over a financing structure that left it holding a worthless bond. Of five pleaded heads of claim — deceit, misrepresentation, breach of the Quincecare duty, breach of regulatory duties, and negligence — the tribunal dismissed every one save a single regulatory-conduct claim worth EUR 1 million, upheld by a majority. The relationship manager's conduct was found to be dishonest, but, as the tribunal put it, his was "not relevant dishonesty". The claim that ultimately succeeded rested on a duty to explain why an indemnity was required — and, on the tribunal's own unanimous finding, it had never been pleaded until the claimant's post-hearing brief.
The award was set aside for unfairness; the judgment upholding it was set aside for opacity. One narrow fact-pattern, two set-asides.
The framework the Court settled
Because there was no prior DIFC appellate authority on Articles 41(2)(a)(ii) and (iii), the Court surveyed the common-law world — Australia, Canada, England, Hong Kong, Ireland, New Zealand and Singapore — and distilled the curial role into four principles:
- Restraint. The Court will be slow to intervene and will not read an award "with a meticulous legal eye" hunting for error.
- A high threshold. The default is minimal interference, in service of arbitral autonomy and finality. The Court deliberately preferred the language of a "high threshold" to the older "only extreme cases" formulation — labels should not do the analytical work.
- Substance over form. The test is not procedural formalism but whether the applicant can show "real unfairness or real practical injustice" — a failure of minimum standards of due process and substantive fairness. Setting aside is emphatically not a merits appeal.
- Fact-sensitivity. Whether that threshold is crossed always turns on the particular arbitration, and often on matters of degree.
Ground II — the inability to present one's case
This was the decisive ground, and the Court took it first, ahead of the jurisdictional ground — a signal that the fair-hearing question is usually where the real vice lies. The Court accepted the tribunal's own consequentials award as conclusive that the successful claim was unpleaded before the post-hearing brief, and found that the "building blocks" of that claim had never been put to the bank.
The distinction that decides most of these cases proved decisive here: this was a party deprived of the opportunity to answer a claim, not a party that simply failed to take an opportunity that existed. The bank had flagged some twenty unpleaded points in its own post-hearing brief and expressly asked the tribunal to direct a supplemental brief if it wished to entertain any of them. The tribunal upheld the claim without taking up that offer. On the causation standard, the bank did not need to prove a different outcome — only that the submissions it was denied (contributory negligence and voluntary assumption of risk) were reasonably arguable and could realistically have made a difference. The majority itself had indicated it might have been sympathetic to exactly those defences.
Ground I — the scope of the submission
Strictly academic once Ground II succeeded, but doctrinally the headline. Here the Court staked out DIFC law's position on a live international split. It declined to follow the stricter Singapore line in CAJ v CAI [2021] SGCA 102, under which an unpleaded matter falls outside the submission unless the pleadings are formally amended. It preferred the Hong Kong approach in C1 v IBS [2025] HKCFI 227 — to "look at matters in the round" across the pleadings, lists of issues, opening statements, evidence and closing submissions, and ask what was genuinely "in play".
Pleadings, the Court held, are the starting point but "not, or not necessarily, the finishing point". On the facts the indemnity was sufficiently in play that the successful claim could not safely be said to fall outside the submission — which is precisely why the fatal flaw was the absence of an opportunity to respond, not any excess of jurisdiction.
Ground III — inadequate reasons
The sleeper point, and the one with the widest reach for practitioners before the DIFC Courts. The first-instance reasons ran to roughly two pages of conclusions — "I accept Parker's submission", "I am not satisfied" — with no explanation of why. Treating the duty to give reasons as a function of due process, the Court held that the losing party must understand why it lost and an appellate court must be able to test whether the judge addressed the determinative issues and reasoned to them rationally. The judgment had also failed to engage at all with the dissenting arbitrator's reasons, on which the challenge heavily relied. The reasons were set aside as inadequate — and, tellingly, the Court suggested that the absence of the "building blocks of the reasoned judicial process" may itself explain why the judge reached a conclusion the Court of Appeal did not.
What this means in practice
For respondents facing a late claim
The bank's playbook worked, and it is worth copying: object contemporaneously, itemise every unpleaded point, and expressly ask the tribunal to direct a supplemental brief if it intends to entertain any of them. That record is what converted a lost arbitration into a successful set-aside. Silence — or substantive engagement that blurs the objection — forfeits the point.
For claimants
Do not assume an unpleaded point is safely "in play" merely because the evidence touches it; but equally do not assume a point first crystallised in closing is automatically out. Re-plead when the case shifts. The tribunal here made the claimant pay 75% of the bank's costs despite its nominal "win" — a reminder of how expensive a diffuse, late-recast case is even before any set-aside.
For tribunals
A majority's instinct to do justice on the facts does not cure a due-process gap. Where a novel theory attracts the tribunal — especially a "dramatic departure" from the pleaded case — the duty to warn is engaged: put the point to the parties and invite submissions, as The Vimeira, Zermalt and X v Y require.
For those drafting Article 41 challenges — and judgments on them
Ground III is a standing reminder that a challenge judgment must show its working. Brevity is welcome; conclusory reasoning is not. Expect first-instance decisions to identify the determinative issues, marshal the evidence on each, explain acceptance or rejection, and engage with a dissenting arbitrator's reasoning where a challenge is built upon it.
Beneath the orthodoxy of minimal interference, the message is that support is not indulgence. Where a party is genuinely shut out of the claim that decides the case, the DIFC Courts will intervene — and they will expect their own judges to explain, in reviewable terms, why they do or do not.
This note is a general commentary on a published decision and is not legal advice. For advice on a specific arbitration or enforcement question, please get in touch.