Lessons developers can apply as data centres enter their major project era

As data centre projects increase in scale and complexity, the sector has an opportunity to apply lessons already learned elsewhere rather than repeating expensive mistakes.

The following six lessons are drawn from observing what consistently differentiates successful infrastructure delivery from project underperformance.

The best data centre developers think like operators

Perhaps the most important lesson from infrastructure delivery is that contracts, procurement models and governance structures are only enablers. The real determinant of success is whether the developer has a sufficiently clear understanding of the operational asset it is trying to create. For a data centre, practical completion therefore has relatively little value on its own. Value is only created when the facility performs reliably throughout its operational life.  

This philosophy is often referred to as the concept of the ‘intelligent owner’ or ‘intelligent client’. At its core is a simple proposition: successful owners start with the operational outcomes they are seeking to achieve and shape project delivery accordingly. 

An intelligent client makes timely decisions, allocates risk realistically and aligns governance, procurement and delivery around clearly defined outcomes, recognising practical completion is a milestone, while operational performance is the outcome. They also recognise that client capability cannot be outsourced. While design, construction and specialist services can be procured, responsibility for defining outcomes, making strategic decisions and maintaining alignment across the project ultimately remains with the client organisation. 

In that context, commissioning becomes the pivotal turning point for a project. Rather than simply a handover activity or the final stage of construction, it is the process through which the client validates that the completed asset can perform as intended in operation and is where the consequences of earlier project decisions become visible. Problems such as poorly defined requirements, unclear interfaces, fragmented accountability and weak governance, which can be managed during design and construction, are exposed when an integrated asset is tested as a functioning system. A developer that thinks like an operator can address these challenges strategically from the outset. 

Interface risk is becoming the defining challenge 

As data centre projects increase in scale, success depends on the seamless integration of multiple systems and disciplines, from grid infrastructure and original equipment manufacturer (OEM) products to operational systems and civil, structural and architectural components. 

The greatest risks often sit between packages rather than within them. This is not a new challenge. Offshore wind, oil and gas, rail and power transmission projects have long recognised that project outcomes are often determined by the quality of interface management rather than the performance of individual contractors. We frequently see developers devote significant time to negotiating risk allocation, when clearly defined interface ownership and operational requirements are often more important to project delivery. 

Individual contractors may successfully complete their own scope, however overall project success depends on the entire system functioning as a single operational asset. Data centre developers should therefore spend at least as much time aligning scope and defining interface ownership as negotiating liability clauses. Many of the most significant risks on data centre projects, including commissioning, operational readiness, long-lead equipment integration and utility dependencies, cut across organisational and contractual boundaries. Managing those interfaces effectively is often a greater determinant of success than the choice of contract form itself. 

Selecting the right delivery model: focus on outcomes, not labels  

One of the most common questions in data centre development is what model to adopt for delivering the project – whether an EPC model, a multi-package strategy, a construction-only model or design and build (D&B). In practice, we see all these approaches being used successfully and unsuccessfully across different geographies and project types. Similarly, pricing structures range from fixed lump sum arrangements to target cost and remeasurement mechanisms. In our experience, projects rarely struggle because the wrong delivery model was selected.  

The real question is whether the chosen delivery model aligns with the project's objectives, risk profile and delivery strategy. Debates around EPC versus D&B often consume disproportionate attention. However, difficulties more commonly arise because delivery requirements, interfaces, governance and commissioning responsibilities were not properly understood at the outset.

For example, an early commitment to an EPC strategy on a project with complex package interfaces may result in prolonged and difficult negotiation periods and attract higher pricing. Early consideration of an alternative, multi-package approach may provide better value for money and a simplified contracting strategy, reducing both cost and programme duration.  

Every project is unique, but in all cases the delivery model should be the consequence of the project strategy, not the starting point. 

Why collaborative contracting could deliver better outcomes 

Many data centre projects already exhibit characteristics of an enterprise approach to delivery. Unlike a transactional approach, which focuses on procuring and managing individual contracts, an enterprise approach brings together owners, designers, contractors and suppliers around a shared objective: the successful long-term performance of the asset. Large data centre developers typically deliver programmes rather than one-off projects, and relationships with specialist contractors, OEMs and commissioning teams are frequently long-term. In an increasingly constrained supply chain environment, access to specialist contractors, critical equipment and commissioning expertise can be just as important as access to capital. 

The question is therefore no longer simply: How much risk can we transfer? Increasingly, it is: How do we become a developer that the best supply chain wants to work with? 

While genuinely collaborative contracting models remain relatively uncommon in data centres, design and construction contracts often operate in a collaborative manner in practice, including through early contractor engagement, joint development of technical solutions and long-term delivery relationships. 

Collaborative contracting does not fundamentally alter the underlying allocation of risk. Rather, it seeks to strengthen alignment between project participants by linking commercial incentives more directly to successful delivery. Some of the most significant risks on large data centre projects, including systems integration, commissioning, operational readiness, utility dependencies and phased delivery, require collective effort. By aligning objectives, improving transparency and creating incentives around shared project outcomes from the outset, collaborative approaches can encourage earlier problem solving, better decision-making and more effective management of interfaces. 

The fastest projects often spend longer planning  

AI, cloud demand and power constraints are increasing pressure to accelerate delivery across the sector. In any project driven by demanding customer programmes, the temptation is to accelerate procurement and start construction as quickly as possible. Yet one of the most consistent lessons from successful infrastructure delivery is that speed is rarely achieved by rushing the early stages of a project.  

As the lessons above demonstrate, successful projects invest early in establishing delivery strategies and implementing effective interface management. The paradox is that investing more time upfront often results in shorter overall delivery programmes because fewer issues emerge later. 

Across almost all our global construction markets, we are seeing an increased use of extended periods of collaborative, multi-disciplinary working before main works commence, whether termed ‘pre-construction’, ‘early contractor involvement’, ’development phase’ or ‘stage one’. 

This need for effective pre-construction planning is even more pronounced in data centres, with the need to accommodate customer driven design changes. Building this reality into delivery and commercial models will help reduce disruption and align parties around the common project goals that matter, even during periods of uncertainty.  

To borrow a phrase increasingly used in major project circles: think slow, act fast. Or, put another way: speed comes from preparation, not acceleration. 

Local stakeholders are not an afterthought 

As data centre development accelerates globally, public awareness of its demands on local energy and water resources continues to grow. In turn, communities, local authorities and environmental groups are becoming more engaged in the development process.  

Developers must therefore be proactive in managing local stakeholder relationships, particularly where projects raise issues of public concern. For example, the effective management of the permitting process, including zoning approvals, environmental consents and planning hearings, requires early and transparent engagement. This type of stakeholder engagement is an important part of project set-up, as opposition at these stages can be just as damaging to the delivery programme as a supply chain disruption or delays during construction.  

It is equally important to engage constructively with concerns about the technology itself. Where objections rest on outdated assumptions or incomplete information, data centre developers should be prepared to explain clearly how their project incorporates upgraded technologies and solutions that address those concerns directly. Successful projects are often those that build trust early and maintain engagement throughout the development process.  

Data centres are entering their major project era

The next generation of successful developments will not be distinguished solely by technology, scale or speed to market, but by delivery maturity. The organisations that succeed will be those that think like operators, manage interfaces rigorously, select delivery models that fit their objectives, engage effectively with their supply chains and invest in project set-up before accelerating delivery. 

The future of data centre contracting may therefore be less about inventing new contractual models and more about applying proven infrastructure delivery principles to an increasingly complex digital infrastructure sector. As data centres become larger, more power-intensive and more operationally critical, the industry's greatest opportunity lies in combining technological innovation with the delivery disciplines that have enabled other complex infrastructure sectors to scale successfully. Those who can do both are likely to define the next generation of digital infrastructure. 

Co-written by Katy Burke of Pinsent Masons.

Anne-Marie Friel will be on the panel discussing project delivery at the Platform Global conference on 9 September 2026. Pinsent Masons is a sponsor of the event.