Kenya’s Data Centre Boom Faces a Power-Wheeling Test

Last Updated 2 hours ago by Kenya Engineer

Kenya has much of what Africa’s emerging data-centre economy needs: a relatively mature electricity market, a largely renewable generation mix, strong connectivity and a growing concentration of digital infrastructure in Nairobi.

What it does not yet have, at least at the scale the industry is beginning to contemplate, is an easy way of bringing all those advantages together.

That was one of the more consequential questions to emerge from ITW Africa and Datacloud Africa in Nairobi this week, where discussions about the future of data centres repeatedly came back to a deceptively simple issue: how will large digital facilities actually get the electricity they need?

The issue is becoming more pressing as artificial intelligence pushes data-centre power requirements higher. Kenya’s Principal Secretary for ICT and the Digital Economy, Stephen Isaboke, put the problem unusually bluntly during the conference, arguing that power planning can no longer be separated from digital infrastructure planning. He described AI itself as physical infrastructure because the applications ultimately depend on fibre, cloud platforms, data centres, electricity and trusted data.

The official programme had already identified power as the biggest constraint to scaling digital infrastructure in Africa. Its Digital Infrastructure Energy Summit was built around the premise that reliability, cost and long-term electricity supply are now central to both deployment and investment decisions.

For Kenya, however, the challenge is not simply a shortage of generation.

Kenya’s unusual advantage

George Aluru, chief executive of the Electricity Sector Association of Kenya, offered useful context during a panel on the rules shaping power for digital infrastructure.

Kenya has spent more than two decades reforming its electricity market. Generation and transmission were progressively unbundled, while independent power producers became part of the electricity landscape. Aluru described Kenya as one of the more advanced African markets in terms of electricity-sector deregulation.

The country also has a significant renewable-energy advantage.

According to Aluru, geothermal accounts for roughly 40 per cent of supply, while hydro, wind and solar together make up a large majority of installed capacity. At times, renewable sources account for 90 per cent or more of electricity actually dispatched.

For data-centre operators facing growing pressure to reduce carbon intensity, that is a potentially important competitive advantage.

The problem is that generation capacity and usable power are not the same thing.

A wind or solar project may be hundreds of kilometres from Nairobi. A data-centre operator may want to contract renewable electricity directly, but the electricity still has to travel through a transmission system governed by technical, commercial and regulatory rules.

That is where wheeling becomes important.

From buying electricity to moving electricity

The principle behind wheeling is relatively straightforward. A customer can contract electricity from a generator in one location while using the transmission or distribution network to deliver that power to a different location.

In practice, it is considerably more complicated.

The ITW discussion pointed to South Africa, where large industrial customers have increasingly used wheeling arrangements to procure renewable electricity. But the same experience has also exposed another infrastructure constraint: transmission capacity.

Aluru noted that South Africa has substantial wind and solar resources but faces transmission bottlenecks between generation areas and major demand centres.

Kenya is beginning to confront a related question.

During the conference discussions, participants pointed to renewable generation in northern Kenya and transmission infrastructure capable of carrying substantially more power towards Nairobi than it currently does. One suggestion was that the area around existing transmission corridors could eventually become attractive for energy-intensive facilities, including data centres, particularly where generation, transmission capacity and land can be brought together.

That is a different way of thinking about the data-centre market.

Instead of asking only where is the fibre?, developers may increasingly have to ask where is the power, where is the transmission capacity and where can the two be economically connected?

The regulator may become part of the commercial model

There is already a regulatory pathway for private transactions. As explained during the conference, a data-centre operator and an electricity producer can structure a bilateral transaction and submit it to the regulator for approval. But industry participants argued that the real test is whether the regulatory framework becomes sufficiently operational to support multiple transactions at scale.

That distinction matters. A regulation can exist on paper while still being difficult to use commercially.

For a data-centre investor making a multi-year, potentially hundreds-of-millions-of-dollars infrastructure commitment, uncertainty over power procurement can become a bigger issue than the headline electricity tariff.

This is why the conference repeatedly linked power, regulation and investment.

Isaboke made a similar point from the government side, telling investors that the country needs predictable rules, faster permitting, predictable demand and confidence that projects will be delivered and operated effectively. “Need does not equal capital flow,” he said in a broader discussion on digital infrastructure investment.

That may be the more important lesson.

The data centre is becoming an electricity customer unlike other customers

A large data centre is not simply another commercial building asking for a connection.

Its load can be large, continuous and highly sensitive to interruptions. The long-term development of AI infrastructure could make the relationship even more demanding as higher-density computing increases both electrical and thermal requirements.

A session on long-term power models for digital growth illustrated the problem from the utility side. One speaker described the difficulty of approaching a utility with a request that may start at 10 MW and eventually scale towards 100 MW. From the utility’s perspective, the land, load profile, timing and actual consumption have to make sense before infrastructure can be planned.

That suggests that future data-centre development in Kenya may need to become much more closely coordinated with power-system planning.

The conversation may therefore move from “Can Kenya generate enough electricity?” to more precise questions: Where will the load be? When will it arrive? Which transmission assets will serve it? Who pays for new infrastructure? How will renewable power be contracted? And how will the system remain reliable when the data centre is consuming continuously?

Those are engineering questions, but they are also investment questions.

Kenya’s opportunity is that it enters this next phase with a relatively sophisticated electricity market and a strong renewable base. Its challenge is turning that advantage into a power-delivery model that is predictable enough for the next generation of digital infrastructure.

The country’s data-centre opportunity may ultimately depend not only on how much electricity Kenya can generate, but on how efficiently it can connect, contract and deliver it.