A data centre project can have a site, an investor, a design team and a construction programme and still be months away from being genuinely ready to start.
The reason is simple: many data centre project delays begin before construction begins.
Power availability may not be as straightforward as the nearest substation suggests. Land may have environmental, flooding, zoning or title constraints. Cooling assumptions may no longer match the planned AI workload. Permitting pathways may take longer than expected. And design decisions made too late can create cost and coordination problems that only become visible during construction.
For developers, hyperscalers, colocation operators and investors, the real challenge is not simply building faster. It is creating enough certainty early enough to make construction predictable.
Why This Matters
Data centre demand is growing rapidly, while the infrastructure required to support it is becoming more complex.
The International Energy Agency, utilities and industry bodies have highlighted growing pressure on electricity systems as large data centre loads connect to the grid. IEEE's 2025 report on data centre grid readiness notes that individual data centre loads can reach the gigawatt scale, creating new challenges for generation, transmission, interconnection and grid management.
At the same time, AI workloads are increasing rack power density and changing data centre cooling requirements.
That means a site that looked viable two years ago may require a very different technical assessment today.
What Is Actually Causing Data Centre Project Delays?
The most common mistake is treating site selection, design, approvals, power, cooling and construction as separate workstreams.
They are not.
A decision in one area can create consequences across the entire project.
For example:
Land → available site area → building footprint → electrical infrastructure → cooling plant → water demand → environmental approvals → cost → schedule
If one assumption changes late, several downstream decisions may need to change with it.
This is why data centre due diligence needs to happen as an integrated technical exercise—not simply as a land purchase checklist.
1. The Site Is Available. But Is It Actually Data Centre Ready?
A parcel of land can appear attractive because it is large, accessible and close to a city.
But data centre site feasibility goes much deeper.
A proper assessment should examine:
Land ownership and title
Zoning and permitted land use
Development restrictions
Topography and grading requirements
Soil and geotechnical conditions
Flood and seismic risks
Environmental constraints
Water availability
Fibre connectivity
Road and logistics access
Proximity to substations
Grid capacity and connection feasibility
Permitting requirements
Future expansion potential
For example, a site may be physically large enough for a 100 MW campus but lack sufficient electrical infrastructure or have environmental restrictions that limit development.
The cheapest land is not necessarily the lowest-cost data centre site.
Desapex's site selection and feasibility approach combines land suitability, power infrastructure, connectivity, environmental risks, regulatory requirements and technical due diligence before a Go/No-Go decision.
2. Grid Power Can Become the Critical Path
For a data centre, power is not simply an operating requirement. It can determine whether the project can proceed at all.
A common assumption is:
"There is a substation nearby, so power should be available."
That is not enough.
The assessment needs to consider:
Existing substation capacity
Available load
Incoming feeder configuration
Transmission and distribution constraints
Interconnection requirements
Redundancy strategy
Utility upgrade requirements
Connection timelines
Power quality
Future load growth
Renewable energy availability
Grid connection delays are increasingly being reported as a major constraint for new data centre developments because utilities are dealing with capacity, permitting and infrastructure limitations.
A simple scenario
Imagine a developer acquires land expecting a 60 MW connection.
During detailed utility discussions, the available capacity is found to be insufficient without infrastructure upgrades.
The project now requires:
Utility study → grid upgrade → equipment procurement → approvals → construction → energisation
The building design may be progressing perfectly, but the project cannot operate without the power connection.
This is how a pre-construction assumption becomes a construction schedule problem.
3. Cooling Decisions Are Moving Earlier
AI is changing the relationship between power and cooling.
Higher computing density means more electrical power is converted into heat.
That creates a chain reaction:
Higher compute density → higher rack power → higher heat density → advanced cooling → more complex MEP infrastructure
This matters during site selection because cooling strategy can affect:
Water availability
Cooling plant area
Electrical load
Mechanical plant rooms
Structural requirements
Piping distribution
Maintenance access
Energy performance
Future expansion
A site with limited water resources, for example, may require a fundamentally different cooling strategy.
Cooling should therefore be evaluated alongside power and site feasibility not after the architectural concept is complete.
4. Environmental and Regulatory Requirements Can Change the Schedule
Data centres interact with multiple environmental and planning requirements.
Depending on location, projects may need to address:
Environmental approvals
Water availability and usage
Wastewater management
Noise
Air quality
Flood risk
Land-use permissions
Energy requirements
Renewable energy considerations
Construction permissions
These requirements can influence the development timeline and even the technical design.
Recent data centre projects have demonstrated how environmental and permitting issues can become material schedule risks.
The lesson is not that every project will face the same issues.
It is that regulatory risk needs to be identified before the project becomes financially committed.
5. Design Coordination Problems Start Before Construction
Another source of data centre construction delays is fragmented design.
Electrical, mechanical, fire protection, structural, architectural, controls and IT infrastructure all compete for limited space.
A late change to one system can affect several others.
Consider a typical MEP coordination issue:
A chilled-water pipe route conflicts with a cable tray.
The routing changes.
That affects a structural opening.
The structural opening changes.
The ceiling coordination changes.
Installation is delayed.
A relatively small design coordination issue has now created a construction problem.
This is where BIM becomes more than a visualisation tool.
A coordinated BIM environment can help teams identify spatial conflicts, validate installation requirements and connect design decisions with construction planning.
Desapex uses BIM-supported site coordination, 4D/5D execution support, digital QA/QC and progress monitoring as part of its digital construction management approach.
6. Cost Overruns Often Begin With Unresolved Assumptions
A data centre cost overrun is not always caused by poor construction productivity.
It can originate much earlier.
Examples include:
Underestimating utility upgrade costs
Inadequate site grading allowances
Late cooling-system changes
Insufficient electrical infrastructure
Design changes caused by AI workload requirements
Rework from MEP clashes
Long-lead equipment substitutions
Regulatory changes
Unplanned enabling works
This is why data centre cost planning should evolve with the design and engineering model.
5D BIM can connect quantities and cost information with the developing design, helping project teams understand the financial effect of design changes before they reach site. Desapex provides 5D BIM and digital project-control capabilities specifically for data centre projects.
A Better Way to Think About Data Centre Readiness
Instead of asking:
"Can we build a data centre on this site?"
ask:
"Can this site support the required power, cooling, connectivity, approvals, construction logistics and future expansion within the required cost and schedule?"
That changes the entire feasibility process.
A useful early-stage framework is:
1. Land
Is the site legally, physically and environmentally suitable?
2. Power
Can the required electrical capacity be delivered within the required timeframe?
3. Connectivity
Are diverse fibre routes and required network connections available?
4. Cooling
Can the site support the required thermal strategy and future rack densities?
5. Approvals
What environmental, planning and statutory approvals are required?
6. Engineering
Can the facility be designed to meet resilience, redundancy and operational requirements?
7. Construction
Can the project be constructed, coordinated and commissioned within the target programme?
8. Expansion
Can the campus scale as IT demand increases?
What Companies Should Do Before Breaking Ground
A practical data centre project-readiness process should include:
01 — Perform integrated site due diligence
Evaluate land, power, water, fibre, environment, zoning, title, logistics and hazards together.
02 — Validate power with technical evidence
Do not rely solely on distance to a substation. Validate capacity, connection strategy and infrastructure requirements.
03 — Define the IT workload early
AI and high-density computing can materially change power and cooling requirements.
04 — Develop a coordinated BIM environment
Use a common digital model to connect site conditions, engineering design and construction requirements.
05 — Identify long-lead equipment
Transformers, switchgear, generators, chillers, CDUs and other critical equipment can influence the programme.
06 — Link design, schedule and cost
Use 4D/5D workflows to understand how design decisions affect construction sequencing and project cost.
07 — Create a decision-gate system
Before moving from land acquisition to design, and design to construction, validate the assumptions that could stop the project.
What Changes Next?
The next generation of data centre development will increasingly rely on digital engineering, AI-ready infrastructure, liquid cooling, modular construction, reality capture and integrated project controls.
AI workloads will continue to push rack densities higher, increasing the importance of power and thermal planning. Industry research is already tracking a shift towards liquid and hybrid cooling for high-density workloads.
At the same time, digital twins and BIM-based workflows can create continuity from site feasibility → design → construction → operations rather than treating every stage as a separate information silo.
That digital thread becomes particularly valuable when projects are large, complex and schedule-sensitive.
The Real Question Is Not "When Can We Start Construction?"
It is:
"How many uncertainties can we remove before construction starts?"
Because a data centre rarely slips only when something goes wrong on site.
It slips when an assumption made months earlier turns out to be wrong.
The earlier those assumptions are tested power, land, cooling, connectivity, approvals, design, cost and constructability the more predictable the project becomes.
How Desapex Helps
Desapex supports data centre developers, hyperscalers, colocation providers and enterprise clients across the project lifecycle from site selection and power infrastructure assessment to mission-critical MEP design and digital construction management.
Our approach connects technical due diligence, BIM, engineering, construction coordination and project controls into a continuous digital workflow.
Explore:
FAQs
Why do data centre projects get delayed before construction starts?
Common causes include insufficient grid capacity, lengthy power-connection processes, land and title issues, environmental constraints, permitting, connectivity limitations, incomplete due diligence and unresolved design requirements.
How does grid power affect data centre project timelines?
A site may have a nearby substation but still lack sufficient capacity or the required connection infrastructure. Utility upgrades, interconnection studies, approvals and equipment lead times can therefore become critical-path activities.
What should data centre site due diligence include?
Data centre due diligence should assess land, title, zoning, power, fibre connectivity, water, environmental constraints, flood and seismic risks, topography, logistics, permitting and future expansion potential.
How does cooling affect data centre site selection?
Cooling strategy influences water requirements, electrical demand, mechanical plant space, energy performance and infrastructure layout. AI workloads and higher rack densities are also increasing the need to evaluate liquid and hybrid cooling strategies earlier in the project.
What causes data centre construction cost overruns?
Cost overruns can result from design changes, utility upgrades, inadequate site preparation, MEP coordination problems, equipment price changes, rework, schedule extensions and requirements that were not identified during early feasibility.
How can BIM reduce data centre construction delays?
A coordinated BIM environment can help identify MEP clashes, validate spatial requirements, support construction sequencing, improve coordination and connect design information with schedule and cost management.
When should data centre feasibility begin?
Ideally, technical feasibility should begin before land acquisition or major development commitments. Early validation of power, land, cooling, connectivity, environmental and regulatory constraints can prevent expensive downstream changes.