Sale-leaseback of a powered shell

TL;DR

A sale-leaseback converts a completed building into capital without giving up its use: the owner sells the envelope, and often the land under it, to an investor and takes back a lease. On a powered shell the instrument is attractive because the asset is long-lived and the power position makes it scarce. What the buyer is really acquiring, though, is the lease covenant — so the seller's own credit does more of the pricing than the building does. The transaction fails on the seams: a power position held in a different entity, transfer restrictions in existing documents, and lender consents that were never sought.

What the structure does, and for whom

A sale-leaseback is a single transaction with two halves. The owner sells the asset to an investor and simultaneously takes a lease back over it, so occupation and use continue uninterrupted while ownership moves.

The economic effect is to release the capital held inside a completed building. A developer that has funded a shell through construction is holding a long-lived asset with a large amount of its own money in it, and that money is not available for the next site. A sale-leaseback returns it, at the cost of a long-term rental obligation and the residual position in the building.

On layer 2 this is a mature, ordinary instrument. What makes the data-center version distinctive is not the mechanics but the asset. A shell with secured power is scarce in a way a warehouse is not, and its scarcity comes from a position that may not sit in the same entity as the bricks. The transaction therefore has to establish something a conventional property sale-leaseback never asks: whether the thing that makes the building valuable is actually included in the sale.

For the buyer, the attraction is a long-dated, contracted income stream secured on a real asset in a sector with structural demand. For the seller, it is capital recycling plus, frequently, a change in what the balance sheet looks like — though that should be treated as a consequence rather than a reason, because the accounting outcome depends on how the transaction is structured and is not guaranteed by calling it a sale.

It is also a live alternative at the end of a build. A completed envelope that cannot convert to permanent financing because the tenant position is not yet settled is still a saleable asset, which is why this route and [construction-to-permanent](construction-to-permanent-for-a-data-center) are usually considered together rather than sequentially.

What transfers, and what does not

The most productive early exercise is a line-by-line inventory of the stack at that address, because a data-center address contains several assets and a sale-leaseback does not automatically carry all of them.

The buyer is underwriting the covenant

This is the point that most changes how a seller should prepare. In a sale-leaseback the investor's return comes from rent, and the rent comes from the lessee. The building matters, but it matters as the fallback rather than as the primary case.

What follows from that:

The lessee's credit is priced into the transaction. A well-capitalised operator with contracted revenue supports different terms from a development company whose income depends on filling the building. Sellers who present the asset in detail and the entity behind the lease in outline have prepared the wrong package.

The lease is the product. Term, renewal options, rent-review mechanics, repair and insurance obligations, permitted use, alienation, and what happens on default are the substance of what is being bought. In the net-lease shape common to this asset class, the lessee carries most of the operating obligations and the investor holds a comparatively passive income position — which is exactly what makes the covenant so load-bearing.

Guarantees follow the credit. Where the lessee is a project-level entity, the investor will generally want the obligation supported by the party with the resources. That is ordinary, and it partially re-couples entities the structure had separated, which is worth understanding before it is conceded.

The residual is genuinely long-dated. Unlike a sale-leaseback of installed compute — where the residual question is how quickly the equipment loses value, worked through in [GPU lease structures](/compute/gpu-lease-structures) — the residual here is a building with a life measured against a generation. The investor's terminal position is a specialised property whose value depends on whether it can be re-let, which depends in turn on whether the power stays with it.

The practical consequence for a seller is that the negotiation is less about the building's replacement cost and more about how firm the income is and how long it runs. Two identical shells with different lessees are two different transactions.

The seams that block the sale

Sale-leasebacks of specialised assets fail on consents and on documents nobody re-read, and on this asset class there is a specific list.

  • The power position sits in a different entity. A shell sold without the interconnection position behind it is a building with an electrical connection whose continuation depends on an affiliate of the seller. Whether that is acceptable depends entirely on what contractually binds the two, and the buyer will read it. The entity question is set out in [who holds the interconnection position](who-holds-the-interconnection-position).
  • The interconnection or supply documents restrict transfer. Consent requirements and change-of-control provisions apply to a sale of the property or of the entity holding it. A transaction structured as an entity sale specifically to avoid an assignment can still trip a change-of-control clause.
  • Existing security has to be discharged. A shell funded through construction carries a mortgage or charge. Release, and the timing of it against completion of the sale, is mechanical but it is a gating item rather than an afterthought.
  • Ground lease terms. Where the land is leasehold, the head landlord's consent and the residual term of the head lease both constrain what can be sold and for how long.
  • Hosting agreements with occupiers. Customers occupying the building may have consent rights, quiet-enjoyment protections, or termination rights triggered by a change of ownership. These are read from the occupier's side in [the hosting agreement as a financeable contract](hosting-agreement-as-a-financeable-contract).
  • Equipment owned by third parties. Compute and generation equipment inside the building may belong to lessors or be pledged to financiers. Their access rights and their acknowledgements have to survive the sale, or the sale impairs a different party's security and will be resisted.
  • Recharacterisation. Where the arrangement functions economically as a financing rather than a sale — a nominal repurchase right, a term covering substantially the whole life of the asset — it may be treated as one, with consequences for accounting, tax and enforcement. That is a question for the parties' own advisers, and it should be asked before signing rather than after.

When it is the right instrument, and when it is not

The structure fits a specific position and is a poor answer outside it.

It fits where a completed, energised or credibly energisable envelope holds capital the owner wants back; where the owner intends to keep operating the facility rather than exit it; where the income supporting the lease is contracted or credibly contractable; and where the power position can move with the building or be bound to it durably enough that a buyer will accept the arrangement.

It fits particularly well as a way of retrofitting the layer separation onto a project that was built without it. A developer that funded land, shell and power out of one pool can use a sale-leaseback to place the long-lived layers with capital that wants long-lived assets, keep the operating position, and free capital for the layers that turn over faster. That is the same logic the whole stack argument rests on, applied after the fact — see [the data-center capital stack](/learn/the-data-center-capital-stack).

It does not fit where the seller's real problem is that the building has no income. A sale-leaseback does not create a covenant; it monetises one. An empty shell with a weak lessee produces a transaction on terms that reflect exactly that, and sellers are frequently surprised by how much of the pricing is about them rather than about the asset.

It also does not fit where the seller wants to exit. The instrument exists to keep the asset in use by the seller. A party that wants out of the building should sell the building, and the lease obligation it would otherwise take back is a liability it does not need.

The test worth applying before starting: if the lease were stripped out and only the building sold, what would change about the price? Where the answer is "most of it", the transaction is a credit sale wearing a property label, and it should be prepared as one.

Continuum structures and arranges these transactions and coordinates the parties in them. It is not a bank, not a broker-dealer, not a direct lender and not a principal in any asset; it does not hold client funds.

Frequently asked

Does the buyer take the power position too?

Only if it sits in what is being sold and is transferable. This is the difference between a sale-leaseback of a powered shell and a sale-leaseback of a shell. Where the position is held by an affiliate of the seller, or by a separate generation entity, the buyer is acquiring a building whose supply depends on a contract with the seller's group — which may be perfectly acceptable, provided the contract is durable, its term matches the lease, and it survives both a default by the lessee and a change of ownership on either side.

Is this a way to finance a building that has no tenant?

Not really, and treating it that way produces a disappointing process. The investor's income is the lease, so the transaction prices the lessee's covenant. An owner-operator with contracted hosting revenue behind it is presenting a genuine income stream; a development entity with an empty building is asking an investor to underwrite a development position in the shape of a lease. The instrument monetises a covenant that already exists rather than substituting for one that does not.

How does this differ from a sale-leaseback of compute equipment?

The mechanics rhyme and the residual behaves completely differently. Compute loses value on a product cycle, so the whole negotiation is about who carries obsolescence risk, and terms are short relative to the asset. A shell lasts a generation, so the buyer's terminal position is a specialised building whose re-lettability depends on the power staying with it. The other structural difference is consent: equipment sale-leasebacks turn on hosting access and existing security, property ones turn on land documents, mortgage releases and occupier rights.

What term does a buyer want on the lease?

Long enough that the income supports the price without depending on a re-letting, which in practice means a term set against the investor's holding horizon rather than against the seller's operating plans. The specific answer is a negotiation and varies by counterparty, but the direction is consistent: the shorter the firm term, the more of the value rests on the residual, and the residual on a specialised building is exactly the thing an investor is least willing to underwrite. Renewal options are not a substitute, because an option the lessee may decline is not contracted income.

Considering a site, a power position, or the capital behind it? Speak with our team.

Submit a transaction for review

Last updated

Continuum Capital is not a bank, not a broker-dealer, and not a direct lender. It acts as arranger and advisor: it structures and arranges capital, does not execute securities transactions, and does not hold client funds. This page is informational and is neither an offer to sell nor a solicitation of an offer to buy any security, nor a commitment to provide financing.

All structures