Santos v Fluor: $692M and a Masterclass in How Courts Test Delay Analysis

The largest Australian construction judgment of 2025 did not come from a trial. It came from a court adopting the work of three referees — a 62-day reference hearing, a final report running to roughly 1,100 pages, and a decade of litigation compressed into a process the court itself could never have run as efficiently. Santos Limited v Fluor Australia Pty Ltd [2025] QSC 184 ended, at first instance, with Fluor ordered to pay approximately $692 million. For anyone who prepares, defends or decides delay and quantum claims, the decision repays close reading twice over: once for what it says about referee processes, and once for what it says about delay-analysis method.

One framing note before the detail: the decision is under appeal, with the Queensland Court of Appeal hearing listed for July 2026. Fluor paid the judgment sum — over $1 billion with interest — in December 2025 on terms that it may be repaid if the appeal succeeds. What follows is the law as it presently stands.

The dispute

In 2011, Santos engaged Fluor under a cost-reimbursable EPC contract to deliver the upstream coal-seam gas facilities of the Gladstone LNG (GLNG) project in the Surat Basin. The works ran from 2011 to 2014. In 2016, Santos sued, claiming more than $1.4 billion in alleged overpayments — costs it said were excluded or not properly payable under the contract — together with liquidated damages for late completion.

The scale defeated conventional trial management. In Santos Ltd v Fluor Australia Pty Ltd (No 2) [2020] QSC 373, the court referred all questions on the pleadings to referees over Fluor's objection. Three were appointed: the Hon Robert McDougall KC, the Hon Richard Chesterman KC and Michael Rudge SC. The reference consumed 171 lay witness statements, 76 expert reports, eight joint expert reports and more than five million disclosed documents. The referees' July 2023 report found Santos entitled to recover approximately $793 million (excluding GST).

Fluor applied to set the report aside — alleging apparent bias, failure to comply with the referral order, excess of jurisdiction, denial of procedural fairness and miscarriage of justice. Santos cross-applied for adoption.

What the court held

Freeburn J dismissed the set-aside application and adopted the report with limited qualifications, producing the ~$692 million order in favour of Santos and its co-venturers.

On the standard for adoption, his Honour stated the orthodox position (at [11]): the court should refrain from reconsidering disputed questions of fact "unless the referee's report reveals some error of principle, an absence or excess of jurisdiction, patent misapprehension of the evidence or perversity or manifest unreasonableness in fact-finding". Questions of law, by contrast, are reconsidered afresh. The bias complaint failed — robust criticism of a party's case is decision-making, not prejudgment — and the procedural fairness complaints failed too: referees need not address every argument, provided the fundamental reasons are explained.

The delay-methodology lesson

For practitioners, the most transportable part of the case is the treatment of delay analysis. Both parties' delay experts started from an agreed baseline — an as-planned versus as-built windows analysis — but diverged over the use of "but-for" counterfactual scenarios. The referees preferred the analysis of Santos's expert, which adhered more closely to the Society of Construction Law Delay and Disruption Protocol; Fluor complained on the set-aside application that the referees had wrongly treated relative adherence to the SCL Protocol as dispositive.

The court's treatment threads a needle worth remembering. The SCL Protocol is an accepted professional benchmark — departing from an agreed methodological baseline invites scrutiny — but it is not a contract document and it is not law. Method must be applied with common sense: on a project of real complexity, a strict "but-for" reconstruction may simply be impossible, and causation can be proved by reasoned analysis anchored in contemporaneous records. That is the same orientation Australian courts have taken since White Constructions Pty Ltd v PBS Holdings Pty Ltd [2019] NSWSC 1166: the label on the method matters less than whether the analysis fits the facts and the records.

What this means in practice

1. Treat reference out as a live procedural option in large disputes. A well-run reference resolved a 1,100-page, $1.4 billion dispute that a court could not have tried economically. But understand what you are buying: once the report lands, the grounds for unwinding factual findings are narrow. The forensic battle happens in front of the referees — not later.
2. Agree delay methodology carefully, then respect the agreement. Where experts agree a baseline method, departures and overlays become the attack surface. If a but-for scenario or other adjustment is genuinely needed, build the justification transparently from project records rather than presenting it as methodological flourish.
3. The SCL Protocol is a benchmark, not a trump card. Adherence helps credibility; mechanical adherence against the grain of the facts does not. Select the method the records can actually support — and say so expressly in the report.
4. Cost-reimbursable does not mean cost-unexaminable. Santos reopened recoverability of costs a decade after the work was done. On reimbursable and alliance-style contracts, the substantiation discipline — cost coding, exclusions mapping, approval trails — is the asset that decides these disputes years later.
5. Watch the appeal. The Court of Appeal's treatment of the adoption standard and the methodology complaints will matter well beyond this dispute. Until then, plan on the basis that referee findings of fact are, for practical purposes, final.

Key takeaways

• Santos v Fluor [2025] QSC 184 adopted a three-referee report on the GLNG dispute; Fluor was ordered to pay ~$692M (referees had assessed ~$793M ex GST). The decision is under appeal, listed for July 2026.
• Factual findings in an adopted referee report are disturbed only for error of principle, jurisdictional excess, patent misapprehension of evidence or perversity (at [11]).
• Delay analysis: agreed baselines bind in practice; the SCL Protocol is an accepted benchmark to be applied with common sense, and causation is proved from contemporaneous records.
• On cost-reimbursable contracts, substantiation discipline during delivery is what survives a decade-later audit of every dollar.

This article is general information only and is not legal advice. For advice on a specific contract or dispute, seek legal counsel or contact Sumit Consulting for commercial and claims advisory support.