A CFO signs a "managed services" agreement for a dedicated bank of cloud servers, a multi-year solar power purchase contract, and a third-party logistics deal that reserves space inside a regional warehouse. None of the three documents has the word "lease" anywhere in the title. Under ASC 842, all three may be leases — and all three may belong on the balance sheet as right-of-use assets and lease liabilities.
This is the embedded-lease problem, and for most mid-market companies it is the single largest source of post-implementation surprises. The accounting team adopted ASC 842 in 2022, ticked the obvious boxes (real estate, fleet, copiers), and moved on. The procurement team kept signing service deals. Two audits later, an auditor pulls a sample of "service contracts" and starts asking inconvenient questions.
If you have ever paged through a 60-page IT hosting agreement wondering whether you just bought a lease, this guide is for you. We will walk through the two tests that decide the question, the most common places embedded leases hide, the practical expedient that can save your sanity, and the workflow that keeps procurement, legal, and accounting from talking past each other.
What ASC 842 Actually Says
A contract is — or contains — a lease if, in exchange for consideration, it conveys the right to control the use of an identified asset for a period of time. That single sentence does most of the work. There is no requirement that the parties call it a lease, no requirement that the asset is real estate, and no requirement that the payment schedule looks like rent. If the substance fits, the form does not matter.
Two questions decide it:
- Is there an identified asset?
- Does the customer have the right to direct that asset's use and obtain substantially all of its economic benefits?
Get a "yes" on both and you have a lease — even if the contract is shelved with the IT vendors, the utility bills, or the freight invoices.
Test 1: The Identified-Asset Test
An identified asset is one that is either explicitly specified in the contract (Server SN-7842 in Rack 14 of the Reston data center) or implicitly specified because the supplier can only practically fulfill the contract using that asset. Capacity portions of a larger asset count as identified assets only if they are physically distinct — a specifically designated floor of a warehouse, for example, but not an unspecified ten percent of the supplier's bandwidth.
Even when an asset is otherwise identified, the analysis collapses if the supplier holds a substantive substitution right. To be substantive, the supplier must both have the practical ability to swap the asset throughout the period of use and economically benefit from doing so. The bar here is meaningful. Theoretical "we reserve the right to substitute" boilerplate rarely passes. If the substitute would have to be installed, certified, or commissioned by the customer — or if the supplier's substitution costs would exceed the gain — the substitution right is not substantive, and you are back to an identified asset.
Practical translation: if your IT vendor would have to send a technician on-site to swap hardware, or if your power contract names a specific solar farm and the developer cannot just re-source generation from a different facility, you almost certainly have an identified asset.
Test 2: The Right-to-Direct-Use Test
Even with an identified asset, no lease exists unless the customer obtains substantially all of the economic benefits and directs how and for what purpose the asset is used during the term. "Directing use" is usually the harder leg. It can come from:
- Deciding the output (what the asset produces, when it produces it, and where it goes)
- Deciding the operating decisions that most significantly affect economic benefits
- Designing the asset in a way that predetermines those decisions before the term begins
When the supplier makes all the day-to-day operational calls — what runs when, what mix of customers shares the asset, when maintenance happens — direction of use sits with the supplier and no lease exists. When the customer specifies the workload, the configuration, the schedule, or the output, direction shifts to the customer and a lease emerges.
A pure software-as-a-service subscription, where the vendor owns the infrastructure and decides everything about how it operates, generally does not contain a lease. A "dedicated tenancy" or "private cloud" arrangement, where you specify the server configuration, control workloads, and have exclusive use, often does.
The Four Places Embedded Leases Hide
Most of the embedded leases auditors find sit in four categories of contracts. If you have not specifically reviewed these populations, you have not finished your ASC 842 work.
IT Hosting, Colocation, and Managed-Services Agreements
This is the number-one hiding place. Look for: dedicated servers, named pieces of network equipment, reserved cages or racks in a colocation facility, dark fiber strands, or private cloud capacity built around specific hardware. The giveaway language is "dedicated," "reserved," "named instances," "private," or specific serial numbers and asset tags.
A multi-tenant SaaS subscription where the vendor moves workloads freely between data centers is not a lease. A contract that says you have exclusive use of Rack B-14 in the Ashburn facility for 36 months almost certainly is.
Logistics and Warehousing Contracts
Third-party logistics (3PL) deals routinely embed leases. Common patterns: a contract that dedicates a specific square footage inside the provider's warehouse to your inventory; a transportation agreement that reserves named tractors, trailers, or rail cars for your loads; a cold-storage contract that gives you exclusive use of a specific freezer room.
The 3PL who runs your goods through a shared facility, mixing your pallets with everyone else's and shifting capacity dynamically, is selling you a service. The 3PL who walls off 20,000 square feet just for your inventory is selling you a service plus a lease.
Power Purchase Agreements (PPAs) and Energy Contracts
Renewable PPAs are a fertile area. When a corporate buyer signs a long-term contract for the output of a specific solar farm or wind project, the analysis almost always lands on "lease." The plant is explicitly identified, the developer cannot meaningfully substitute a different facility, and the buyer is taking substantially all of the economic output. The same analysis applies to dedicated generation facilities at industrial sites and to specific take-or-pay contracts that name particular wells, plants, or transmission lines.
A grid power supply contract from a utility that draws from a pool of generating assets is not a lease — there is no identified asset. A contract to take all the output from "the 80-megawatt solar facility located at 1234 Desert Road, Mojave, CA" is.
Equipment-Embedded Service Contracts
The classic example: medical-device contracts where the hospital pays per-test or per-procedure and the equipment is delivered, maintained, and ultimately replaced by the vendor. The same structure shows up in printing-as-a-service deals, industrial gas supply with dedicated tanks on-site, vending and coffee-service arrangements with named machines, and POS-system contracts that include reserved hardware.
If the equipment is physically located at the customer's site, the customer effectively decides when and how it is used, and the vendor cannot freely swap units between customers, the equipment is leased even when the contract is structured as a per-use service fee.
Separating Lease and Non-Lease Components
Once you have identified an embedded lease, you face a new question: the contract bundles a lease (the dedicated server) with non-lease services (24x7 monitoring, patching, security). ASC 842 requires you to allocate consideration between the components based on relative standalone prices.
The lease component gets capitalized as a right-of-use asset and lease liability. The non-lease components are expensed under their normal accounting model — generally as services rendered.
Standalone prices are not always available. The colocation provider does not publish a list price for a bare rack with no managed services. In practice, controllers triangulate using vendor rate cards, comparable arrangements, internal pricing models, or — when nothing else exists — a judgment-based allocation that is documented well enough to survive an audit.
This work is genuinely tedious, which is why FASB gave you an escape hatch.
The Practical Expedient: Combine and Move On
Lessees may elect, by class of underlying asset, to combine each lease component with its associated non-lease components and account for the entire arrangement as a single lease component. The election simplifies the accounting dramatically: no standalone-price allocation, no split journal entries, no annual reconsideration.
The trade-off is real:
- Higher right-of-use asset and lease liability. The entire contract price flows to the balance sheet, not just the lease portion. That can move debt-to-equity ratios and tighten covenant headroom.
- More leases classified as finance leases. The 90-percent test for finance lease classification (where the present value of payments equals or exceeds substantially all the asset's fair value) is easier to trip when the lease consideration includes service charges.
- Lumpier P&L geography. Service costs that would have been operating expenses become amortization and interest in finance-lease scenarios.
For arrangements with relatively small service components — say, IT hardware leases with light maintenance — the expedient is usually worth electing. For arrangements where the service portion dwarfs the lease (a major managed-services contract where the dedicated hardware is incidental), the math may push you toward separating components.
Decide once per class of underlying asset and apply consistently. Document the decision before your auditors ask.
A Workable Identification Process
Most embedded-lease misses are not analytical failures. They are population failures: the accounting team never knew the contract existed. A workable process puts identification upstream, where the contracts are signed.
Build a Contract Inventory
Pull contracts from procurement, legal, IT, real estate, energy management, and operations. Focus on multi-year service agreements above a materiality threshold (often $50K–$100K in annual spend). Sort by counterparty type — IT vendors, logistics providers, energy suppliers, equipment-as-a-service providers — because patterns repeat within categories.
Use a Standard Screening Checklist
For each contract, work through six questions in order:
- Does the contract reference specific physical assets (serial numbers, addresses, named equipment, dedicated capacity)?
- If yes, can the supplier substitute a different asset, and is that right substantive (practical ability and economic benefit)?
- Does the customer obtain substantially all of the economic benefits from the asset over the term?
- Does the customer direct how and for what purpose the asset is used?
- If a lease exists, what is the lease term (including reasonably certain renewals and considering termination options)?
- What is the appropriate discount rate (the rate implicit in the lease if determinable; otherwise the lessee's incremental borrowing rate)?
A "yes" on questions 1 (without substantive substitution), 3, and 4 means you have a lease. Move to measurement.
Get Procurement and Legal in the Loop
The most durable fix is contract-level. Add an embedded-lease question to the procurement intake form so any new agreement above a threshold gets routed for accounting review before signature. Train procurement and legal to flag language like "dedicated," "reserved," "exclusive use," "named equipment," or asset-specific commitments.
When the accounting consequence is undesirable, contract changes often resolve it: adding substantive substitution rights, removing exclusivity, or moving from a dedicated-asset model to a shared-capacity model. These conversations are easier at signing than at remediation.
Reassess at Modification
Embedded-lease determinations are not one-time. When a contract is modified — added capacity, extended term, changed counterparty — reassess whether the conclusion still holds. A service contract that did not contain a lease at inception can become one when the parties amend it to dedicate specific assets.
Common Mistakes and How to Avoid Them
Treating all SaaS as non-lease. Most SaaS is not a lease, but private cloud, dedicated tenancy, and "named instance" arrangements often are. Read the entitlement language, not the marketing.
Accepting boilerplate substitution rights at face value. Suppliers love to include "may substitute equipment" language. Test whether the right is actually exercisable and economically rational for the supplier. Most are not.
Ignoring contracts below the real-estate threshold. Embedded leases hide in $30K-per-month IT hosting deals and $80K logistics agreements. The materiality threshold for in-scope leases is generally much lower than the threshold you use for, say, real estate searches.
Forgetting that the discount rate matters. A 10-year embedded lease discounted at a 4 percent incremental borrowing rate produces a very different liability than the same lease at 7 percent. Document your rate derivation; auditors will ask.
Skipping the practical-expedient decision. Failing to elect by class of underlying asset means defaulting to component separation — which is more work and more documentation, not less.
Letting embedded leases stay off-system. Once identified, embedded leases need to live in the same lease accounting system as your real estate and equipment leases. Tracking them in side spreadsheets virtually guarantees a future restatement.
Why Accurate Lease Records Matter at Tax Time
The implications run past the balance sheet. Lease classification drives book-versus-tax differences (finance leases generate interest deductions and ROU amortization; operating leases generate level rent expense), affects debt covenant calculations under most modern credit agreements, and shows up in EBITDA-based valuation multiples. An embedded lease found late may trigger a material weakness disclosure, restate prior-period financials, and force renegotiation of covenant calculations with lenders.
Clean, complete lease records also matter for the next migration. The FASB has signaled continued attention to leases, and any successor standard will start from your current lease population. Companies with disciplined inventories adopt easily; companies whose lease registers are best-guess approximations spend the next adoption cycle reconstructing history.
Keep Your Lease and Service Contracts Auditable from Day One
Embedded-lease identification only works when your underlying records — vendor contracts, payment streams, modifications, classification decisions — are organized, queryable, and version-controlled. Beancount.io provides plain-text accounting that gives you complete transparency over every transaction and supporting note: lease payments tagged by contract, classification decisions documented inline, and a full git history of every change for auditor review. Get started for free and see why controllers and CFOs are switching to plain-text accounting that scales from a single subsidiary to a multi-entity consolidation.