Will Nigeria get its first AI data centre in 2026? The data says it is likely

Nigeria’s push to position itself as a serious player in the global digital economy is increasingly converging on a single piece of infrastructure: data centres. As artificial intelligence workloads begin to reshape computing demand worldwide, the question is no longer whether Nigeria will build more data centres, but whether it is ready to host facilities capable of supporting AI at scale. Based on current projects, investment timelines, and technical constraints, the data suggests that Nigeria is closer than ever to commissioning its first true AI-focused data centre by 2026—but with important caveats.

Nigeria already hosts 17 operational data centres, with at least nine more either under construction or at advanced planning stages. One of the next additions is Equinix’s LG3 carrier-neutral facility on Victoria Island, Lagos, a 1-megawatt site scheduled for commissioning in the first quarter of 2026. This rapid build-out places Nigeria among Africa’s fastest-growing data infrastructure markets, driven by rising demand from cloud service providers, financial institutions, telecom operators, and digital-native businesses. While installed capacity currently stands between 65 and 86 megawatts, industry projections suggest it could climb beyond 400 megawatts within the next three to five years as new facilities are completed.

That build-out is anchored in a compelling investment story. Nigeria’s data centre market was valued at about $1.4 billion in 2025 and is projected to grow to $2.7 billion by 2035, representing an estimated compound annual growth rate of 7%, according to a Verraki report published in December 2025. Rising enterprise digitalisation, increased cloud adoption, the growth of fintech and e-commerce, and the emergence of AI-driven workloads are all fuelling demand. Consequently, local and international investors view Nigeria not only as a large domestic market but also as a strategic entry point into the wider West African digital economy.

Why AI changes the equation

Traditional enterprise data centres are designed around relatively modest rack densities, typically between 10 and 15 kilowatts per rack. AI workloads, particularly those involving large language models and GPU clusters, demand far more power, cooling, and network performance. At scale, AI-focused racks can require anywhere from 60 to 100 kilowatts per rack, often relying on liquid cooling and highly resilient power infrastructure.

Krish Ranganath, Regional Executive for West Africa at Africa Data Centres, a Cassava Technologies subsidiary with an operational 10 MW facility designed to scale up to 20.65 MW to meet growing AI and cloud demand, explains the difference clearly: “An AI data centre is fundamentally about high rack density.”

While supporting technologies such as cooling, networking, and power redundancy can be engineered over time, density is the defining constraint. Under current conditions, Ranganath notes that a 25-kilowatt rack can support entry-level AI workloads if the facility is designed to scale. However, truly AI-native deployments require expandability and long-term power certainty.

The challenge is that data centres, whether conventional or high-density, take time to build. Typical construction timelines range from 16 to 20 months, even before accounting for power connections, imported equipment, and commissioning. While prefabricated solutions can shorten deployment cycles, they are difficult to deploy at hyperscale and often come with technical limitations.

Economic impact strengthens the case

Beyond digital capability, the economic impact of data centres strengthens the argument for accelerated investment. Modelling by Verraki shows that a $10 million, 1-megawatt Tier III data centre generates approximately $17 million in economic output during the construction phase alone. When operational expenditure and refresh capital expenditure are included, cumulative economic output exceeds $39 million over ten years.

Employment effects are equally significant. A single 1-megawatt facility supports roughly 700 construction jobs and 20 to 30 operational roles annually, resulting in more than 1,600 cumulative jobs over a decade. These roles span engineering, power management, cooling, cybersecurity, and facilities operations, aligning closely with Nigeria’s push for skilled technical employment.

At an industry level, costs remain high. Modern Tier III data centres typically require between $10 million and $15 million per megawatt to build. Open Access Data Centres’ 24-megawatt Lagos facility, for instance, carries a reported cost of $240 million. Yet the multiplier effects—job creation, tax revenue, and demand for energy and ICT services—continue to attract capital.

Big players are already committing

Global and regional operators are no longer treating Nigeria as a speculative market. Open Access Data Centres announced a $500 million investment across Africa in 2021 and has since established operational facilities in Lagos and South Africa. In 2024, Equinix committed $390 million to the continent over five years. MTN Nigeria is building a 1,500-rack Tier IV facility, while Airtel Africa’s Nxtra project in Nigeria is expected to go live by the first quarter of 2026 as part of a broader hyperscale strategy.

Airtel’s Lagos facility is notable because it is being designed specifically for AI compute rather than traditional cloud storage. The project represents a $120 million investment, with early shipments of high-performance GPUs already delivered in late 2025. This marks a shift from “AI-ready” marketing toward explicit AI workload support.

Kasi Cloud and the strongest AI signal yet

The strongest signal that Nigeria could host its first true AI data centre by 2026 is Kasi Cloud’s flagship campus in Lekki, Lagos. As of January 2026, the LOS1 facility is in its final stages of completion, with some parts of the facility already in use, and is widely regarded as one of the most ambitious hyperscale developments on the continent. Backed by a $250 million investment and supported by the Nigeria Sovereign Investment Authority, the project was conceived to handle the extreme power densities and cooling demands associated with AI and advanced cloud workloads.

Built on a 4.2-hectare site with roughly 172,000 square feet of white space, the campus is designed to scale significantly. At full build-out, it can host between 3,000 and 4,000 racks across multiple data halls. Power infrastructure is a defining feature: the site is anchored by the largest dedicated data centre substation in Africa, with a total capacity of up to 100 megawatts. The first phase is designed to deliver between 32 and 44.4 megawatts of critical IT load, supporting rack densities that range from conventional 8 kilowatts to peaks of 100 kilowatts per rack, a threshold typically associated with liquid-cooled AI systems.

Location and resilience further strengthen the case. The campus is situated along the Lekki corridor, adjacent to several submarine cable landing stations, providing low-latency connectivity to global networks. It is engineered to Tier IV reliability standards and is targeting up to 95% renewable and carbon-free energy usage. 

According to Alex Tsado, co-founder and COO of Ahura AI and a founding member of the Alliance for Africa’s Intelligence (Alliance4AI), the facility has already opened its doors. He said Kasi Cloud is optimised for AI GPUs and is working in partnership with UduTech, a GPU cloud platform that accelerates AI innovation across Africa, founded by Tsado, to provide GPU cloud services tailored to regional AI demand. 

“UduTech plans to partner with them and MSI to link the GPUs to its cloud platform,” Tsado told TechCabal. “When the GPUs aren’t being used for gaming, others can rent them to run AI workloads at low cost, generating revenue for the GPU owners as well. Essentially, it’s a model where distributed GPUs earn money for everyone involved.”

GPUs, power, and the remaining constraints

While infrastructure readiness is improving, constraints remain. Almost all specialised equipment, from GPUs to cooling systems, is imported, exposing projects to currency volatility and supply chain delays, according to Ranganath. High-quality, reliable power remains the single most critical factor, alongside dense network connectivity capable of handling large data flows.

There are also transitional models emerging. Rack Centre’s 12-megawatt LGS2 facility, launched in 2025, is marketed as AI-ready, while partnerships such as NVIDIA and Cassava Technologies’ $700 million pan-African initiative aim to deploy thousands of GPUs across Africa Data Centres facilities, including in Nigeria. These deployments are designed to close the computing gap for startups that previously depended on expensive foreign cloud credits.

“There are probably other AI-ready data centres or GPU system operators that I’m not aware of, but I believe Rack Centre is AI-ready,” Tsado said.

The evidence suggests that Nigeria is unlikely to flip a switch overnight into full hyperscale AI dominance. However, at least one AI-focused facility is likely to go live by late 2026, particularly as projects like Kasi Cloud and Airtel Nxtra move from construction to commissioning. As Ranganath cautions, many projects remain in redesign stages, and timelines are sensitive to power availability and execution risk.