Investment & Technical PaperPublished · 9 September 2026

Data Centre Technical Due Diligence: What Investors Should Test Before Capital Is Committed

Testing the investment case, not confirming that documents exist.

  • Published 9 September 2026
  • 15 pages
  • Written for investors, lenders, developers and operators
Abstract

Technical due diligence is not a checklist confirming that drawings, equipment and permits exist. Its purpose is to test whether the asset, site and delivery plan can support the investment case at the required capacity, resilience, programme, cost and operational standard. A technically feasible project can still be a poor investment when critical dependencies remain unproven.

Market signals

Strong capital appetite raises the standard of evidence.

95%

Investor intentions

of surveyed investors expected to increase or maintain data centre investment activity.

CBRE 2025 Global Data Center Investor Intentions Survey

Capacity already committed

57%

of data centre capacity under construction is already pre-leased, which moves diligence towards deliverability and handover quality.

JLL 2026 Global Data Center Outlook

Uptime Institute

Continues to report that power and human or procedural failure remain leading causes of serious outages.

Colliers

Reports that power availability and connection timing now lead site selection constraints in major markets.

Figures are reported as published by their authors and remain attributable to those publishers. Qualitative statements are directional summaries of published industry commentary. No percentage on this page is CMx research.

From the paper

Feasibility is not the test. Evidence is.

A campus valued on four phases and consented for two is a different asset from the one being purchased.

If no single party owns an interface, the investor owns it.

Evidence pathway

Every critical assumption sits in one of four states.

A number in a model carries no record of how it was established. The first task of diligence is to place each critical assumption and to name what would move it forward.

01

Stated

Asserted in a document, presentation or model. No instrument behind it.

02

Contracted

Secured by an agreement with a counterparty, a date and a consequence.

03

Deliverable

Achievable within the programme, supply chain and consents actually held.

04

Proven

Demonstrated by test, energisation or operation under design conditions.

Confidence should follow the state of the evidence, not the confidence of the presentation.

Design, capacity and resilience

The intended design and the installed reality are two different sources of evidence.

Engineering drawings under technical review
The intended designDrawings and schedules describe intent, capacity and resilience as designed.
Installed switchboard aisle in an energised electrical room
The installed realityOnly maintenance and fault conditions on the installed system describe behaviour.
What the paper covers

The tests that decide whether capital should follow.

01Why a technically feasible project can still be a poor investment
02Stated, contracted, deliverable and proven: the four states of any critical assumption
03Power capacity, connection instruments, energisation dates and upstream reinforcement
04Land, planning conditions, environmental constraints and the honesty of phasing
05Connectivity, workload fit, pre-lease evidence and tenant concentration
06Design maturity, failure domains and resilience claims tested against maintenance and fault
07Programme logic, long-lead equipment and the commissioning time genuinely allowed
08Commissioning evidence, operational readiness and handover quality
09Capex, opex, change exposure, governance and interface ownership
10How to report a decision rather than a document set
Selected highlights

Three positions the paper argues.

This page is a preview. The full argument, the diagnostic diagrams, the risk view and the practitioner checklists are set out in the 16-page paper.

01

Test the case, not the file

Confirming that drawings, consents and equipment schedules exist proves completeness of a data room. It does not prove that the asset can support the investment case at the required capacity, resilience, programme, cost and operational standard.

02

Separate the stated from the proven

A capacity figure can come from a firm connection agreement, a reserved allocation, a non-binding indication or an estimate. All four appear in a model as the same megawatts. The paper sets out how to categorise each critical assumption and what would move it.

03

Follow the dependency to its owner

Where an outcome depends on works outside the developer's control, that dependency belongs in the investment case with a date, an owner and a consequence if it slips. Risk that sits between organisations usually returns to the investor.

The paper cites selected market material from CBRE, the Uptime Institute, JLL and Colliers for context. Figures are reported as published by their authors, remain attributable to those publishers, and are not restated as CMx findings.

Who it is written for

Investment committees, lenders, asset managers, developers and operators who need to know whether a proposition is evidenced rather than plausible.

CMx works across advisory, engineering, technical assurance, commissioning and operational readiness. If a live proposition would benefit from that scrutiny, you are welcome to start a conversation.

Download the full paper

The full 16-page paper is issued directly to named readers. Provide your details to receive the PDF.

Your details are used to provide this publication and understand who is engaging with CMx research. We do not treat a publication download as consent to receive marketing. See our privacy notice.