What Hyperscale Operators Actually Expect from Construction Teams

They expect you to understand the commissioning sequence before you break ground

Commissioning is where hyperscale builds are won or lost. It is also where the evidence of every earlier decision shows up, for better or worse.

Most contractors think of commissioning as something that happens at the end. Operators know it is the outcome of everything that came before. The way MEP systems are installed, the way inspections are signed off, the way defects are tracked and closed, all of it feeds directly into how smoothly or how painfully commissioning runs.

The numbers bear this out. According to a 2022 Uptime Institute study, up to 80% of data centre outages involve components or sequences that were never directly tested during commissioning. They were assumed to work based on similar equipment that was tested. That assumption is where most commissioning failures live.

The five-level commissioning process used in hyperscale data centres is not a formality. It is a structured gate system. Each level must be proven complete before the next begins. Level 1 validates equipment at the factory. Level 2 confirms installation against drawings and specifications. Levels 3, 4, and 5 progressively test individual systems, integrated systems, and full-load scenarios before handover. Any defect that surfaces at Level 3 or beyond carries a cost multiplier that dwarfs what it would have cost to catch it during construction.

Operators expect construction teams to understand this. Not just conceptually, operationally. They want to see quality workflows that are designed around the commissioning sequence, not bolted on after the fact.

They expect progressive quality evidence, not end-of-project documentation

The financial stakes make this non-negotiable. Research from STL Partners puts the cost of a one-month delay on a typical 60 MW data centre at $14.2 million, made up of lost lease revenue, compounding construction overruns, and tenant penalty exposure. A three-month slip can cut project IRR almost in half. When commissioning runs long because quality was not managed progressively through the build, those costs fall on someone. Usually the GC.

One of the most common complaints from hyperscale clients is not that contractors fail to produce documentation. It is that the documentation appears all at once, at the end, in a format that cannot be interrogated.

A folder of completed ITPs uploaded to a shared drive the week before practical completion is not quality assurance. It is paperwork. The difference matters enormously to an operator who needs to verify that critical systems were installed correctly, inspected at the right stages, and defect-free at the point of energisation.

What operators want is a live, traceable record that builds throughout the project. They want to be able to see at any point in the build which inspections have been completed, which defects are open, and which systems are cleared for the next commissioning level. They want evidence that is current, not compiled.

This is not about surveillance. It is about confidence. When a hyperscale operator is managing a 200MW campus with multiple contractors working across civil, structural, and MEP packages, they need an assurance layer that survives subcontractor changes, site turnover, and package handoffs. A system that tracks quality progressively across every contractor and every system is what gives them that.

As one detailed analysis of hyperscale construction governance noted, weekly or monthly reporting cycles no longer provide adequate signal. Leading teams operate with near real-time data, allowing deviations to surface within days rather than weeks.

They expect consistency across phases and across sites

The phased delivery model is standard in hyperscale construction. An operator building a 300MW campus does not commission the whole thing at once. They bring phases online sequentially, often with aggressive timelines between them.

That creates a specific demand on construction teams: every phase must be built to the same standard as the last. Not approximately. Not close enough. The same inspection protocols, the same commissioning gating, the same defect management process, repeated reliably across every hall, every phase, every contractor.

This is harder than it sounds when builds span multiple subcontractors, multiple geographies, and teams that rotate between projects. The operators who have been doing this longest have learned that the only way to maintain consistency at scale is to standardise the quality system, not just the design. A repeatable QA process, documented and enforced across every package, is what separates a contractor who can deliver phase one brilliantly from one who can deliver phase one through phase six without regression.

Hyperscale operators also increasingly build across global portfolios. They want contractors who can replicate their delivery standard across sites, not just demonstrate it once.

They expect the build record to stand alone at asset transfer

This is a less-discussed expectation, but it matters more than most contractors realise.

Data centres have become an investable asset class. Global institutional capital is flowing into digital infrastructure because the demand profile is compelling. But capital wants assurance, not just that the building is standing, but that the systems inside it were built correctly, tested to the right standard, and will perform as the technical specifications promised.

When an operator divests an asset, sells a campus, or brings in institutional co-investment, the quality record becomes part of the technical due diligence. Future buyers and their technical advisors will ask hard questions: Can you show me that every critical system was inspected? Can you trace the defect history on the MEP infrastructure? Can you demonstrate commissioning completion at the level required for Tier III certification?

If the answer involves hunting through PDFs, chasing subcontractors for paperwork, and manually reconstructing a timeline from email threads, that asset is exposed. The technical due diligence process will surface the gaps, and those gaps translate into caveats, risk adjustments, and deal friction.

The most commercially sophisticated operators are already thinking about this during construction, not at divestment. They want a build record that can be interrogated by someone who was not on the project, a historian of how the asset was built that holds up under scrutiny years after handover.

They expect you to manage your subcontractors to their standard

Hyperscale projects are multi-contractor by nature. A large campus typically involves civil and ground works contractors, structural specialists, MEP packages split across mechanical, electrical, and plumbing trades, and specialist commissioning agents who work independently of the construction team.

The hyperscale operator does not want to manage quality across all of those parties. That is the builders job. They expect one consistent quality standard to apply across the entire site, regardless of which subcontractor is working in which hall.

This is where many General Contractors run into trouble. They have strong internal quality systems but limited visibility into subcontractor QA. Defects logged in one system, inspection records in another, commissioning data in a third. The result is a fragmented picture that serves no one, not the General Contractor, not the operator, not the commissioning agent trying to work from a complete system history.

What operators want is a unified quality record. One system, one inspection standard, one defect workflow applied consistently across every package. Not because it makes reporting easier, but because it is the only way to guarantee that a missing inspection on a subcontractor’s MEP installation does not become a commissioning failure six months later.

The expectation gap, and what it costs

There is a gap between what many construction teams think they are delivering and what hyperscale operators actually need. It is not usually a gap in intent. Most General Contractor’s working on data centres take quality seriously. The gap is in the system.

The cost of that gap is well documented. Research from CII, McKinsey, and Autodesk consistently puts rework at 5–15% of total project cost. On a $500 million data centre build, that is between $25 million and $75 million in avoidable waste. And that is before accounting for the commissioning delays that rework causes downstream.

Spreadsheets break. Shared drives fragment. Paper-based ITPs disappear. Manual processes that work on a 50,000 square metre commercial project break down entirely on a 100,000 square metre hyperscale campus with 250 subcontractors and a commissioning window that cannot slip.

The operators who are building at scale have already worked this out. They are actively seeking contractors who can evidence their quality management capability, not just assert it. They want to see the system before they commit the contract.

For contractors who want to win data centre work and keep it, the question is straightforward: can your projects quality and commissioning process match the expectations of the client you are trying to serve? If not, the gap will show up. It always does.