The State of Workforce Learning in UK Data Centres
Competence, procedure discipline and operational resilience in the UK's critical national digital infrastructure.
Prepared for boards, operations directors and workforce leaders.
© 2026 Nuerofy Ltd. All rights reserved. Published July 2026. Industry intelligence series · UK edition
Competence Is Now an Infrastructure Question
In September 2024 the UK government designated data centres as Critical National Infrastructure, placing them alongside energy, water and emergency services. The designation formalised something the sector already understood. When a data centre fails, the consequences travel well beyond the fence line, into health systems, payment networks, logistics platforms and the public services that people assume will simply work.
That designation changes the nature of the workforce conversation. In most industries, training is a human resources matter with a budget line and an annual completion rate. In critical national infrastructure, the competence of the individual standing in front of the switchgear at two in the morning is an operational control, in the same category as redundancy, monitoring and maintenance regimes. It belongs on the risk register, not only in the people plan.
The evidence supports treating it that way. Uptime Institute's outage research has found consistently that human error contributes to a large share of serious incidents, and that the overwhelming majority of those errors trace back to procedures that were either not followed or not adequate in the first place. Meanwhile the sector is expanding faster than its talent pipeline, its most experienced engineers are approaching retirement, and the skills required of a modern critical environment technician are broadening to span mechanical, electrical, digital and security domains at once.
This report examines those pressures and what the most capable operators are doing about them. It is written for boards, operations directors, engineering leaders and the workforce professionals who advise them. It is not a technical manual and it is not a procurement document. It sets out the structural challenges facing the UK data centre workforce, the evidence base behind them, and a practical framework for assessing where an organisation currently stands.
Nuerofy publishes this series because we work daily with the organisations that operate critical infrastructure, and we see the same patterns recur across operators of very different scale. Our aim here is to be useful to the sector, whether or not an organisation ever becomes a customer.
Nuerofy Industry Intelligence
July 2026
The Data Centre Workforce Imperative
Critical national infrastructure, accelerating demand, and a constrained talent pipeline
The UK operates the largest data centre market in Europe. Since the sector received Critical National Infrastructure status in September 2024 it has attracted substantial committed private investment, and techUK reported a development pipeline worth approximately £36.4 billion across close to 100 planned schemes as of August 2025. Every megawatt in that pipeline requires people who can commission it, operate it, maintain it and recover it when something goes wrong.
The Economic Case Is Established. The Workforce Case Is Not.
techUK's Foundations for the Future report, produced with Henham Strategy, estimates that UK data centres currently contribute £4.7 billion in gross value added annually, generate £640 million in tax revenue and support 43,500 jobs across the economy. If capacity growth continues above trend, the report projects an additional £44 billion in GVA between 2025 and 2035, alongside 40,200 additional operational roles and 18,200 additional construction roles.
Those projections are conditional. They assume the roles can be filled by people competent to perform them. Uptime Institute's 15th Annual Global Data Center Survey, drawing on responses from more than 800 owners and operators, found that close to two thirds report difficulty retaining staff, finding qualified candidates, or both. Within that, 46% specifically cited difficulty finding qualified candidates for vacant roles, while 37% reported problems retaining the staff they already have.
The binding constraint on UK data centre growth is increasingly not power, land or capital. It is the availability of people competent to operate what gets built.
1. Critical National Infrastructure Expectations
CNI designation raises the bar on resilience, security and operational governance. It brings closer coordination with government on cyber and continuity risk, and it sharpens what clients, insurers and regulators expect by way of evidence that an operator's people are competent to work in a critical environment. In practice this moves competence records out of the HR filing system and into the operational assurance picture, where they are examined alongside maintenance regimes and incident history.
2. Demographic Attrition and Knowledge Loss
A substantial proportion of the UK's experienced critical environment engineers are within a decade of retirement, and industry analysis points to a significant share of the current workforce leaving the sector before 2030. Each departure removes knowledge that was never written down: the behaviour of a particular switchgear line-up, the sequence that works on a specific chiller plant, the reason a procedure was drafted as it was. Unlike a vacancy, this loss is not visible on a headcount report until the moment it matters.
3. Converging Skill Requirements
Roles once cleanly divided between IT and facilities are merging. A 2025 IDC Info Brief cited by Colt Data Centre Services identified persistent UK skills gaps in cybersecurity (44%), networking (40%) and data, AI and automation (33%). Modern critical environment roles increasingly combine mechanical and electrical engineering with monitoring and DCIM tooling, energy management and security awareness. Very few candidates arrive holding the full combination, which makes structured internal development the only realistic route to a fully competent team.
Rising rack densities driven by AI workloads compound all three pressures. CBRE's UK Data Centres Outlook 2026 reports that combined new supply to the London market across 2025 and 2026 will reach 373MW, more than double the 147MW delivered in the preceding two-year period. Capability therefore has to be built while the estate expands, not afterwards.
Questions for the board
- What is our current time from hire to independent unsupervised working, and is it trending in the right direction?
- How many of our critical operational activities depend on a single individual at any one site?
- What proportion of our most experienced engineers will reach retirement age within five years, and what is the plan to capture what they know?
Critical Environment Safety & Electrical Competence
The legal and operational foundation of working inside a live data hall
A data centre is an occupied high-energy electrical environment that must not be switched off. Work takes place live, alongside energised switchgear, standby generation, high-capacity battery systems and pressurised cooling circuits. Safety competence in this setting is not a general workplace obligation discharged by an annual e-learning module. It is an operational control that protects both people and continuity of service.
The Legal Framework
UK operators carry duties under the Health and Safety at Work etc. Act 1974, the Electricity at Work Regulations 1989, the Management of Health and Safety at Work Regulations 1999, the Provision and Use of Work Equipment Regulations 1998 and, where relevant, the Confined Spaces Regulations 1997 and the Control of Substances Hazardous to Health Regulations 2002.
Regulation 16 of the Electricity at Work Regulations shapes most of the sector's competence architecture. It requires that no person works on or near an electrical system unless they possess the technical knowledge or experience, or are under an appropriate degree of supervision, to prevent danger and injury. Authorised Person appointments, safe systems of work, permit regimes and switching authorisations all exist to give effect to that requirement and to evidence it afterwards.
After a serious incident, the training record ceases to be an administrative detail. It becomes the principal evidence that the person doing the work was competent to do it.
The Competence Framework in Practice
- Electrical safety awareness: the behaviour of LV and HV systems, arc flash risk, approach boundaries, and the limits of an individual's authorisation.
- Safe isolation: proving dead, lock-off and tagging, stored energy, and the verification discipline that must hold under time pressure.
- Permit to work: permit types, issue and receipt responsibilities, cancellation, and the expectation of working strictly within the scope described.
- HV and LV switching: switchgear types, interlocking, transfer sequences, and the responsibilities of the Authorised Person within a switching programme.
- Working at height and confined spaces: plant rooms, roof-mounted condensers, tank spaces and underfloor voids, including rescue arrangements.
- Manual handling, PPE selection and inspection: rack and battery handling, arc-rated clothing, and gloves rated for the task in hand.
- Emergency response: casualty management, the specific implications of electrical injury, and the location and use of site emergency equipment.
- Contractor and visitor control: site rules, escorting, restricted areas, and the reporting route for anything that looks wrong.
Where Operators Typically Sit
Across the operators we work with, safety competence management follows a recognisable progression. Most sit at stage two or three and assume they are at four.
| Stage | Characteristics | Typical Exposure |
|---|---|---|
| 1 · Manual | Paper sign-off sheets, classroom-only induction, local record keeping | Evidence cannot be assembled quickly |
| 2 · Digitised | Online modules and a basic LMS, certification tracked manually | Renewals depend on individual diligence |
| 3 · Centralised | One platform across sites, contractors partially included | Gaps concentrate in contract labour |
| 4 · Automated | Automated renewals and expiry alerts, role-based pathways | Content currency lags procedure change |
| 5 · Assured | Live competence visibility, predictive expiry management, board reporting | Residual risk is known and managed |
The Contractor Question
Most facilities are maintained by a mixed workforce of directly employed staff, facilities management partners, specialist M&E contractors and manufacturer engineers. Competence governance frequently stops at the boundary of direct employment, with contractor evidence held by the contracting company and accepted on trust. This is where audits most reliably find gaps.
The exposure is not theoretical. A contractor working under a permit issued by the operator is working within the operator's safe system of work, and the adequacy of that system will be assessed as a whole after an incident. Operators that verify contractor competence directly, rather than relying on a supplier declaration, consistently report fewer surprises during client assurance visits and a shorter path to closing audit actions.
Questions for the board
- Could we evidence, for every person on site today, the authorisation under which they are working and the training that supports it?
- Who holds accountability for judging competence for live electrical work, and on what basis is that judgement made?
- Are contractors and agency staff held to the same evidenced standard as our own employees, and how do we verify it rather than assume it?
Systems Competence: Power, Cooling & Suppression
Understanding the plant, not just the button sequence
Safety competence prevents harm to people. Systems competence prevents harm to service. The distinction matters because organisations routinely invest heavily in the first while leaving the second to informal shadowing, on the assumption that a technician who has watched the task performed twice can perform it alone.
Power: The Dominant Failure Domain
Uptime Institute's research has repeatedly identified UPS failures as the single biggest cause of power-related incidents. The competence question is whether technicians understand topology and consequence rather than sequence alone: how a static UPS behaves in double conversion against eco mode, what placing a unit on bypass does to the protection scheme, and what a given alarm means for the resilience the design was meant to provide.
Standby Generation
Start-up and load acceptance, fuel management and quality, load bank testing regimes and the sequencing logic of automatic transfer switching. Much of this is exercised only during scheduled testing, which means real competence has to be built deliberately rather than accumulated through routine exposure.
Cooling: Narrowing Margins
As AI workloads drive higher rack densities, the time available to respond to a cooling failure has shortened. CRAC and CRAH operation, chiller plant sequencing, containment discipline and airflow fundamentals all sit within shift decision-making. High-density deployments introduce liquid cooling concepts, with leak detection and containment expectations many teams have not previously needed.
Suppression: Expensive Good Intentions
Staff need to understand very early smoke detection, two-stage alarm logic, the discharge behaviour of systems such as FM-200 and NOVEC 1230, the function of hold and abort controls, and safe evacuation. Equally significant is prevention of false discharge, which can take a hall offline as effectively as a fire.
A facility can be designed for concurrent maintainability and still fail on a Tuesday afternoon because the person on shift did not understand what the design assumed they would do.
Security as an Operational Discipline
In a CNI environment, physical security competence extends beyond the security team. Access control discipline, mantrap operation, visitor and contractor escorting, CCTV awareness, tailgating prevention and incident reporting all fall within the expected competence of operational staff, and all are routinely examined during client assurance visits and certification audits.
Questions for the board
- Do our technicians understand the systems they operate, or only the sequences they have been shown?
- What training accompanied our last major plant installation, and was it delivered before or after handover?
- Who on each shift is trained on suppression abort protocol, and when was that last verified?
Procedure Discipline & Human Error
The single highest-return area of workforce investment in the sector
The most consequential finding in data centre resilience research concerns people and procedures rather than equipment. Uptime Institute's Annual Outage Analysis 2025 reported that nearly 40% of organisations had suffered a major outage caused by human error in the preceding three years, and that 85% of those incidents stemmed from staff failing to follow procedures or from flaws in the procedures themselves. Uptime also recorded a ten percentage point rise in failure-to-follow-procedure outages in 2025 compared with 2024, and its 2026 analysis found that failures to follow established procedures remain the leading driver of human error-related outages.
Read the other way around, this is the most encouraging statistic in the sector. The dominant cause of serious downtime is a category that training and procedure review can directly address.
Why Procedures Fail
Uptime's data separates two distinct failure modes, and they call for different responses. Where staff did not follow a sound procedure, the issue is competence, culture or conditions: the individual did not know the procedure existed, did not understand why it mattered, or was under pressure that made deviation feel reasonable. Where the procedure itself was inadequate, the issue is authorship: the document was written by someone who knew the task too well to describe it fully, or was never revised after the plant changed.
Both failure modes are addressable, and neither is addressed by asking staff to be more careful. The first requires structured training and a culture in which stopping work is supported. The second requires that procedure writing is treated as a skill in its own right, with peer review and a revision trigger tied to plant change.
What Competent Procedure Practice Looks Like
- Method of Procedure authorship: objective, risk assessment, step sequence, abort criteria and back-out plan, written to be executable by a competent technician who has not performed this exact task before.
- Peer review before first use, ideally by someone who will not be executing the work.
- Execution discipline: read the step, perform the step, record the step, and stop when observed reality diverges from the document.
- Standard Operating Procedure adherence across rounds, monitoring, alarm response and escalation.
- Emergency Operating Procedures, including explicit authority to deviate and who holds it.
- Change management: risk classification, approval routes, change freeze periods and post-change verification.
- Incident response: immediate actions, evidence preservation, escalation and handover into root cause analysis.
- Root cause analysis that closes with a procedure revision and a training update, not only a report.
The Loop Most Organisations Leave Open
Nearly every operator conducts root cause analysis after a significant incident. Far fewer can demonstrate that the finding reached every affected technician, on every shift and across every relevant site, within a defined period, with a record of who received and understood it. That interval between identifying a lesson and embedding it across the workforce is the point at which the same failure becomes available to happen again, at another site, to someone who was never told.
Questions for the board
- What proportion of our incidents and near misses over the past 24 months involved a procedure that was not followed or was not adequate?
- How long does it take, following a root cause analysis, for the lesson to reach every affected technician with an auditable record?
- Who reviews our MOPs for quality and currency, how often, and what triggers a revision?
Conversion: From Trades Talent to Technician
Reframing a national skills shortage as an internal capability question
The UK is not short of qualified electricians, HVAC engineers and mechanical fitters. It is short of people who can apply those trades safely inside a live, high-availability, procedure-driven critical environment. The distinction is strategically significant, because it converts the problem from one of national skills supply, which no individual operator can solve, into one of conversion speed, which every operator can influence directly.
Operators that can convert a qualified electrician into a competent data centre technician in weeks will staff their halls. Those that cannot will keep competing for the same experienced technicians as everyone else, at rising cost.
What Conversion Actually Requires
A time-served electrician arriving from construction or industrial maintenance already holds the fundamentals: electrical theory, isolation practice, test instrument competence and a safety mindset. What they lack is a specific and finite overlay, which is precisely why it can be taught as a structured programme rather than absorbed over years.
- Concurrent maintainability and fault tolerance: why work is performed live, and what redundancy does and does not guarantee.
- Procedure culture: MOPs, SOPs, permits, and the expectation that nobody improvises.
- Site-specific topology: this facility's switchgear, UPS configuration, cooling plant and containment strategy.
- Alarm literacy: interpreting what the BMS and DCIM are reporting, and what warrants escalation rather than observation.
- Client and audit awareness: why the record of the work matters alongside the work itself.
- Security discipline: access control, escorting, and the reporting obligations attached to CNI status.
Designing for a 24/7 Rota
Data centres operate continuously, which makes classroom-based induction structurally unfair to the people working nights and weekends. Where training is available only in a scheduled weekday session, the night shift receives a compressed or deferred version, and the training record becomes quietly less reliable than it appears on a dashboard. Any serious conversion programme has to be deliverable across the rota, in short modules that fit a genuine break, with completion captured automatically.
Retention Is the Other Half of the Equation
Conversion investment only pays if people stay. Uptime's finding that 37% of operators struggle with retention should be read alongside the recruitment figure rather than separately from it, because in a market this tight, a well-trained technician is a portable asset. Structured development pathways, visible progression and CPD recognition are among the few retention levers available to operators who cannot simply outbid hyperscale competitors on salary.
Questions for the board
- Is time-to-competency tracked as an operational KPI, and does it reach this board?
- Do night and weekend shifts receive a demonstrably equivalent training standard to the day shift?
- What is our attrition rate among technicians in their first two years, and what does exit data tell us about development and progression?
Site-Specific Knowledge at Scale
The content problem that constrains almost every operator's training capability
Generic critical environment training conveys principles. It cannot convey a specific facility. The training that prevents incidents is local: this switching sequence, this containment strategy, this escalation matrix, this client's access requirements. That content almost always exists already, written as MOPs, SOPs, commissioning documentation and site handbooks. It simply exists in a form that nobody learns from.
Where Training Capability Actually Sits
Data centre training is rarely owned by a dedicated learning function. It typically sits with an operations manager, a critical environment lead or an HR generalist who is simultaneously carrying compliance, recruitment, audit preparation and rota management. Traditional content development, in which a subject matter expert briefs an instructional designer over several weeks, cannot keep pace with a facility that commissions new plant, revises procedures and onboards contractors continuously.
The consequence is predictable and widely observable. Generic mandatory training is current, because it can be bought. Site-specific training is out of date, because it has to be made. The training that matters most is the training least likely to reflect the facility as it exists today.
Most operators do not have a training content shortage. They have a conversion shortage: the knowledge exists, but not in a form that can be delivered, assessed and evidenced.
The Case for Automated Content Conversion
The practical development over the past two years has been the maturing of AI-assisted authoring, which changes the economics of site-specific content. Where converting a procedure into structured training with assessment previously took days of specialist effort, it can now be drafted in minutes and refined by the subject matter expert who owns the procedure. The expert reviews rather than authors, which is a far better use of the scarcest resource in the building.
The governance implication deserves board attention. Automated drafting does not remove the need for technical review and approval, and any operator adopting it should be explicit about who signs off content before it reaches the workforce. The gain is speed of conversion, not removal of accountability.
Where Conversion Delivers Most Value
- Site-specific induction reflecting actual topology, access rules and escalation routes for each facility rather than a generic estate-wide document.
- Task training derived directly from the MOP, so a technician learns the procedure before executing it live and the learning is recorded.
- Post-incident briefings issued within days of a root cause analysis, closing the loop described in section 04.
- Commissioning handover training delivered at the point new plant goes live, rather than months later once problems have surfaced.
- Client-specific requirements for colocation customers and managed service contracts.
- Contractor briefings delivered and evidenced before arrival on site rather than at the gate.
- Retirement knowledge capture, recording how experienced engineers actually perform critical tasks while they are still available to explain it.
Questions for the board
- How much of our operational knowledge exists only in individuals' heads rather than in documented, teachable form?
- What is the current lag between a procedure changing and the affected workforce being trained on the change?
- Who owns training content creation in this organisation, and what else are they responsible for?
Competence Assurance, Audit & Governance
From scattered records to evidence that holds under scrutiny
Competence evidence is tested at predictable moments: an ISO 27001 or ISO 45001 surveillance audit, a SOC 2 examination, a colocation client's technical assurance visit, an insurer's review following a claim, or an investigation asking who performed a switching operation and what training they held on the day. In each case the question reduces to the same form. Can the organisation demonstrate, for every person who worked in this facility, what training they completed, when, and whether it remained current at the relevant time?
Why Fragmented Records Fail
Competence evidence in data centre operations is characteristically scattered: authorisation letters in a site folder, external course certificates in personnel files, toolbox talk attendance sheets in a drawer, contractor inductions held by the facilities management partner, and a spreadsheet maintained by whoever has capacity. Each element may be individually valid. Collectively they are slow to assemble, difficult to reconcile and unpersuasive at precisely the moment persuasion is required.
The reputational asymmetry is worth noting. An operator with excellent training practice and poor record architecture will present badly in an audit, while the evidence itself is what the auditor is entitled to rely on. Good practice that cannot be demonstrated is, for assurance purposes, indistinguishable from absent practice.
Competence that cannot be evidenced quickly is treated, by auditors, clients and insurers, as competence that may not exist.
What Good Assurance Looks Like
- A single competence record per individual, covering directly employed staff, agency workers and contractors on the same basis.
- Timestamped history of every assignment, completion, assessment outcome and renewal, retained for the period required by the relevant certification scheme.
- Forward visibility of expiry, so certification lapses are managed as a forecast rather than discovered as an exception.
- Role-mapped requirements, so it is clear what a given role must hold and what is missing rather than only what has been completed.
- Multi-site aggregation with drill-down, allowing portfolio-level oversight and site-level investigation from the same source.
- Rapid export in a form a client or auditor will accept without further manual assembly.
- CPD recognition where relevant, supporting individual professional development records alongside organisational compliance.
From Evidence to Governance Intelligence
The same data that satisfies an auditor answers more valuable questions for the board. Which sites carry the thinnest cover of authorised persons for a given switching activity, and what happens if one individual is unavailable? Where do certification expiries cluster over the coming quarter, and does that coincide with planned maintenance activity? Do assessment outcomes in a particular discipline correlate with where near misses are being reported? Treated this way, workforce competence becomes a managed operational risk with leading indicators, rather than a compliance percentage reviewed annually.
Questions for the board
- How long would it take today to produce complete competence evidence in response to a major client assurance request?
- Do we have a single source of truth across every site and every employment type, or several partial ones?
- Which competence metrics currently reach this board, how often, and are they leading or lagging indicators?
A Board-Level Framework & Priorities
Assessing where the organisation stands and what to address first
The framework below is intended as a self-assessment instrument for boards and executive teams rather than a maturity model to be scored for its own sake. In our experience most operators place themselves one stage higher than the evidence supports, and the exercise is most useful when the assessment is made against what could be demonstrated tomorrow morning rather than what is believed to be true.
The Five-Stage Competence Readiness Framework
| Stage | Focus | Characteristics | Test of Attainment |
|---|---|---|---|
| 1 · Baseline | Visibility | Training mapped, gaps documented, exposure identified | We know what we do not know |
| 2 · Foundation | Consolidation | Single platform, migrated records, automated renewals | One source of truth exists |
| 3 · Conversion | Capability | Defined trades-to-technician pathways, shift-accessible delivery | Time-to-competency is measured and falling |
| 4 · Procedure | Resilience | MOPs converted to training, post-incident loop closed, procedure competence tracked | Lessons reach the workforce within days |
| 5 · Assured | Governance | Live competence visibility, predictive expiry management, board-level reporting | Evidence is available on demand |
Metrics Worth Reporting at Board Level
Training completion percentage is the metric most commonly reported to boards and among the least informative, because it measures activity rather than capability or risk. The following set is more diagnostic and can generally be produced from the same underlying data.
| Metric | Definition | Why It Matters |
|---|---|---|
| Time-to-competency | Days from start date to independent unsupervised working in a defined role | Sets the ceiling on how fast the estate can be staffed |
| Authorised person cover | Authorised persons per critical activity, per site, per shift | Exposes single points of failure in switching capability |
| Certification currency | Proportion of the workforce with all role-required training in date | Direct regulatory and client exposure |
| Expiry concentration | Certifications due to lapse within the next 90 days | Converts compliance from reactive to forecast |
| Procedure-related events | Incidents and near misses where a procedure was not followed or was inadequate | Leading indicator of outage risk |
| Contractor compliance | Proportion of contractor workforce with verified, current site induction | Where audits most frequently find gaps |
| Knowledge concentration | Critical tasks documented and trained against those held informally | Quantifies retirement and attrition exposure |
Ten Priorities for 2026 and 2027
- Move competence onto the risk register. In critical national infrastructure, the competence of the individual performing the work is an operational control and should be governed as one.
- Test your evidence rather than your intention. If a client assurance team arrived tomorrow, could complete records be produced for every person on site, including contractors, within 30 minutes?
- Treat procedure discipline as a resilience investment. With failure to follow procedures identified as the leading driver of human error outages, this is the workforce intervention with the clearest link to availability.
- Convert existing MOPs and SOPs into training. The documents are already written; the gap is that nobody learns from a PDF in a shared drive.
- Close the incident-to-training loop with a defined service level. Every root cause analysis should produce a training update delivered within a stated number of days, with a record of receipt.
- Build and measure a conversion pathway. Define the route from qualified electrician or M&E engineer to competent data centre technician, and report time-to-competency as an operational KPI.
- Extend competence governance to the contract workforce. Clients, auditors and investigators do not distinguish by employment type when something goes wrong.
- Make delivery genuinely shift-accessible. If night and weekend staff cannot complete training on their own device at a time that suits the rota, the training record is less reliable than it appears.
- Capture retiring engineers' knowledge deliberately and now. Record how the most experienced staff actually perform critical tasks and convert it into training while they are still available to be asked.
- Replace the annual completion report with live visibility. Competence data across every site, shift and discipline turns an unmanaged dependency into a governed risk.
About This Report
Sources, method and publisher
Method
This report synthesises published industry and government research with observations drawn from Nuerofy's work with organisations operating critical infrastructure, including hyperscale and colocation operators, facilities management providers and M&E engineering firms. Where the report characterises common practice, that characterisation reflects patterns we observe across those engagements and is presented as informed observation rather than survey finding. Quantitative claims are attributed to their published sources throughout.
Sources
UK economic and employment figures draw on techUK's Foundations for the Future report (November 2024), developed with Henham Strategy, which estimates £4.7 billion in annual gross value added, £640 million in tax revenue and 43,500 jobs supported, alongside projections of an additional £44 billion in GVA, 40,200 additional operational roles and 18,200 additional construction roles between 2025 and 2035. The designation of data centres as Critical National Infrastructure was announced by the Department for Science, Innovation and Technology in September 2024. Development pipeline figures draw on techUK reporting from August 2025, and London supply figures on CBRE's UK Data Centres Outlook 2026.
Staffing and resilience findings draw on Uptime Institute's 15th Annual Global Data Center Survey (2025), based on responses from more than 800 owners and operators, together with its Annual Outage Analysis reports for 2025 and 2026. UK skills gap percentages draw on a 2025 IDC Info Brief cited by Colt Data Centre Services. Regulatory descriptions reflect publicly available information from the Health and Safety Executive, including the Health and Safety at Work etc. Act 1974, the Electricity at Work Regulations 1989 and related legislation. All cited sources are publicly available from their respective publishers.
Limitations
Data centres do not hold a discrete UK industrial classification, which means sector employment and economic figures are estimates derived from modelling rather than direct measurement, and should be treated as indicative of scale rather than precise. Uptime Institute's survey data is global; where it is cited here, readers should note that UK conditions may differ from the international average. This edition is not derived from a Nuerofy primary research survey.
Forthcoming Research
Nuerofy is commissioning a UK sector survey covering hyperscale operators, colocation and carrier-neutral providers, facilities management contractors, M&E engineering firms and edge and regional operators, with a full data-led edition of this report scheduled for 2027. Operators interested in participating are welcome to contact us.
About Nuerofy
Nuerofy is a UK-based AI-powered Learning Management and Experience Platform working with organisations across regulated and safety-critical sectors, including the operators, contractors and engineering firms that run the UK's digital infrastructure. We publish this industry intelligence series because the workforce challenges described here are structural and shared, and are better addressed with a common evidence base.
Ready to Transform Your Staff Training?
Join leading educational institutions already using Neurofy to deliver compliant, effective online training while enhancing safeguarding and teaching excellence