Equipment finance for turbines and gensets

Generation equipment is unusually good collateral by the standards of a data-center stack, and understanding why explains most of how the paper is written.

It is long-lived. Where accelerators turn over on a product cycle, generating plant works for decades with maintenance. The residual after a normal financing term is real rather than nominal, which is what allows structures that do not fully amortise and lease structures where a lessor genuinely underwrites a return value.

It is movable, in a way buildings are not. Smaller and mobile units in particular can be relocated and redeployed. That gives a financier an alternative to the single site, which is worth a great deal on an asset class whose primary risk is that one project fails. Larger installed plant is far less portable, and the difference is a genuine credit distinction rather than a technical one.

It is slow to arrive. Manufacturing lead times on this equipment are long, and the consequence for a financier is that a delivered, commissioned unit is worth more than an equivalent order slot to a buyer that needs power sooner than the slot allows. That supports secondary value in a way that is unusual for capital equipment.

It has a functioning market. Units are traded, remarketed, leased and moved between projects by parties who do that for a living, which is what makes a residual position underwritable at all.

Against that, three characteristics pull the other way, and they are what the structure has to answer: the equipment takes a long time to arrive, it usually has to be paid for substantially before it does, and its value in use depends on a site, a fuel supply and permits that may not travel with it. The rest of this page is about how paper is written around those three facts.

What is deliberately not here is anything about how the equipment performs. Specification, selection and engineering belong to the hardware side of the fleet; this page is about who holds the unit, for how long, against what credit.

The defining structural problem is that money moves long before collateral exists. Long-lead equipment is ordered against a schedule of progress payments, and for the entire period between the first payment and delivery there is a substantial amount advanced against a manufacturing slot rather than a machine.

What capital is exposed to changes at each stage, and so does what mitigates it.

A workable package over generation equipment is assembled from more than the units, because the units alone are worth less than the working plant they form part of.

• The equipment itself, identified at serial level and maintained as a live schedule. Components are replaced, spares are rotated, and a schedule captured once at closing describes an estate that no longer exists.
• The supply contract, assigned. It carries the delivery obligation, the warranties, the performance and availability commitments and any refund rights, and in the pre-delivery period it is frequently the only thing there is to take.
• Site rights sufficient to reach and remove the equipment. A financier that cannot enter the site cannot inspect and cannot recover, and the plant sits on land that may be controlled by another party in the structure.
• The service and maintenance arrangements, including long-term service agreements. Generation equipment without a maintenance regime is a depreciating liability, and the ability of a successor to keep the service arrangement in place is part of what makes the collateral worth anything.
• Spares and consumables, which are frequently a material part of the estate and are frequently forgotten in the schedules.
• Insurance, noted with the interested parties named, covering transit and construction as well as operation. Transit and installation are when physical loss is most likely and are the periods most often left uncovered by an operating policy.
• The revenue contract, where the units are supplying a facility under a power services agreement. The equipment and the cash flow it generates should not be separable at enforcement.

Where the equipment is leased rather than financed, the same list applies from the other side. A lessor holds title and needs the access, service and site rights that make its return position real — and it should establish, before the units are shipped, whether the site arrangements permit equipment to change hands at all.

On this layer, the covenant that does the most work is not a financial ratio. It is availability — how much of the time the plant is capable of producing at contracted output.

The reason is direct. Where the units are supplying a facility under a contract, the revenue that repays the financing depends on the plant performing to a stated standard. A shortfall is not a performance inconvenience; it is a reduction in the cash flow the structure was sized against, and it may also be a default under the supply agreement with its own damages attached.

That makes several ordinarily technical documents into credit documents:

The service agreement. Who maintains the plant, to what standard, with what response times, and whether the arrangement survives a change of ownership. A successor that cannot inherit the service arrangement inherits an asset it cannot keep running.

Spares and parts availability. A unit awaiting a component is unavailable, and on scarce equipment that wait can be long. Contracted access to spares is worth more than it appears on a schedule of assets.

Fuel. Supply and delivery arrangements have to run at least as long as the availability obligation. A plant contractually obliged to be available and commercially unable to obtain fuel is exposed on both sides at once.

Permits. Operating consents can restrict how and when a plant may run. Where the availability commitment assumes operation the permit does not allow, the commitment is the one that carries damages.

Metering and reporting. Whether availability is measured, by whom, and on what basis. An obligation nobody measures is an obligation that will be argued about at exactly the wrong moment.

The general point is that on layer 3 the boundary between operations and credit is thin. A financier underwriting generation equipment is underwriting a working plant, and every arrangement that keeps it working is part of the analysis.

The recurring ways equipment paper on this layer disappoints:

• Delivery slips and everything downstream slips with it. The unit is the critical path for energization, so a delay is not contained within the equipment financing; it moves the conversion test on the building's facility and the commencement date under customer contracts. This is why the [bridge to energization](bridge-to-energization) exists as a separate instrument.
• Pre-delivery exposure was never properly secured. Substantial payments made against a supply contract with no title-passing mechanism, no refund arrangement, and no security over the contract itself. This is the largest unmitigated exposure in most transactions on this layer.
• The units arrive and the site is not ready. Equipment delivered ahead of civil works, permits or fuel connections sits in storage, uninsured for the right risks and unavailable to generate anything.
• Orphaned units. Equipment ordered for a project that does not proceed. Its value then depends entirely on whether it can be redeployed or sold, which is a question about portability and market demand rather than about the paper.
• The service arrangement does not transfer. A successor takes the machines and not the maintenance regime, and the availability the collateral value assumed is no longer achievable.
• The equipment is financed on the facility's tenor. Bundled into the building's capital, a twenty-year asset is funded on a term set by the shortest thing in the structure. That is the [tenor mismatch](/compute/tenor-mismatch-compute-and-infrastructure) argument at layer 3, and it is the most common structuring error here.

Continuum structures and arranges equipment transactions and coordinates the parties in them. It does not supply, own, lease or operate equipment, is not a bank, a broker-dealer or a direct lender, and does not hold client funds.