Current trends and developments in procurement of data centre delivery

This is part of a series of articles put together by our Construction team, called ‘BREAKING GROUND’ where we share the latest thinking on commercial issues for construction projects

Data centre delivery across the APAC region is scaling rapidly, propelled by cloud adoption, AI workloads and heightened data‑residency requirements. However, persistent bottlenecks in skilled contractors and critical equipment are amplifying delivery risks. These factors are compressing delivery timelines and favouring procurement strategies that repeat reliably across jurisdictions. 

This article looks more closely at the rationale for the data centre procurement models currently being used in the market.

An overview of the latest developments in the sector and their impact on delivery models and contract form will be discussed in our forthcoming articles in this series.

Risks and Factors Affecting Procurement  

Key data centre delivery risks

Energy

Power is now the defining constraint in data centre delivery. Grid capacity, regulatory approvals and upstream network reinforcement can extend programmes by many months (and sometimes years) because substation availability rarely aligns with a developer’s schedule (or chosen location).

This requires developers to engage utilities at the point of site selection, map credible upgrade pathways, and importantly build realistic float into the programme.  

Where renewable power purchase agreements are part of the solution, they need to be treated as a parallel workstream with clear milestones, dependencies and risk ownership, which are mirrored in commissioning and testing regimes.

In Australia, the approach of network operators varies from State to State, needing a tailored solution.

Water

Effective cooling depends on reliable water supply and compliant discharge. In many APAC jurisdictions, connection volumes and discharge limits are tight and early engagement with water and environmental regulators is critical to secure the required rights and any necessary pre‑treatment.

Developers also need to consider seasonal variability, utility upgrade timing and embed clear contractual responsibilities for approvals, metering, water quality and testing directly aligned to the chosen cooling strategy.

Sustainability objectives and WUE targets increasingly drive design and procurement choices. Approaches such as closed‑loop systems and adiabatic solutions can materially reduce water consumption, but they also reshape supplier and equipment selections and commissioning requirements.

Long‑lead equipment (LLE)

Factory capacity constraints, shipping dynamics and export rules can make delivery dates volatile. One mitigant is for developers to secure production slots early and consider owner‑procured LLE of critical items for free‑issue.

However this is not without significant upfront investment and has the potential for delay if not managed well or if the owner is inexperienced in this market. 

Site risks

Unforeseen site conditions are among the most common causes of budget and programme blowouts. Contamination risks (including asbestos and PFAS) together with hard rock and unknown underground services can materially disrupt delivery. These risks require rigorous geo‑environmental due diligence and intrusive investigations to effectively manage, often followed by targeted early works to clear obstructions and “prove” the site before main works can commence.

Permits and approvals

Many jurisdictions classify data centres as critical infrastructure, which frequently triggers foreign investment reviews, security clearances and detailed environmental assessment, alongside land‑use, utilities, traffic, noise and emissions approvals.  Even if a data centre does not meet this classification, various levels of government in Australia are increasing seeking to regulate delivery of this infrastructure, and at the same time, provided tailored pathways for smoother delivery.[1]

Contractor capability

Experienced data centre contractors, especially those proficient in high‑density MEP and integrated commissioning, remain in short supply across APAC, pushing up pricing and limiting appetite for full EPC ‘wrapped’ risk. Thin capacity also elevates interface and safety risk on multi‑phase campuses where construction overlaps with live operations, requiring developers to address clear scope demarcations, master scheduling, coordinated access protocols and integrated commissioning.

For offshore suppliers, developers also need to consider counterparty creditworthiness (and the need for parent guarantees or bonds) and cross‑border enforceability.

Managing the above risks requires realistic scheduling across the contractual suite, with appropriate contractual mechanisms to capture regulatory shifts, procurement and market changes, and bespoke risk allocation to narrow any gaps with developers' upstream obligations (e.g under lease arrangements).

Factors affecting procurement choices

Hyperscalers

Hyperscale developers generally operate under master service agreements that fix hard completion deadlines, impose stringent liquidated damages and allow only narrowly circumscribed extensions of time.

Those commercial pressures and the above risks tend to drive a disaggregated delivery model, breaking works into concurrent packages such as core-and-shell, MEP fit-out, and direct procurement of long‑lead equipment.

Early Contractor Involvement (ECI) is being frequently deployed to crystallise design intent, align procurement strategies and validate constructability while head terms are still being finalised.

While this strategy mitigates schedule risk, and allows securing factory slots early, ahead of letting the main works, it tends to elevate interface risk and may erode negotiating leverage if not done properly.

Where designs are highly standardised and repeatable (such as campus blocks built to a settled template), hyperscalers may opt for selective turnkey EPC delivery for modules to obtain single‑point commissioning accountability and bankability. Even in those cases, contracts are generally customised to pass through MSA obligations, embed enhanced performance regimes and secure strong step‑in rights if milestones slip.

Given the limited pool of EPC contractors prepared to absorb full risk (and the resulting price premia), a pragmatic hybrid often emerges: turnkey delivery for repeatable blocks to anchor schedule certainty, paired with disaggregated ECI‑led delivery for bespoke elements to reflect supply chain realities and maintain agility.

Colocation developments

The procurement approach for colocation developments typically balances two imperatives: preserving flexibility for phased tenant fit‑outs and meeting lender expectations on bankability. A design & construct powered shell, coupled with two‑stage procurement for the fit‑out, is a common structure because it allows an early start on the base building while keeping tight control over tenant interfaces as occupiers come online.

When supported by clear scope demarcations, robust performance testing and appropriate direct agreements, this model is generally attractive to financiers. It also lends itself to staged handover, partial possession and practical rules of access as tenants mobilise, which in turn reduces operational disruption and mitigates the risk of claims in subsequent phases.

In several emerging APAC markets, developers are turning to wrapped DBOM/operator models to secure single‑counterparty accountability across design, build, operation and maintenance. The wrap can simplify financing and lifecycle risk allocation, but it must be underpinned by a carefully calibrated performance regime, lender‑friendly step‑in rights and alignment with the lease structure and regulatory approvals.

Importantly, any operator wrap should be future‑proofed to accommodate technology refresh (including liquid cooling and higher rack densities) while harmonising warranties across the supply chain.

Limited recourse financing – Lender’s expectations

In limited recourse structures, lenders place particular emphasis on how delivery risk is allocated and controlled, leading to a greater interrogation of interface risk, commissioning certainty, and the degree to which the build and operations profile is aligned with the revenue model set out in the customer MSA.

Where the procurement strategy disaggregates delivery across multiple packages, financiers will look for an umbrella coordination deed that hardwires responsibility for interfaces, a clear framework for attributing delay and liquidated damages, and harmonised performance warranties and remedies across all packages to safeguard revenue start dates and protect financial covenants.

Some of the trends being seen in response to the increase in limited recourse project financings of data centre developments will be discussed in our forthcoming articles in this series.

Remote sites and/or utility constraints

Where a site is remote or has constrained access to utilities, the project will typically involve bespoke power and water solutions (e.g. new substations, private wire arrangements, on‑site generation, and tailored discharge consents) that may require packaging of works in ways that do not align neatly with standard procurement models with the need to consider interfaces with clear milestones, inter‑dependencies, and pass‑through of regulatory obligations. This requires early engagement with regulators and network operators and realistic scheduling to accommodate.

Pricing and risk sharing

Fixed-price D&C/EPC contracts remain popular if scope is well defined and developers are prepared to pay a premium. In this context, price escalation mechanisms are generally tightly drawn and limited to clearly defined inputs or indexation to recognised benchmarks. This structure aligns with lender expectations around cost certainty, but it places pressure on delivery in periods of supply chain volatility, for example where input prices or lead times move materially after contract close.

By contrast, target cost arrangements with pain/gain sharing are gaining ground on larger or fast-tracked programmes because they temper price risk and strengthen incentives to keep the programme on time. When combined with open‑book transparency, objective milestones and sensible caps and collars, these models can encourage collaboration without relaxing accountability for critical dates, energisation, and commissioning outcomes.

Pricing decisions should be grounded in a clear view of market capacity as a target cost with pain/gain sharing can attract more bidders and better align behaviours, especially when deployed alongside standardised designs and repeatable modular building blocks.

Early long lead equipment (LLE) commitments

The current 12–24 month lead times for critical plant (including transformers, switchgear, generators, chillers, liquid‑cooling systems, UPS units and batteries) make early capacity reservations a pragmatic necessity to secure manufacturing slots and de‑risk delivery.

Effective reservation agreements do more than hold a place in the queue: they clearly allocate commercial and operational risk through reservation fees, objective conditions precedent, and tiered cancellation mechanisms with pre‑agreed break costs. They also preserve continuity if a counterparty falters, by incorporating robust step‑in and assignment rights for developers, lenders and other project stakeholders.

Where developers elect to procure long‑lead equipment directly, the interface risk increases materially unless the approach is underpinned by a rigorous coordination deed and, where appropriate, novation to the general contractor. Effective management of this interface risk, by including appropriate contractual frameworks and insurance programmes, will be discussed in our forthcoming articles in this series.

Procurement models – a critical analysis