As Kenya accelerates its rollout of Artificial Intelligence (AI) across key sectors like health, education and agriculture, experts are warning that the country’s energy infrastructure may not be ready for the demands of AI-powered data centres.
The government’s AI strategy also targets public service, creative industries, and small businesses—but the physical backbone of this digital transformation hinges on reliable electricity.
Global studies now show that the power needed to run generative AI systems could soon surpass the energy consumed by Bitcoin mining.
Earlier this year, iXAfrica launched a 4.5-megawatt (MW), AI-ready data centre in Nairobi, part of a 22.5MW campus along Mombasa Road. Once complete, the facility could serve over 300 million users in the region.
Other players like Africa Data Centres, PAIX, and icolo.io are positioning Kenya as a regional hub—but powering these operations remains a challenge.
By the end of 2024, Kenya’s installed electricity capacity stood at 3,811.6MW, with peak demand nearing 2,316MW. However, a government freeze on new power purchase agreements (PPAs) since 2018 has stalled clean energy expansion.
The freeze, based on a presidential taskforce recommendation, aims to align future energy projects with the Least Cost Power Development Plan.
Despite these constraints, President William Ruto last year announced plans for a one-gigawatt (1GW) green-powered data centre in Naivasha, in partnership with global tech giants including Microsoft and G42.
Globally, AI is projected to account for nearly half of all data centre electricity use by 2026, according to the International Monetary Fund (IMF).
AI systems like ChatGPT and Claude require massive computational resources—both during model training and when responding to users—making electricity a critical cost.
A related IMF report warns that AI-driven energy demand could push petroleum prices up by nine percent by 2026.
“As substantial computational resources are required during both stages, electricity consumption represents a critical input for companies delivering AI services,” reads the report.
The International Energy Agency (IEA) also notes that AI adoption and rising air-conditioning use are now the biggest contributors to the global rise in electricity demand.
“Global electricity consumption is expected to increase at the fastest pace in years over the 2025-2027 period, fueled by rising use of air conditioning and the expansion of data centres worldwide,” wrote IEA.
While AI is improving energy efficiency and spurring uptake of renewables, experts caution that these gains are being outpaced by the sheer volume of new hardware and data centre builds.
Kenya’s challenge is now twofold: keeping pace with digital ambitions, while expanding power capacity to support the AI future already knocking on its door.
