In short
- Design and permits take the same time either way; modular saves time after permits by building in the factory in parallel.
- Reported energisation: 12–18 months modular against 24–36 months conventional.
- Cost: about 10–15 % lower when labour, weather and rework are counted.
- On a 20 MW campus, opening 15 months earlier at $150 per kW-month is worth about $45M of revenue.
Where the time is saved
| Phase after 6 months design + 6 months permits | Conventional | Modular |
|---|---|---|
| Factory build | — | 6 months, starting during permits |
| Civil, structure, site works | 8 months | 4 months |
| MEP fit-out / install and connect | 10 months | 2 months |
| Commissioning | 3 months | 2 months |
| Total programme | ≈ 30 months | ≈ 19.5 months |
The saving depends on starting factory work early, before permits are final, and on the factory slot being available. The schedule comparator lets you change design and permit times.
Cost
Equipment costs about the same either way. Builders report 10–15 % lower total cost for modular, from less site labour, fewer weather delays and less rework. At the 2026 average of $11.3M per MW, a 20 MW build costs about $226M conventionally and $192–203M modular. Transport, cranes and factory margin eat into the saving on small or remote projects.
The value of opening earlier
For most operators this is the bigger number. A 15 % build saving on $226M is about $34M; a year of earlier revenue can exceed it. If the grid connection, not construction, sets the opening date, the value of speed falls to zero.
Side by side
| Criterion | Conventional | Modular |
|---|---|---|
| Time to energisation | 24–36 months | 12–18 months |
| Build cost | Baseline | 10–15 % lower reported |
| Quality control | Site conditions, many trades | Factory process, integrated tests |
| Design freedom | High | Limited to the product range and road sizes |
| Site labour | High, skilled MEP trades | Low; mainly connections |
| Scaling | Large phases | Add pods and skids as demand grows |
| Logistics | Many small deliveries | Fewer, larger loads; cranes, permits |
What changes for commissioning
Modular moves much of the testing to the factory. Integrated factory tests of whole skids replace part of site L3 and L4. Site testing still proves the joints and, above all, the connected system in the integrated systems test. Guidance on the levels is on allinCX.
How to choose
- Is construction or the grid on the critical path? If the grid, speed is worth little.
- Will you repeat the design? Repetition is where modular pays.
- Is skilled site labour scarce or expensive?
- Can the roads and the site take large modules and cranes?
- Does a standard product meet your requirements without heavy customisation?
Three or more "yes" answers point to modular, at least for the power and cooling plant.
Questions about modular vs traditional
Is modular always faster?
Only after permits and only if the factory starts early. Design, permitting and grid connection take the same time either way.
Is modular cheaper?
Builders report 10–15 % lower cost, mainly from site labour and rework. The larger financial gain is usually earlier revenue.
Can I mix modular and conventional?
Yes, and most large projects do: a conventional shell with prefabricated power skids, e-houses, cooling modules and pre-cabled rack rows.
What are the risks of modular?
Factory capacity and slots, transport damage, crane and road limits, and dependence on one supplier's product range.
Sources
- Datacentres.com, Sep 2026: 24–36 vs 12–18 months; 10–15 % lower cost; $11.3M per MW average.
- Phase durations follow the schedule comparator's planning assumptions; revenue rate is an assumption.