Nigeria is building more data centres, but power and water could decide who survives 

Nigeria’s data centre market is rapidly becoming the backbone of West Africa’s digital economy, and the growth is accelerating faster than most observers anticipated.

Driven by cloud adoption, new international submarine cable capacity, and growing regulatory pressure to keep data onshore, the market is projected to more than double – from US$374.15m in 2026 to US$782.82m in 2031.

Hyperscalers, telecom operators, and colocation providers are racing to build data centre facilities to capture the twin waves of enterprise cloud migration and government e-governance investment. The opportunity is real. But so is the risk of building it wrong.

The CBN’s data localisation directive – mandating local storage of all payment transaction data by January 1, 2027 – has transformed that opportunity into demand certainty. Industry estimates suggest that over 90% of Nigerian financial institutions currently host data on foreign cloud platforms, representing more than $1.1 billion in annual cloud spending subject to mandatory repatriation.

For the first time, the question is not whether demand exists – it does. The question is whether the industry can build sustainably enough to capture it without being destroyed by its own operating costs. Every megawatt of capacity added generates far greater indirect value across engineering, construction, power, telecommunications, and the wider digital economy. The socio-economic stakes are significant.

Yet the scale of the opportunity should not obscure the scale of the challenge. Nigeria’s installed IT load capacity is expected to nearly quadruple – from approximately 56 MW today to more than 218 MW – a fourfold increase that will stress every assumption about power availability, cooling capacity, and operational efficiency. Building more data centres is necessary but insufficient. What Nigeria needs is data centres capable of operating efficiently, reliably, and sustainably in one of the world’s most energy-constrained environments.

The energy constraint: between an expensive grid and a volatile alternative 

The energy calculus for Nigerian data centre operators has fundamentally changed. Grid power, once subsidized, now reflects full market tariffs. Diesel, the traditional fallback, is also no longer subsidized by government, moving instead with global oil prices and naira fluctuations.

For many Nigerian operators, energy already accounts for more than 35% of operating expenditure – the highest ratio in the world. At self-generated power costs of between self‑generated diesel power costs US$0.28–$0.33 per kWh, this is not merely a cost challenge – it is an unhedged structural exposure.

Power Usage Effectiveness (PUE) has therefore become more than an operational metric. Every 0.1 improvement in PUE translates directly into millions of naira saved annually per facility, materially influencing cash flows, asset valuations, refinancing potential, and investor returns.

Water: the overlooked crisis 

Energy is not the only elephant in the room. Water represents an equally significant and largely underestimated threat. Lagos already faces chronic water scarcity – as many as 60% of residents lack access to clean water. Yet a single large-scale data centre using conventional cooling technologies can consume approximately 5 million gallons per day, equivalent to the daily needs of nearly 50,000 people.

Investors, regulators, lenders, and host communities are now demanding greater transparency over social and environmental performance. Water can no longer be treated as an invisible operating input. Just as the industry has learned to optimise PUE, it must now apply the same discipline to Water Usage Effectiveness (WUE)

Why structure matters more than ever 

The convergence of rising energy costs, water scarcity, and AI-driven computing density has fundamentally changed what a data centre represents. It is no longer simply a container for technology. It is a strategic infrastructure asset whose engineering determines its economic performance over decades.

The engineering choices made today – floor loading, cooling architecture, and material selection – must be treated as capital investment decisions, because they will directly influence operating costs, financing, insurance, regulatory compliance, and long-term asset value. An AI server may be refreshed every two to four years, but the facility must deliver for ten years or more.

While servers, chips, and cooling technologies will continue to evolve, the physical structure will remain largely fixed for decades. The structure is therefore the constraint to future growth and the most critical investment decision an operator can make. This is resilient engineering.

Resilient engineering as competitive advantage: five critical dimensions 

The case for resilient engineering is clear: Nigeria’s data centre boom cannot be sustained by capacity alone. Efficiency, adaptability, and structural intelligence are now the true markers of competitive advantage. But what does resilient engineering look like in practice? It is not a single design choice, but a series of deliberate decisions across five critical dimensions that separate future‑ready assets from stranded ones

1. Design for density, not just floor space: Traditional facilities were designed around racks weighing 1,500–2,500 pounds. Fully configured AI racks can exceed 4,000 pounds. Many new AI-ready facilities are abandoning traditional raised floors in favour of slab-on-grade construction, where equipment sits directly on reinforced concrete foundations. Strengthening raised floors after the fact often costs more than designing the structure correctly from the outset.
2. Cooling as infrastructure: AI is rapidly pushing conventional air cooling towards its physical limits. Operators are moving towards direct liquid cooling, immersion cooling, and hybrid architectures. Liquid cooling requires different pipework, floor penetrations, and equipment layouts. Chillers, cooling towers, and liquid distribution units introduce substantial structural loads that must be accommodated from the earliest stages of design. Cooling architecture can no longer be treated as mechanical equipment installed after the building is complete – it has become a structural design consideration. The most resilient facilities will be those engineered with sufficient flexibility to accommodate multiple cooling pathways as technologies evolve.
3. Retrofit first, build second: The first instinct in Nigeria is often to build new. That instinct deserves to be challenged. Around the world, some of the most successful data centre developments are adaptive reuse programmes that transform former factories, warehouses, and manufacturing facilities – structures that often already possess the floor loading capacity, ceiling heights, and utility connections required for modern computing infrastructure. Adaptive reuse reduces development timelines, lowers embodied carbon, avoids lengthy permitting processes, and delivers capacity to market faster than new construction. In a market where regulatory changes are accelerating demand, speed is itself a competitive advantage. Viewed through an investment lens, it is not simply a sustainability strategy, it is a capital efficiency strategy.
4. Materials as long-term investments: The materials selected during construction influence far more than the initial capital cost of a project. They shape maintenance requirements, operational resilience, lifecycle costs, and environmental performance for decades. Innovations such as glass fibre reinforced polymer offer corrosion resistance and electrical insulation that improve durability in mission-critical environments. Engineered timber products are increasingly being adopted globally for non-critical support buildings, reducing embodied carbon without compromising functionality. Yet materials decisions are often evaluated on upfront procurement cost alone. They should instead be evaluated on lifecycle value.
5. Future-ready, not future-proof: No developer can confidently predict the evolution of AI chips, cooling technologies, or power densities over the next ten years. The practical objective is therefore to become future-ready rather than future-proof. Future-ready infrastructure is designed for adaptation rather than prediction: modular power systems, flexible cooling architecture, standardised interfaces, scalable plant rooms, and structural capacity that exceeds immediate operational requirements. The objective is not to build for one technology, but to preserve options as technologies evolve

Six questions every data centre investor should ask 

These five engineering dimensions translate into six investment questions that should now accompany every development or acquisition decision:

1. Is the structure designed for future computing density, not just opening-day requirements?
2. Does the cooling strategy preserve future options across both air and liquid technologies?
3. Have PUE and WUE been established as design objectives, not post-commissioning targets?
4. Have existing assets been evaluated before committing to new construction?
5. Are materials being evaluated on lifecycle value rather than procurement cost?
6. Is the CBN data localisation directive being treated as a strategic demand signal, not merely a compliance obligation?

Developers who deliver efficient, scalable, and sustainable capacity ahead of demand will compete on value. Those who delay will compete primarily on price.

Conclusion 

Nigeria’s digital economy needs more data centres. But the next generation of winners will not be those who own the largest facilities, they will be those who own the most resilient ones: assets that consume less, adapt faster, and continue generating value long after less flexible infrastructure becomes obsolete.

In a high-cost operating environment defined by rising energy prices, chronic water scarcity, and rapid AI-driven obsolescence, resilient engineering is not a nice-to-have. It is the defining competitive advantage of the next decade.

The decisions that will determine whether a data centre remains profitable in 2035 are not being made when the servers are installed. They are being made today, by investors, developers, engineers, lenders, and policymakers, long before the first rack arrives. The most valuable infrastructure is not simply the infrastructure that works. It is the infrastructure that continues to work, economically, operationally, and sustainably, for the lifetime of the asset.

Dr. Ulohomuno Eze Afieroho is an infrastructure investment and capital project delivery specialist who leads complex adaptive programmes at the intersection of infrastructure, governance, and economic development.