Data Center Acquisitions: Underwrite Expansion Rights

Data center acquisitions are often presented as shortcuts to scale. A buyer acquires an industrial site, an operating facility or a brownfield asset, then points to a much larger number described as future capacity. The investment case appears simple: buy a small powered footprint today and expand it into a major campus tomorrow.

The underwriting is rarely that simple.

The critical distinction is between what exists, what has been applied for, what is contractually reserved, what can be delivered, and what could eventually become IT load. Those categories are not interchangeable. Treating them as one number can inflate supply, distort valuation and create a false sense of development readiness.

A recent Swedish example illustrates the issue. On 11 September 2026, Vertical Data announced a definitive agreement to acquire an existing industrial data center site at Hamre Industrial Park in Sollefteå Municipality. The company reported 4.5 MW of existing power capacity and said an application was underway for up to 100 MW of total power capacity. It also stated that closing remained subject to financing, corporate, regulatory and other customary conditions. The announcement did not describe either figure as IT load. Read the company announcement.

This is not a criticism of the project. It is a useful reminder of how acquisition-led expansion should be analysed.

Start with the capacity that exists today

The first question is not the headline expansion number. It is what can physically and contractually be used now.

Existing capacity should be broken into utility supply, site electrical capacity and IT load. A facility described as having 4.5 MW of power may support less than 4.5 MW of IT load after mechanical systems, electrical losses, redundancy and ancillary loads are considered. Without a published IT figure, the disciplined approach is to preserve the 4.5 MW as reported site or electrical power.

This distinction is consistent with PowerlandMap's broader guidance on AI factory MW versus data center IT load. It also matters when comparing an acquisition target with competing supply. A 4.5 MW powered building is not directly comparable with a 4.5 MW IT-ready facility, and neither is equivalent to a 100 MW campus application.

The buyer should verify the metering point, connection agreement, available fault level, import limit, redundancy configuration and any conditions attached to the existing supply. The right to consume power may depend on the current industrial use, the legal entity holding the contract or a specific operating profile.

Treat a grid application as a milestone, not capacity

An application for additional power is evidence of intent and process. It is not evidence that capacity has been reserved or will be delivered on a usable date.

The underwriting team should identify the application date, requested capacity, connection voltage, responsible network company and current stage. It should then test whether studies have started, whether reinforcement is required, whether a formal offer exists and whether security or milestone payments have been made.

Each step changes the probability and timing of delivery. A project with an acknowledged application is more advanced than a concept with no grid engagement. A signed and secured connection agreement is more advanced than an application. Even a signed agreement can remain exposed to dependencies, works and long-stop dates.

The practical framework is set out in what makes a data center site power-ready and in PowerlandMap's analysis of why a grid connection is no longer the whole power strategy. The key is to record the milestone, not translate it prematurely into IT MW.

Underwrite the expansion rights behind the land

Expansion depends on more than power. The buyer needs legal and physical control of a site that can host the proposed development.

At Hamre, Sollefteå Municipality markets the industrial park as a location with existing grid infrastructure and says the wider area can provide up to 500 MW of grid capacity. That is relevant market context, but it should not be attributed automatically to one project or one parcel. See the municipality's Hamre Industrial Park information.

A buyer should therefore separate area-level potential from project-level rights. The diligence should confirm the acquired boundaries, title, easements, access, utilities, environmental constraints and the ability to phase construction. It should also determine whether adjacent land is owned, optioned, reserved or merely available in principle.

Planning must be tested against the actual development envelope. Noise, water, backup generation, visual impact, heat rejection, transport and biodiversity can constrain a technically attractive parcel. The acquisition agreement should allocate the risk if the envisaged use or expansion cannot be permitted.

This is why reserved matters and buildable supply deserve separate treatment. A broad planning principle or industrial designation is valuable, but it is not the same as an executable design.

Convert technical potential into a staged delivery case

A credible acquisition model should use stages rather than one terminal number.

Stage 1: existing usable capacity

Confirm what can be operated immediately, the condition of the building and infrastructure, and the cost of bringing it to the required standard. Record electrical capacity and IT capacity separately.

Stage 2: secured near-term expansion

Include only capacity supported by sufficiently advanced grid, land and permitting milestones. Apply a realistic programme for design, procurement, enabling works and commissioning.

Stage 3: conditional future capacity

Keep applications, wider-area grid potential and unoptioned land visible, but probability-weight them. They are strategic upside, not committed supply.

This staged method makes portfolio comparisons more useful. It also supports the data center portfolio planning framework, where budget, power, land and delivery dates are intentionally separated.

Make the acquisition price reflect conditions

The acquisition structure should mirror the evidence.

Where closing depends on financing or regulatory approvals, the buyer needs clear conditions precedent, termination rights and responsibility for interim costs. Where value depends on future power, deferred consideration or earn-outs may be more appropriate than paying today for uncertain capacity.

The model should isolate the purchase price, refurbishment, grid works, land expansion, permitting, shell and core, mechanical and electrical systems, and tenant-specific fit-out. Servers, GPUs, network equipment and storage belong in a separate technology budget unless the transaction explicitly includes them.

The same discipline applies to financing. Debt sized against operating cash flow should not be confused with capital available for speculative expansion. PowerlandMap's article on how to read the debt signal in data center financing explains why the financing instrument and its permitted use matter.

Test the customer case independently

A site can be technically expandable and still lack a bankable route to demand.

The buyer should examine the regional operator landscape, latency requirements, fibre diversity, cloud availability zones, local enterprise demand and the economics of attracting AI or HPC workloads. Claims of customer interest should be separated into enquiries, letters of intent, reserved capacity, signed contracts and energised occupancy.

An acquisition announcement may say that customer demand will be assessed after closing. That is a meaningful disclosure: demand remains part of the post-acquisition work, not proof of a committed tenant.

The PowerlandMap market intelligence workspace can help compare announced supply, development maturity, power evidence and counterparties across markets. For a direct view of the product perimeter, see coverage.

PowerlandMap view

In our view, the most valuable acquisition target is not necessarily the site with the largest expansion headline. It is the site where existing capacity, land control, grid milestones, permitting and customer strategy can be traced through a coherent delivery path.

A small operating base can be strategically attractive because it provides infrastructure, local knowledge and a route into a constrained market. But its valuation should reflect the difference between existing power and conditional expansion. An application for up to 100 MW should be recorded as an application until the relevant milestones change.

This approach does not eliminate uncertainty. It makes uncertainty explicit and comparable.

A practical acquisition checklist

Before treating an acquisition as scalable data center supply, verify:

• the current utility import limit and the capacity actually available to the site;
• reported electrical capacity versus measured or designed IT load;
• the legal holder, duration and transferability of the connection rights;
• the exact status and dependencies of every expansion application;
• land ownership, options, easements and phased-development rights;
• planning status and constraints on the intended technical design;
• refurbishment and infrastructure CAPEX, separated from IT equipment;
• credible delivery dates with dependencies and contingencies;
• customer evidence, separated by level of commitment;
• closing conditions, deferred consideration and downside protections.

The result should be a capacity stack with evidence and dates, not a single promotional number.

For investors, developers and operators assessing acquisition-led expansion, PowerlandMap provides structured market, project and power intelligence while keeping reported facts, inferred potential and unresolved diligence items separate. Request access to review how the platform supports site and portfolio qualification.

*Matthieu Gallego*