Where engineering ends and the contract begins

The covered story is volumes and prices. The under-covered one is a contractual boundary: when does a plant stop being "in commissioning" and become "commercial"? That switch — the declaration of the commercial operation date (COD) — is where billions get captured, because the seller largely controls the declaration.

Venture Global is the visible case of a recurring class. Against Shell, an ICC tribunal read the contract literally and VG won — commissioning cargoes fell outside the supply obligation as drafted. Against BP, a tribunal applied the "reasonable and prudent operator" standard and VG lost — it should have declared COD on time. Same facts, two readings. That boundary is being redrawn across three arenas that do not converge: arbitration (the contract), FERC (the regulator), and the equity market (pricing the contingent liability).

The boundary — where the engineering record decides a legal question — is the thread worth pulling.

The timeline — the events laid over realized price

Plot the event sequence against realized price and the timing question becomes visible. War opens the spread (Feb 2022); VG sells 400+ commissioning cargoes at spot — earning more than $19B in spot LNG sales over three years, almost entirely from Calcasieu (Reuters, Apr 2025) — while it stays wide; buyers file ICC arbitration (2023); the spread collapses (2024, Henry Hub at a real low); and Calcasieu's COD is declared April 2025 — near the trough, three months after the IPO. FERC's commissioning deadline — 31 Dec 2027 (extension granted Oct 2025) — bounds how long the window can stay open. After the effective closure of the Strait of Hormuz — legally open, but with tanker traffic collapsed and transit conditioned by Iran — Asian spot peaked above $20/MMBtu (~$21.65 average, March 2026) and held elevated through Q2; the spread, not the price level, is where the rent lives.

The window is the distance between when a facility appears to have moved past a transient testing regime and when COD was formally declared. A facility shipping 400+ cargoes over multiple years begins to look less like testing and more like sustained commercial operation — though the contractual threshold is a separate question. Whether the timing reflected economics, technical constraints, or both is an evidentiary question, not a rhetorical one — and the answer isn't in either party's narrative but in the engineering record: commissioning logs, performance-test dates, the repair timeline against the production curve.

The governance asymmetry. The deeper issue is not whether COD was early or late — it is governance. "Completion" had four parallel definitions and five independent verifiers: module FAT (Baker Hughes), mechanical completion and performance tests (the EPC — Kiewit at Calcasieu, KBR at Plaquemines), the lenders' Independent Engineer, and FERC. The one party whose 20-year contract turned on the milestone — the offtaker — had no seat at that table and no audit rights over it. The lender and regulator sign-offs that COD can hinge on are themselves sequenced by the operator, so a "pending sign-off" is not, on its face, an external constraint. Every major stakeholder held verification rights over completion except the one whose economics depended on it. That asymmetry is observable, contractual and replicable — and it is the question for any future SPA.

Honest caveat. Correlation isn't causation — the full treatment of what muddies the read is in Confidence & limits below; the point here is only that the sequence is verifiable by dates, not by narrative.

Plaquemines currently exhibits a similar economic and contractual pattern: 92 commissioning cargoes in Q1-2026, COD deferred — Phase 1 now targeted for Q4 2026 and Phase 2 for mid-2027, with FERC's 15-month extension moving the full-service deadline to 31 Dec 2027 — and a DOE capacity uprate four days after the Hormuz shock. The economic outcome resembles Calcasieu's; whether that reflects strategy, circumstance or both is what the litigation is now testing.

What changed is the legal exposure, not the underlying economic pattern. The BP award (October 2025) put a price on deferred COD: BP seeks in excess of $1B (ICC partial award, 8 Oct 2025) — and the foundation buyers' combined claims reached ~$7B at their peak (VG 3Q2025 disclosure; since guided to $4.8–5.5B), while VG has disclosed the contract's liability cap likely won't apply. The BP damages hearing is anticipated in 2026 — not yet scheduled — and VG has stated its arbitration costs are "fully accounted for" in guidance; the quantum will set the reference for Plaquemines' buyers, while Galp's claim (alongside Edison's and Repsol's) remains pending. Calcasieu's buyers had to wait years for a precedent; Plaquemines' buyers already have one.

The shock re-prices the switch

The Hormuz shock closes the loop. By reopening the price spread that makes an early-versus- late COD worth billions, it re-prices the same switch across every commissioning window still open — Plaquemines' included. The instrument is not historical; it is live, and it is being re-valued in real time. (The shock also redrew the shipping map that carries the gas — a chokepoint-and-freight story in its own right, treated in a companion field note.)

What this means — by seat at the table

Every seat reads the same instrument differently. Reading it well is the difference between watching the dispute and acting on it.

  • Buyers · offtakers. Write an objective COD trigger and audit rights into the SPA — completion tests, not the seller's declaration. The BP award shows the liability cap may not hold; the precedent is now yours to cite.
  • Lenders · Independent Engineers. Your sign-off can be the hinge COD turns on — and the operator sequences it. Time-stamp the independent record so a "pending sign-off" cannot double as a commercial lever.
  • Investors. Price the contingent-liability tail, not just the ramp: >$1B on BP alone, ~$7B in foundation claims at peak (VG 3Q2025), the cap likely disapplied. That quantum sets the reference for Plaquemines.

Confidence & limits

Correlation isn't causation. Earnings misses, the FID cycle and the macro all muddy the timing read — but the sequence fits spread economics better than any documented technical constraint, and it is verifiable by dates. The one read this piece does not make is intent: whether a given COD reflected economics, technical constraints or both is what the arbitration is testing, not something this analysis resolves. The structural reads — the controllable switch, the governance asymmetry, the live re-pricing — do not depend on it.

References

Selected, not exhaustive — key figures are attributed inline. Figures refreshed July 2026.

  • Arbitration & litigation. International Chamber of Commerce (2025), partial awards: Shell (7 Aug 2025); BP (8 Oct 2025) — BP seeking damages in excess of $1B. U.S. District Court, Southern District of New York (2025), related enforcement proceedings.
  • Corporate disclosure. Venture Global, Inc. (2024–2025), Forms 8-K and 10-Q, Q1-2026 results (92 Plaquemines commissioning cargoes), and 3Q2025 disclosure (customer claims peaked ~$7B, since guided to $4.8–5.5B); U.S. Securities and Exchange Commission.
  • Spot-sales revenue. Reuters / Yahoo Finance (Apr 2025), "New LNG plant to boost Venture Global profits" — >$19B earned in spot LNG sales over three years, almost entirely from Calcasieu Pass.
  • Regulatory. Federal Energy Regulatory Commission (2025), commissioning-deadline extension (17 Oct 2025); U.S. Department of Energy (2025), LNG export authorizations.
  • Supply & prices. International Energy Agency (2026); U.S. Energy Information Administration (2026); LSEG (Asian spot / Henry Hub price series — Asian spot $21.65/MMBtu, Mar 2026).