Kenya’s Data Centre Setback - a lesson in the AI Value Chain
- Eastern Legacy
- May 7
- 6 min read
The reported stalling of Kenya’s $1 billion Microsoft-G42 data centre project is more than a local infrastructure story. It is a window into one of the most important realities of the AI era: artificial intelligence is not only a software revolution. It is an infrastructure race.
In 2024, Microsoft and G42 announced a major digital ecosystem initiative in Kenya. The package included a green data centre, a new Azure East Africa cloud region, local-language AI research, AI skills, connectivity investment, and work on trusted cloud and data governance. Reuters confirmed that the project was intended to expand cloud-computing services in East Africa and would be powered by geothermal energy.
Now, reports suggest the project has stalled because Kenya’s electricity system cannot yet support the required scale. The precise contractual status remains unclear; there is no clear public evidence that Microsoft, G42, or the Kenyan government has formally cancelled the project. But the strategic signal is already visible. AI infrastructure is colliding with national power systems.
The AI race is becoming an electricity race
The IEA projects that data centre electricity consumption could roughly double by 2030, while AI-focused data centres could grow even faster. That changes the economics of digital transformation.
For the past decade, countries could participate in the digital economy through connectivity, mobile money, e-government, software talent, and platform adoption. Those remain important. But frontier AI introduces a heavier infrastructure layer: accelerated computing, GPU clusters, high-density data centres, advanced cooling, resilient grids, and huge volumes of firm power.
This means energy policy and AI policy can no longer be separate.
A country that wants AI infrastructure must answer difficult questions. Can the grid absorb large new loads? Is power available at competitive cost? Is it clean, firm, and reliable? Can transmission infrastructure connect generation to data centre sites? Can the permitting system move quickly? Can water and cooling requirements be managed sustainably?
Kenya has strong renewable credentials, especially geothermal. But the regulatory data show the country’s interconnected installed electricity capacity remains limited relative to a possible gigawatt-scale data centre.
The real lesson: countries must choose their AI value-chain position
The biggest mistake emerging economies can make is to treat AI as a single industry. AI is a value chain.
At the bottom are physical inputs: electricity, chips, land, water, fibre, cooling, and capital. Above that sit compute infrastructure and cloud regions. Then come foundation models, data ecosystems, applications, adoption, regulation, and value capture.
Very few countries can dominate all layers.
For Kenya and many comparable economies, the strategic question is not whether to participate in AI. The question is where participation creates the most durable national value.
A hyperscale data centre may create strategic visibility, but it may not automatically create broad domestic value if the country lacks energy surplus, local cloud demand, AI firms, procurement capacity, and data governance. By contrast, applied AI in agriculture, fintech, logistics, health, public services, education, and climate resilience may generate more immediate productivity gains.
What Kenya still has
The reported stalling of the Microsoft-G42 data-centre project should not be mistaken for the disappearance of Kenya’s digital opportunity. Kenya may not yet be ready for the most power-intensive end of the AI infrastructure stack, especially if the requirement approaches hyperscale or gigawatt-scale compute. But Kenya still has several strategic assets that matter deeply in the AI value chain.
First, Kenya has a credible renewable-energy base, and geothermal remains its most distinctive advantage. EPRA’s latest available data shows that renewable sources supplied 78.79% of the energy to Kenya’s national grid between July and December 2025, while geothermal alone accounted for 40.06% of total interconnected-grid generation. Geothermal is especially important for AI infrastructure because it can provide steadier baseload power than intermittent renewables. However, the strategic issue is not simply whether Kenya has green power; it is whether that power can be delivered at the right scale, price, location, reliability, and contractual certainty for high-density compute.
Second, Kenya has strong regional connectivity. The Communications Authority of Kenya reported total equipped/lit international internet bandwidth capacity of 24,161.332 Gbps in Q2 FY2025/26, with capacity distributed across systems including SEACOM, TEAMS, EASSy, Lion2, DARE1 and PEACE. This matters because AI and cloud infrastructure do not depend only on power; they also depend on latency, redundancy, peering, and cross-border traffic flows. Kenya’s digital geography gives it a plausible role as an East African cloud and data-exchange hub.
Third, Kenya has a maturing broadband and digital-demand base. The National Broadband Strategy 2025–2030 states that broadband subscriptions rose from 22.08 million in June 2020 to 45.79 million in June 2025, while available international bandwidth rose from 7,392.26 Gbps to 22,311.445 Gbps over the same period. The strategy explicitly links broadband expansion to e-government, digital learning, e-commerce, mobile financial services, cloud computing, innovation, and smart service delivery. That makes Kenya more than a potential hosting location; it is also a demand market and regional distribution platform for digital services.
Fourth, Kenya has an existing and expanding data-centre ecosystem. Africa Data Centres describes its Nairobi facility as connected to carrier networks across Kenya and long-distance fibre routes to Uganda, Tanzania, Rwanda, Burundi, Ethiopia and Somalia, with diverse fibre routes to Mombasa cable landing stations. iXAfrica also positions its Nairobi campus as hyperscale, carrier-neutral and AI-ready, with industry reporting describing the launch of East Africa’s first large hyperscale AI-ready facility in Nairobi. These assets do not yet make Kenya a frontier AI-training cluster, but they do support a more realistic role in colocation, regional cloud, enterprise hosting, sovereign cloud, inference workloads and edge AI.
Fifth, Kenya has policy momentum. The National AI Strategy 2025–2030 is important because it shows that Kenya is not treating AI only as a foreign-investment opportunity. The strategy frames AI around digital infrastructure, data ecosystems, research and innovation, governance, and socio-economic development. This is the right direction: countries capture AI value not only by hosting compute, but by building institutional capacity, local datasets, sectoral applications, standards, skills, and public-sector adoption.
Finally, Kenya has a strong innovation and application layer. Its comparative advantage may be less about competing with the Gulf, the United States or China for massive AI-training campuses, and more about becoming East Africa’s sustainable AI adoption and inference hub. That means local-language AI, fintech infrastructure, agricultural intelligence, logistics optimisation, public digital infrastructure, healthcare applications, SME cloud adoption, cybersecurity, and regional sovereign-cloud services. Microsoft’s original announcement with G42 explicitly included local-language AI model development, skills, connectivity, cybersecurity, and a new Azure East Africa cloud region — a reminder that the project was never only about a data centre.
So Kenya’s opportunity remains substantial, but it should be framed correctly. Kenya may not yet be the natural home for the most electricity-intensive AI training clusters. But it can still be one of Africa’s strongest candidates for green regional cloud, AI inference, sovereign data infrastructure, applied AI, public-sector AI capability, and East African digital services integration. The strategic lesson is not that Kenya lacks an AI future. It is that Kenya’s strongest AI future may sit in the middle layers of the AI value chain — where connectivity, renewable power, data governance, regional demand, and applied innovation combine.
Implications for governments
Governments need national AI value-chain audits.
Before offering incentives to hyperscalers, they should evaluate power availability, transmission capacity, data governance, talent, water constraints, cybersecurity, cloud demand, fiscal exposure, and domestic spillovers.
They should ask:
Can this investment strengthen local firms?
Will it improve public-sector capability?
Does it deepen energy resilience or crowd out other users?
Will it create data sovereignty risks?
Can the country capture value beyond land, power sales, and publicity?
AI policy must become industrial policy, energy policy, education policy, and infrastructure policy at the same time.
Implications for investors
Digital transformation financing can no longer focus only on broadband, startups, and e-government. The next generation of digital infrastructure includes and requires power-sector investment, grid modernisation, transmission, renewable baseload, battery storage, and institutional capacity.
AI readiness is now partly a power-system readiness question.
What’s next
Kenya’s reported data centre setback should not be interpreted as failure. It should be interpreted as diagnosis.
The country discovered that its AI infrastructure ambition is ahead of its current power-system capacity. That is valuable information.
The deeper lesson applies globally: countries must stop thinking about AI participation as a prestige contest. They must identify where they can realistically contribute to the AI value chain: compute, data, models, applications, infrastructure, governance, or adoption.
The winners will not simply be the countries that announce the biggest AI projects.
They will be the countries that align energy, infrastructure, talent, regulation, capital, and national development strategy into a coherent position in the AI economy.
SOURCES & FURTHER READING
Microsoft and G42 announce $1 billion comprehensive digital ecosystem initiative for Kenya
Microsoft, G42 to invest $1 billion in Kenya data centre
Energy shortfall problem scuppers Kenya’s $1B Microsoft data center
Kenya suspends $1 billion Microsoft data centre as energy shortfall raises concerns
Microsoft and G42 Announce $1B Geothermal Data Center in Kenya
Biannual Energy & Petroleum Statistics Report 2025/2026 — Energy and Petroleum Regulatory Authority, Kenya
https://www.epra.go.ke/sites/default/files/2026-03/Biannual%20Statistics%20Report%202025-2026_1.pdf
Key questions on energy and AI — International Energy Agency
https://www.iea.org/reports/key-questions-on-energy-and-ai/executive-summary
Country Case: Kenya taps the Earth’s heat — International Monetary Fund
https://www.imf.org/en/publications/fandd/issues/2022/12/country-case-kenya-taps-the-earth-heat
Comments