Part 1: Commissioning Is No Longer an End-Stage Activity

I recently hosted a panel, one of the first of its kind, at one of the industry’s biggest events, to discuss the role of commissioning in the data centre industry. The room brought together perspectives that are usually kept in separate conversations: design, the OEM, the main contractor, the commissioning authority, and the client. This mix mattered, because the questions we are now facing in this industry do not belong to any one of those seats, but they sit across all of them.

I set the discussion against a simple framing. The industry still talks about commissioning as a milestone at the end of construction, while modern data centre delivery increasingly depends on commissioning-led thinking from concept design onward. Underneath that sits a problem we don’t voice often enough: we are still trying to deliver these projects with the methods used for the last ten years, on a scale those methods were never built for. 

What struck me over the course of the session was how little disagreement there was on that point, and how much work the industry still has to do to act on it. These are the takeaways that stayed with me.

Commissioning is no longer an end-stage activity, and pretending otherwise is now a risk

The traditional model is well understood. In very simple terms, you complete construction, hand the building over to commissioning, and prove the facility at the end. That sequence quietly depended on three things being true: 

- Time at the back end to absorb problems

- A building that was substantially complete before systems went live

- A single clean handover to work towards.

None of those assumptions holds for the projects being delivered today. Clients are asking for significantly larger campuses and far greater capacity, with timelines that have not materially expanded since five or ten years ago, while equipment lead times have grown, delivery has become phased, and grid constraints now dictate when and how power even becomes available. When you compress all of that together, the back-end buffer that the old model relied (whether officially or unofficially) upon is simply not there. Commissioning has to start influencing the project from the design stage, or it will spend the whole programme trying to recover ground it was never given in the first place.

This is the part the industry has been slow to confront. The scale has changed beyond recognition, but the methods and thought process have not. Doing more of the same, faster and with more people, is not a method for a different scale. It is the old method under strain. A tenfold increase in capacity is not a bigger version of the same project. It is a different problem and requires a different approach to sequencing, assurance, and operational control, not just a longer task list.

It starts at design, not at energisation

The clearest consensus on the panel was that the discipline begins in design. Designing for commissionability is not a refinement you add later. It determines whether you can stage energisation, properly prove redundancy, run partial-load testing, and re-commission in the future without taking the facility apart. When that thinking is present at Stage 2 and Stage 3, the rest of the programme has somewhere to stand. When it is absent, you discover the gaps at the worst possible moment, when systems are live, and the cost of change is highest.

This is where the industry still confuses the two definitions. A commissioning plan written after the design is complete is not lifecycle commissioning. It is a closeout with an earlier start date. The discipline means the assurance logic shapes the design, not the other way around.

Commissioning has become an operational safety discipline as much as a technical one

This was the point I most wanted the audience to sit with. On a phased delivery, parts of the facility are live while other parts are still being built, and the moment that happens, commissioning is no longer only about proving that equipment works. It becomes the function that controls risk in a live environment through energisation management, switching procedures, safe systems of work, and clear accountability between the construction, commissioning and operations teams, who are now sharing the same site at the same time.

When commissioning is treated as a back-end activity in that environment, the safety exposure is real, not theoretical. You have live systems, incomplete construction, evolving client requirements, and teams under sustained programme pressure. Commissioning and operational control is the thread that holds that together: control of energisation, control of sequencing, control of operational boundaries, control of live systems. That is a safety discipline, and the industry needs to resource it as one.

The front line of assurance is moving off site

As facilities get larger and more modular, more is being proven before the equipment reaches the site. Factory testing is no longer a checkbox exercise. When done properly, it removes significant integration risk from the programme. That shift demands more from OEM capabilities, more trust from clients, and a commissioning process mature enough to carry factory evidence forward rather than repeat the work on-site. It does not remove the need to physically prove things on site before operational acceptance. It changes what remains to be proven and when.

Where this leaves us

The industry is being asked to deliver significantly more capacity, with far greater complexity, on timelines that have not expanded at the same pace. The old sequence of build first and prove later does not survive that reality.

Commissioning is no longer a back-end activity. It is a discipline that has to shape the project from the point decisions are made: in design, procurement, factory certification, sequencing, energisation planning and operational safety.

That does not mean commissioning takes ownership of every risk. It means the assurance logic has to be present early enough to influence the decisions that determine whether those risks can be managed well.

The businesses that deliver well over the next few years will be the ones willing to share information earlier, challenge assumptions honestly and accept that assurance is a shared responsibility rather than a baton passed down the line.

Part 1 of this discussion has focused on why commissioning must be present earlier.

Part 2 will look at why it must also continue longer, because a facility can be considered  ‘complete’ without being proven.