Bridge to energization: short-dated capital to firm power

Between the moment a site is controlled and the moment it can serve load, a project consumes capital and produces nothing. Land is carried, options are exercised or extended, deposits are paid, studies are commissioned, permits are pursued, long-lead equipment is ordered. None of it generates revenue, and most of the capital that will eventually fund the project cannot arrive yet, because the thing it underwrites does not exist.

That gap is a financing problem in its own right, and it is the reason a distinct instrument exists for it. The bridge is not a smaller version of the permanent capital. It is a different structure with a different security package, a different tenor logic and a different exit, and it is priced and negotiated as such.

What makes it particular to this sector is the size of the gap. In most real-asset development the pre-revenue period is a permitting and construction question, and both are broadly within the sponsor's control. Here the binding constraint is frequently energization, and energization sits with a utility, a system operator or an equipment supplier's delivery schedule. The sponsor can do everything correctly and still not move the date — which is the condition speed-to-power describes and this instrument is built to survive.

The consequence is that a bridge here is underwritten less on the asset and more on a schedule. The questions are what has to happen, who has to do it, what evidence exists that it will happen when claimed, and what happens to the capital if it happens later. Everything below follows from that.

Before energization, a data-center bridge loan is usually a controlled-draw facility rather than a single advance against a finished asset. At closing, the lender establishes which pre-revenue costs the facility may fund and which have to be paid with sponsor equity. Draws then follow evidence that the project has preserved or advanced the position: site-control payments have been made, utility deposits are due, permits or studies have reached the stated stage, and the borrower remains on the milestone path that supports the expected exit.

The borrowing base, where the parties use one, is not a simple percentage of completed value. It may give different recognition to land value, refundable deposits, eligible development costs and equipment payments, while giving little or no value to soft costs that a successor could not use. A cost being necessary does not make it good collateral. That is why the sources-and-uses schedule and the collateral schedule should be read separately.

Interest and fees also need a source before the project produces revenue. They may be paid currently by the sponsor, reserved at closing or capitalised into the balance, subject to the facility's terms. Each route changes the amount of runway. Capitalising interest preserves cash but causes the claim to grow while the collateral remains a development position; a reserve solves only the period it was sized to cover.

The operating cadence is milestone-driven. The lender receives updated utility or operator correspondence, evidence that deposits and site obligations are current, budget-to-actual reporting, and notice of any movement in energization or takeout conditions. A delayed date can reduce availability before maturity if the remaining commitment is no longer enough to carry the project to exit. A bridge can therefore tighten at exactly the moment the project needs more time.

The final draw does not repay the bridge. Repayment occurs when the agreed takeout funds, the project is sold, or another permitted exit closes. Energization may be the milestone that makes that exit possible, but the documents should not treat the physical event and the cash repayment as though they occur simultaneously.

The security package is assembled from what exists at the time, which is not much: a site, a position, some paper, and possibly some deposits on equipment that has not been delivered.

In practice it is built from some combination of:

• The land, by mortgage or charge, or where control is contractual, an assignment of the option or lease. This is the only conventional real-asset collateral in the package.
• The entity holding the power position, generally by equity pledge rather than directly — the position itself usually cannot be assigned without consent, which is the point worked through in who holds the interconnection position.
• The permits, studies and development work product. Of limited standalone value, but the difference between a site a successor can continue and one that has to start again.
• Deposits and progress payments on long-lead equipment, together with the benefit of the supply contracts they sit under.
• Sponsor support, in whatever form is negotiated. On a pre-revenue position this frequently does more work than the asset collateral.

The honest reading of that package is that it is a position in a development, not a claim on cash flow. Recovery in a downside is a sale of the site to somebody who wants to finish it, which is only attractive while the power position remains intact — and the power position is precisely the thing that a stalled project puts at risk, because deposits go unpaid and milestones get missed. The collateral and the risk are the same object. That correlation is the defining feature of the instrument, and it is why the takeout matters more than the security does.

The factory schedule can be as important as the utility schedule. US Department of Energy data show that distribution-transformer demand and lead times moved sharply above their pre-2020 levels, with large substation and generator-step-up units reported on a multi-year clock. (US Department of Energy, Office of Electricity, as of September 18, 2026) That matters to a bridge because a project may have to pay to preserve equipment availability before ordinary construction capital is ready to fund. The interest reserve, commitment period and maturity then need to cover not only a utility date, but also the interval between a supplier payment and usable equipment on site.

OEM disclosures confirm both the scarcity and the financing consequence. Siemens Energy reported a €51 billion Grid Technologies backlog and customer advance payments that included reservation fees. (Siemens Energy Q3 FY2026 earnings release, as of September 18, 2026) GE Vernova separately reported a large Electrification equipment backlog and, in Gas Power, distinguished new slot reservations from reservations converted into orders and from units shipped. (GE Vernova Q2 2026 earnings release filed with the SEC, as of September 18, 2026) That sequence is the useful point. A reservation, a firm purchase order, work in progress, delivered equipment and commissioned equipment are different assets, even when a market graphic groups them under one backlog theme.

A bridge should therefore attach different evidence and collateral value to each stage. A refundable reservation may preserve optionality but offer little recovery. A non-refundable deposit is a real use of funds, yet still may not give the project title to equipment or the right to transfer the slot. A firm order may improve certainty while introducing cancellation liability. Work in progress becomes better collateral only if ownership, identification, insurance, inspection and assignment are documented. Delivery still leaves installation, testing and acceptance before the equipment contributes to energization.

The practical response is not to assume that a large OEM backlog guarantees delay or pricing power on a particular order. It is to read the supplier paper: refundability, milestone payments, price escalation, title, storage, transport, warranty, cancellation, assignment and the consequences if the site's utility schedule slips. Those provisions decide whether an equipment draw preserves a valuable position or merely converts bridge cash into a claim that cannot travel with the project.

Every other structure in the stack is sized against how long an asset lasts. This one is sized against when something happens, and that difference drives the whole negotiation.

The date being bridged to is usually one of a small set: energization from the grid; commissioning of on-site generation that removes the dependence on the queue; completion of a building sufficient to draw a takeout; or the commencement date under a hosting or compute contract that triggers revenue. Which one it is has to be stated precisely, because the whole structure is calibrated to it.

Three things then have to be established, and they are the substance of the transaction:

How firm is the date? A date derived from an executed agreement with defined obligations on a counterparty is a different input from a date derived from a study, a forecast, or a queue position whose treatment is at the operator's discretion. The document behind the date is read for the same reason the document behind a power claim is read.

What margin sits between the bridge's own maturity and that date? A bridge maturing on the projected energization date has no margin at all. Anything that slips — and something usually does — lands directly on the borrower. Margin costs money and is worth it.

What are the extension mechanics? Whether extension is available at the borrower's option on stated conditions, or requires a fresh credit decision, is the difference between a schedule risk that was allocated and one that will be renegotiated under pressure. Extension terms agreed at the outset are worth more than optimism about the date.

The general discipline: a bridge should be sized against the date the project can defend on paper, not the date it hopes for, and the difference between those two should be visible to everyone signing.

A bridge is repaid by something else. What that something else is, and how conditional it is, is the most important term in the transaction — more important than the security, and more important than the tenor.

The distinction worth drawing is between a takeout that is committed and one that is available. A committed takeout is a facility already documented, with conditions the project expects to satisfy, held by a party obliged to fund on satisfaction. An available takeout is a market — capital that exists, would plausibly be interested, and is under no obligation whatsoever. Bridges are routinely written against the second while being discussed as though they were written against the first.

The test that separates them is mechanical: read the takeout's conditions precedent alongside the bridge's maturity date, and ask whether every condition can be satisfied before that date by parties who are obliged to act. Where a condition depends on a third party's schedule, or on a document nobody has started, the takeout is a plan rather than an exit.

The possible takeouts, and what each requires to be real:

The instrument fails in recognisable ways, and most of them are visible at signing to anyone looking for them.

• The bridge to nowhere. No committed or credibly available takeout, on the theory that one will exist by the time it is needed. This is the dominant failure mode, and it converts a schedule risk into a refinancing risk in a market that may not be open.
• The date was never the sponsor's to give. A maturity set against an energization date controlled by a utility, a system operator or an equipment delivery slot. The sponsor bears a risk it cannot manage, and the only lever left is an extension it has to ask for.
• The takeout's conditions are not the bridge's milestones. A takeout conditioned on completion and occupancy, bridged by an instrument that matures at energization, leaves a gap nobody funded. The two sets of conditions have to be read against each other line by line.
• Carrying cost outruns the position. A site with no revenue accrues obligations. Where the bridge capitalises them, the balance grows against an asset whose value has not moved, and the terms that looked comfortable at signing do not look comfortable at maturity.
• The collateral decays with the delay. Missed milestones and unpaid deposits can put the power position itself at risk. The one asset the package depends on is the one the delay erodes.
• Contracted demand with a commencement date that slips. A customer contract with a firm start date, bridged by a facility that does not deliver power in time, can hand the customer a termination right. The financing gap becomes a contract gap.

Continuum structures and arranges around these positions and coordinates the parties in them. It is not a bank, not a broker-dealer and not a direct lender; it does not provide bridge capital or hold client funds.