Petronet Dahej case study: why 66% utilisation is not a 26-point demand collapse

Start with the change in the asset, not the percentage

Petronet disclosed that additional Dahej regasification facilities were commissioned at 23:30 on 31 March 2026, expanding nameplate capacity from 17.5 to 22.5 MMTPA. Its subsequent Q1 FY2026-27 results explicitly state that utilisation uses the new 22.5 MMTPA base. A researcher comparing the new percentage with an earlier period must therefore account for a denominator change, not simply subtract the two reported utilisation percentages and call the result lost demand. [S4] [S3]

This is a particularly useful public case because the source identifies the change directly. It removes the need to guess whether an apparent utilisation deterioration reflects lower flow, a larger asset base or both. The remaining work is analytical: reconstruct the physical change and show what each comparison means. No confidential operating information or customer relationship is required to demonstrate the method.

The reported operating observations

The year-on-year physical-energy change is 192/207−1, or approximately−7.25%. The quarter-on-quarter change is 192/201−1, or approximately−4.48%. Neither calculation equals the 26 percentage-point difference between reported 92% and 66%. A percentage-point change in utilisation is not a percentage decline in throughput, even before an expansion is involved. Treating the two as interchangeable would overstate the physical deterioration.

A simple denominator normalisation

One way to expose the denominator effect is to keep the reported current numerator implicit and rescale the capacity base. The result is approximately 84.9% on the old denominator, rather than 66%. Another way is to hold the old 92% utilisation constant but expand its denominator:92%×17.5/22.5 gives about 71.6%. Under that mechanical comparison, roughly 20.4 percentage points of the original 26-pointgap arise from the larger capacity base before considering the physical-volume change.

This decomposition is approximate. Reported percentages are rounded, quarters differ in calendar days, and the precise company annualisation may use details not reproduced in the press release. The exact energy-based throughput change is therefore the cleaner public observation. The normalisation helps explain the direction and scale of the denominator effect; it should not be presented as a formally audited utilisation restatement.

What the data does not establish about the business

Lower physical throughput deserves investigation, but it does not by itself establish customer loss, underperformance against contracted quantities or a permanent impairment of expansion economics. Those questions require information about bookings, nominations, cargo timing, maintenance, network evacuation and demand conditions. In its February 2026 conference-call transcript, management discussed downstream evacuation as an important aspect of terminal performance. That is useful company commentary to investigate, not independent proof of the cause of the later quarter’s volumes. [S9]

Similarly, company earnings and terminal throughput are not one interchangeable measure. Petronet’s Q1 release reports higher standalone profit year on year despite the lower Dahej physical volume. This is evidence that a simple volume-only earnings narrative is incomplete; it does not establish which commercial mechanism caused the result. A proper earnings bridge would need the relevant financial disclosures and contractual/accounting context. [S3]

Apply the lesson to other infrastructure expansions

For a B2B research service, the value lies in catching the misleading inference before it reaches an investment or procurement committee. A labelled series, a commissioning event and two transparent calculations can change the discussion from apparent collapse to a more precise question about ramp-up and physical demand. That does not remove commercial risk. It ensures that the risk is being described with the right denominator and the right evidence.