Nigeria’s AI Boom Could Run on Gas. Can It Find a Greener Way?
Nigeria’s digital ambitions face an important energy decision Nigeria’s push to build data centres and attract artificial intelligence investment is creating an energy challenge that cannot be solved simply by buying mor…
Nigeria’s digital ambitions face an important energy decision
Nigeria’s push to build data centres and attract artificial intelligence investment is creating an energy challenge that cannot be solved simply by buying more servers. The facilities that store information, run cloud services and process AI tasks need dependable electricity every hour of the day. With the national grid unable to guarantee uninterrupted supply, operators are turning to gas-fired plants and diesel generators. That may keep the computers running, but it risks tying the country’s digital growth to fossil fuels just as the global technology industry faces pressure to reduce its carbon footprint.
A report published by BusinessDay on November 16, 2026 describes the dilemma facing Nigeria’s expanding data-centre industry. The question is not whether the country should develop AI infrastructure. It is how to power that infrastructure reliably without creating expensive, long-lived energy systems that undermine environmental goals.
Why artificial intelligence needs so much electricity
AI depends on computing equipment that performs enormous numbers of calculations. Those calculations run on processors, particularly graphics processing units (GPUs), installed in server racks inside data centres. The more demanding an AI model becomes and the more people use it, the more computing work the facility must perform.
That equipment consumes electricity and produces heat. Data centres therefore need power not only for servers but also for cooling systems, networking equipment, lighting, security and backup infrastructure. Cooling is especially important in Nigeria’s hot climate because equipment must remain within safe operating temperatures. If cooling fails, servers can overheat, slow down or shut down.
The scale is larger than many people imagine. The International Energy Agency (IEA) estimates that data centres used about 415 terawatt-hours of electricity worldwide in 2024, equivalent to around 1.5 per cent of global electricity consumption. Its 2025 report projects that demand could rise to approximately 945 TWh by 2030, with AI among the main drivers. The IEA says a typical AI-focused data centre can consume as much electricity as 100,000 households.
Think of a data centre as a large digital factory that never closes. A normal factory may pause some operations at night, but cloud platforms, banking services, online communication and AI applications are expected to remain available around the clock. Even a short power interruption can disrupt services, damage operations or force companies to switch to backup systems. Reliability is therefore a business requirement, not a luxury.
Why Nigeria faces a difficult choice
Nigeria’s electricity challenge makes the expansion more complicated. Data-centre operators cannot depend entirely on public supply when outages and fluctuations threaten continuous operations. They need backup generation and, in some cases, dedicated power plants.
BusinessDay reports that natural gas and diesel remain important in operators’ power strategies. Gas can produce electricity with lower direct carbon dioxide emissions than diesel for an equivalent amount of power, but it is still a fossil fuel. Methane leaks during production and transportation can also increase its climate impact. Diesel generators add local air pollution and carbon emissions, particularly when used frequently.
The scale of planned projects shows why the choice matters. BusinessDay reports that Tetracore Energy Group is planning a 20-megawatt data centre in Ogun State supported by a proposed 100MW gas-fired power plant. Open Access Data Centres has discussed gas-powered capacity for its Lekki campus, while its chief executive has indicated that future AI-related requirements could raise power needs substantially. Airtel Nigeria is developing a 38MW hyperscale facility at Eko Atlantic, although BusinessDay says its specific energy arrangement has not been publicly established.
These are not all completed facilities or operating power plants. Some are plans, proposals or capacity targets. That distinction matters because projected electricity demand should not be reported as power already being consumed.
The risk is that investments made to solve today’s reliability problem could remain in use for decades. If new data centres are built around long-term gas infrastructure without a credible route to cleaner electricity, Nigeria may find it harder and more expensive to reduce emissions later.
There is also a question of fairness. Many households and small businesses still struggle to obtain reliable electricity. Data centres are essential to digital services and economic growth, but their power arrangements should not worsen pressure on an already constrained electricity system or divert scarce resources from wider public needs.
What greener power could mean
Greener data centres do not necessarily mean switching off generators overnight or relying on solar panels alone. The practical option is likely to involve a combination of energy sources, storage and efficiency measures.
Solar power can supply electricity during daylight hours, while batteries can store some of that energy for use after sunset or during short interruptions. Hydropower can contribute through the national grid where available. Grid electricity, solar, battery storage and carefully managed backup generation can be combined in a hybrid system. The right mix will depend on location, cost, reliability requirements and the amount of power a facility needs.
Nigeria already has experience with distributed solar systems and mini-grids. The World Bank’s Nigeria renewable-energy programme supports expanding electricity access through decentralised renewable solutions, including solar. That programme is not specifically designed for data centres, but it demonstrates that private investment and distributed clean-energy systems can help address gaps that the central grid does not immediately solve.
Efficiency is another part of the answer. Operators can use more efficient servers, improve airflow, adopt advanced cooling technologies and monitor electricity use closely. Liquid cooling, for example, can remove heat directly from high-performance computing equipment and may be useful for dense AI workloads. Better software and hardware management can also reduce the amount of computing power wasted on underused equipment.
BusinessDay reports that Rack Centre is pursuing a hybrid approach involving gas and solar, while the National Information Technology Development Agency has highlighted the use of solar, renewable-energy storage and grid connections. These examples suggest that the choice need not be framed as gas or renewables alone. However, a hybrid system is only as green as its actual energy mix. Operators should disclose how much electricity comes from each source rather than describing a facility as sustainable simply because it has installed some solar panels.
A global push for more responsible AI
The environmental question is gaining international attention. On November 16, the International Telecommunication Union and the COP31 Presidency launched the Antalya Artificial Intelligence Pledge. It calls for more sustainable AI infrastructure, including greater use of low- and zero-emission energy, improved energy and water efficiency, responsible hardware management and clearer reporting of AI’s environmental impact.
The pledge is not a Nigerian law or an automatic requirement for every local data centre. It is a global initiative that provides a framework for governments and companies to consider the environmental cost of AI development. Its relevance to Nigeria is straightforward: investment in computing infrastructure should be accompanied by investment in cleaner and more reliable energy.
What Nigeria should prioritise
Nigeria needs data centres to support local cloud services, financial technology, research, businesses and the wider digital economy. Building local computing capacity can reduce dependence on overseas infrastructure and create opportunities for skilled employment. But the country should measure success by more than the number or size of facilities announced.
Government agencies, electricity providers and data-centre operators should coordinate infrastructure planning so that new facilities are located where power can be supplied without undermining other users. Operators should publish credible information about energy consumption, emissions, water use and the share of electricity supplied by gas, diesel, solar and the grid. New projects should also assess battery storage, renewable power contracts and energy-efficiency measures from the design stage.
For Nigerians, the issue is not simply whether AI will use more electricity. It is whether the investment will strengthen the country’s energy system or create another set of expensive facilities dependent on private fossil-fuel generation. A reliable power supply is essential to the AI economy, but reliability and sustainability should be planned together.
Nigeria does not have to choose between digital development and environmental responsibility. It does, however, need to make its energy decisions early. If the country builds its AI future around cleaner power, efficient cooling and transparent energy reporting, data centres could support economic growth without locking the digital economy into a more polluting path.
Sources
- businessday.ng/technology/article/nigeria-risks-fossil-fuel-trap-as-global-ai-industry-turns-to-greener-data-centres/
- www.iea.org/reports/energy-and-ai/executive-summary
- www.worldbank.org/en/news/press-release/2023/12/15/nigeria-to-expand-access-to-clean-energy-for-17-5-million-people
- www.itu.int/en/mediacentre/Pages/PR-2026-11-16-COP31-Antalya-AI-Pledge.aspx
Version 1.0. Kobo Daily stories are edited before publication and updated as facts change. Spotted an error? Tell our editors.