LNG supply for on-site generation

The generation analysis in [fuel supply and tolling for on-site generation](fuel-supply-and-tolling-for-onsite-generation) assumes a pipeline connection exists and asks how firmly the gas moving through it is contracted. That assumption does not always hold. Land that is otherwise well suited to a data center — available power capacity, an interconnection position in hand, room to build — is regularly land with no pipeline of adequate capacity anywhere near it, and building one is its own multi-year infrastructure project with its own permitting path, not a fuel-contracting exercise.

That is a hard stop on financeability, not a soft one. A generation entity's revenue obligation under a power services agreement does not abate because it could not obtain fuel, so a plant with no route to gas has no route to a firm output obligation either, and the strength of the site's interconnection position or its offtake contract does nothing to change that. The [what makes a site financeable](/sites/what-makes-a-site-financeable) question includes fuel access as a first-order fact about the land, not as a detail to be solved once everything else is in place.

LNG delivered by truck or rail is the answer to that specific gap, not a general alternative to pipeline supply chosen for convenience or cost. It exists to make a site fuelable where the alternative is that it cannot be fuelled at all, and the structure that follows should be read with that as the premise: this is what gets built when the ordinary path is unavailable, and it carries a different set of risks in exchange.

Where gas cannot reach the plant through a pipe, it reaches the plant as a supply chain instead — sometimes described as a virtual pipeline, though the phrase describes a series of contracts rather than any single arrangement. Reading it as a structure means reading each link as a separate counterparty relationship, because that is what it is:

• A supply source, contracting to make a quantity of LNG available, generally from a liquefaction facility with its own capacity constraints and its own queue of customers.
• Transport, moving the LNG from that source to the site by truck or rail. This is logistics capacity, contracted separately from the supply itself, and subject to its own scheduling, routing and equipment availability.
• On-site storage, which receives deliveries and holds inventory against consumption between them.
• Regasification, converting the stored liquid back to gas at the rate the generation equipment consumes it.

Each link is a distinct contract with a distinct counterparty, and each can fail independently of the others. A supply agreement that is entirely sound does not help if the transport arrangement cannot move the volume contracted, and firm transport does not help if the site's storage and regasification capacity cannot receive and process what arrives. The chain is only as strong as its weakest link, and the weakest link is rarely the one most recently negotiated.

What this structure actually does, in credit terms, is convert a pipeline-access problem into a logistics-and-storage contract. That is a real conversion — it makes an otherwise unfuelable site fuelable — and it is also a narrower one than it first appears, because logistics and storage introduce dependencies (scheduling, route availability, on-site buffer capacity) that a direct pipeline connection does not carry at all. The structure should be underwritten as what it is: a workable substitute with its own failure modes, not a like-for-like replacement for pipeline supply.

On-site storage is the one component of the chain that sits physically at the plant, under the generation entity's own control, and it does the most important work in the whole structure: it is what stands between a disruption anywhere upstream — at the supply source, in transport, at any point before delivery — and an outage in the generation entity's own performance.

That makes the amount of storage held on site a credit question rather than an operating preference. The relevant measure is not tank volume or any physical specification — those are Pantheon's territory — but storage expressed in days of run-time: how long the plant can continue to meet its availability obligation on stored inventory alone, with no further deliveries arriving. A generation entity with a firm output obligation and only a thin margin of stored days has, in substance, made its ability to perform a firm contract as fragile as its worst historical gap between deliveries, whatever the transport contract promises on paper.

Storage days should be read against the same yardstick as any other availability provision: what disruption is it sized to absorb, and does that sizing match what the delivery chain has actually shown it can fail on — a missed delivery window, a transport disruption of a given duration, a supply-source outage. A storage covenant sized against an optimistic delivery record is not a mitigant; it is an assumption wearing a number. This is the sharpest distinction the LNG structure introduces relative to pipeline supply, where the pipe itself is effectively continuous inventory and the question of buffer days barely arises.

Storage also interacts with the chain around it rather than substituting for it. Deep storage relaxes how quickly a transport failure becomes an operational one; it does not relax the requirement that transport and supply be contracted firm in the first place, for the same reason a large cash reserve does not excuse an entity from having a real revenue contract.

Delivered LNG is generally a more expensive way to receive a unit of energy than gas arriving by pipeline, once supply, transport, storage and regasification are all paid for as separate services rather than folded into a single connection charge. This page does not put a number on that difference, and no number belongs on it — the delivered cost of fuel is exactly the kind of dated, market-specific fact this analysis keeps out, and it should be established for a given site the way any commodity cost is: current, sourced, and reviewed against the actual chain contracted for it.

What matters structurally is where that cost goes. Fuel cost feeds directly into the delivered cost of the power the generation entity sells, which feeds into whether the compute the site was built to host remains economic against its alternatives over the contract's life — the same chain of consequence set out in [GPU residual value and depreciation](/compute/gpu-residual-value-and-depreciation), applied upstream from the racks to the fuel that powers them. A site that is fuelable only through a materially more expensive delivery chain can still be a sound financing; whether it is one depends on whether the power it produces stays competitive across the life of the obligation, not on whether fuel could be obtained at all.

The delivery chain also introduces a tenor question the pipeline case does not raise in the same way. LNG supply and transport arrangements are frequently shorter-dated than the generation equipment they fuel, and considerably shorter than the facility itself — logistics contracts of this kind are not typically written on multi-decade terms, where a pipeline connection, once built, effectively is. That gap matters for the same reason it matters everywhere else in this stack: an obligation that has to be renewed inside the life of the asset it supports is a repricing and re-availability event sitting in the middle of the structure, on terms nobody can fix today. The general form of that problem, and how it is tested, is set out in [what is tenor](/learn/tenor); here it applies to every link in the chain — supply, transport, storage service agreements and regasification arrangements alike — and each should be checked against the generation entity's own contracted term independently, because they will not all run out at the same time.

The security and diligence questions on an LNG-fuelled generation entity track the chain, link by link, rather than a single fuel contract:

• Supply firmness at the source. Whether the liquefaction source is contracted on a firm basis for a stated quantity, with a remedy if it under-delivers, on the same firm/interruptible distinction drawn on the pipeline side.
• Transport-leg redundancy. Whether delivery depends on a single carrier, a single route or a single mode, and what happens when any one of those is unavailable. A supply contract with no alternative path to site is a single point of failure dressed as a delivery schedule.
• Storage adequacy, read as days of run-time against the generation entity's own availability obligation and against the delivery chain's demonstrated failure modes, not against its best-case performance.
• Regasification capacity and redundancy — whether the on-site equipment converting stored LNG back to gas can sustain the plant's full draw, and what backup exists if it cannot, tested the same way any other single piece of critical equipment is tested for a redundant path.
• The credit of every counterparty in the chain, not only the supply source. A structure with four contracts has four counterparties whose non-performance can stop fuel reaching the plant, and each is examined on its own credit rather than assumed sound because the lead supply agreement is.
• Tenor across the whole chain, checked link by link against the generation entity's contracted term, per the previous section, because a chain is only as long-dated as its shortest contract.

The practical exercise is the same discipline applied on the pipeline side, extended across more counterparties: read every document in the chain against the generation entity's own output obligation, and identify every point at which a promise made on the output side depends on a promise, on the input side, that nobody has actually secured.

How an LNG-fuelled generation position behind a financeable-looking site fails:

• One link in the chain is firm and the others are not. A firm supply contract at the source with interruptible or unpriced transport, or firm transport with no committed regasification capacity at the far end. The chain's strength is set by its weakest link, and diligence that stops at the headline supply agreement misses it.
• Storage sized against an assumption rather than a record. A stated number of storage days that was never tested against the delivery chain's actual failure history — a missed truck, a route closure, a supply-source outage — is a number, not a covenant.
• Tenor mismatch across the chain. Supply, transport, storage-service and regasification arrangements running out at different times, none of them matched to the generation entity's own contracted term, so a recontracting event sits somewhere in the structure's life with no fixed date and no fixed price.
• Counterparty concentration mistaken for diversification. A chain with four contracts can still depend on very few actual counterparties once affiliates, shared logistics providers and common carriers are traced through, and a chain that looks distributed on paper can fail as a single point on the day it matters.
• Delivered cost treated as fixed rather than reviewed. A structure financed against an assumed delivered fuel cost that is not revisited as the chain, the route or the source changes carries an economic risk that was priced once and never rechecked.
• Permits or site constraints that cap storage or regasification below what the plant needs. Where the operating consent limits on-site inventory or throughput below what the availability obligation assumes, the permit governs, and the contractual obligation is the one with damages attached.
• The chain was diligenced as a single fuel contract. The security package lists one supply agreement and treats transport, storage and regasification as implementation detail beneath it. A successor stepping in inherits four counterparty relationships, not one, and the collateral is only as good as whichever of the four nobody read.

Continuum structures and arranges around generation and its fuel arrangements and coordinates the parties in them. It does not supply, market, transport, trade or hedge fuel, does not own or operate generation or storage, does not sell power, and is not a bank, a broker-dealer or a direct lender; it does not hold client funds. LNG storage, regasification and generation equipment are covered by Pantheon, not here.