Microsoft’s plan to build a new Azure cloud region in Kenya, announced with much fanfare in May 2024 as part of a $1 billion digital investment package, has reportedly stalled over disagreements on payment commitments and power supply guarantees. The delay throws the future of East Africa’s first hyperscale cloud region into uncertainty, leaving businesses and developers waiting for local Microsoft cloud services.

The announcement, made jointly with Abu Dhabi-based artificial intelligence firm G42, was hailed as a transformative moment for Kenya’s digital economy. It included a pledge to build a state-of-the-art data center, deliver AI and cloud skilling programs, and spur innovation across the region. At the heart of the package was the Azure region, which would bring Microsoft’s full cloud stack—compute, storage, networking, AI, and machine learning—closer to East African users.

Now, sources familiar with the negotiations say that talks between Microsoft and the Kenyan government have stalled over two critical issues: who guarantees the electricity supply required to run the facility, and how the multi-year cloud service contracts will be financially secured. Neither party has officially commented on the impasse.

The Power Struggle

Data centers are voracious consumers of electricity. A single hyperscale facility can require tens of megawatts of stable, uninterrupted power—often backed by on-site generation and redundant grid connections. Microsoft’s standard design for a new Azure region assumes a minimum of 30 MW of reliable power, with expansion potential up to 100 MW or more to accommodate AI workloads, which are especially energy-intensive.

Kenya’s national grid, while the most robust in East Africa, faces challenges. The country relies heavily on a mix of geothermal, hydro, and thermal generation, but peak demand already strains capacity. Meeting the 99.99% uptime demands of a Tier IV facility would require not just grid upgrades but also dedicated industrial power lines and on-site backup generation. Who foots the bill for that infrastructure is a central point of contention.

Microsoft typically seeks long-term power purchase agreements (PPAs) with clear government guarantees that the grid will deliver. In previous deals, such as in Ireland and Sweden, Microsoft has partnered directly with energy developers and grid operators to build out renewable energy capacity. In Kenya, however, the government is wary of shouldering the upfront infrastructure costs, especially when the economic benefits of the data center—job creation, tax revenue, and market stimulation—are delayed and uncertain.

An insider familiar with the negotiations indicated that the Kenyan government wants Microsoft to provide guarantees that the power consumption will actually materialize at the scale promised, effectively insuring the state against building underutilized grid connections. Microsoft, on the other hand, needs assurance that the power will be there before it breaks ground on a $500–700 million facility.

Payment Commitments: Who Takes the Risk?

The second sticking point revolves around cloud service consumption. Governments in emerging markets often negotiate anchor tenancy agreements—promising to be a major customer for a set period—to de-risk the massive capital outlay hyperscalers must make. In Kenya’s case, the government was expected to guarantee a certain volume of cloud consumption, potentially through a sovereign-backed contract, to make the project viable.

According to sources, negotiations have included demands for advance payment commitments or sovereign guarantees that the state is reluctant to provide. Kenya’s fiscal position, strained by debt repayments and currency volatility, makes long-term financial commitments to a foreign tech company politically and economically contentious. The government is reportedly pushing for a model where cloud services are paid for only as they are consumed by individual ministries and agencies, without a blanket guarantee.

For Microsoft, a region without an anchor tenant is a risky bet. While it can sell to private enterprises and startups, the public sector is often the largest spender on cloud in African markets, and a government that won’t commit is a warning sign. The company has walked away from projects before when the financial case didn’t close—in 2019, it paused plans for a cloud region in Ohio, though it later resumed after revised incentives.

The East African Cloud Vacuum

The delay is a setback not just for Kenya but for Microsoft’s broader African cloud strategy. The continent remains a battleground for hyperscalers. Amazon Web Services opened a Cape Town region in 2020, and Google Cloud launched in Johannesburg in 2023. Microsoft, through its Azure network, operates in South Africa (North and West) and is expanding to Nigeria, but East Africa has been a glaring omission.

A Kenyan region would serve a vast and fast-growing market that includes Ethiopia, Tanzania, Uganda, Rwanda, the Democratic Republic of Congo, and even parts of the Indian Ocean islands. For local businesses, the absence of a local Azure region means continued reliance on data centers in Europe or South Africa, bringing latency of 100–200 milliseconds—unacceptable for real-time AI inference, financial trading, or cloud gaming. Data sovereignty concerns also persist: sensitive government and financial data must often stay in-country by law, forcing organizations to use on-premises hardware or smaller, non-hyperscale local providers.

G42: The AI Wildcard

The involvement of G42 adds another layer of complexity and potential. The UAE-based company has deep ties to Middle Eastern sovereign wealth and a focus on artificial intelligence infrastructure. In April 2024, Microsoft invested $1.5 billion in G42 as part of a broader alliance to accelerate AI capabilities globally. The Kenya project was to be a flagship for this partnership outside the Gulf.

G42 brings expertise in building energy-efficient AI training clusters—exactly the kind of infrastructure that could turn Kenya into an AI hub for the continent. The company has already deployed large-scale NVIDIA GPU clusters in the UAE and has begun exporting that model. A person familiar with the talks noted that G42’s participation was meant to future-proof the Kenyan region for AI workloads, from training large language models to running inference for mobile-first applications. Stalled talks now jeopardize that vision.

Microsoft and G42 had planned to build not just a standard Azure region but one with a dedicated AI infrastructure, potentially including a zone rich in GPUs and AI accelerators. That would have been a first for Africa, leapfrogging the continent from basic cloud to cutting-edge AI services. Without the power and payment guarantees, however, the economics of such a capital-intensive buildout fall apart.

A Pattern of Hyperscale Hurdles

The Kenya standoff echoes challenges seen in other emerging markets. In India, Microsoft took years to negotiate power and land for its Hyderabad region. In Indonesia, the government had to provide special economic zone status and tax incentives to attract a Google Cloud region. Africa adds its own layers: currency risk, political uncertainty, and infrastructure deficits.

Yet the prize is enormous. The African cloud services market is projected to grow at over 20% annually, driven by mobile-first populations, fintech, and e-government initiatives. A local Azure region could reduce latency to single-digit milliseconds for millions of users, enable advanced AI services, and keep data sovereign. The delay threatens to cede this market to competitors or to homegrown alternatives like Liquid Intelligent Technologies, which operates data centers in Nairobi and already offers Azure Stack edge solutions.

Economic and Strategic Fallout

Kenya has ambitions to become a digital leader, with a vibrant startup scene dubbed “Silicon Savannah.” An Azure region was seen as the infrastructure foundation for that goal. The $1 billion package included AI innovation labs, digital skills training for millions of Kenyans, and support for startups. If the Azure region is delayed, the entire package could unravel, as the data center was the linchpin.

Some components—like training programs—might proceed independently, but the transformative effect would be diluted. Without a local cloud, Kenya risks falling behind Nigeria and South Africa in the race for hyperscale investment. That would have cascading effects on job creation, GDP growth, and the ability to attract global tech companies that rely on local cloud infrastructure.

What Happens Next?

Behind the scenes, negotiations are ongoing. Both parties have too much at stake to walk away entirely. There is talk of a revised timeline, perhaps pushing the region’s launch to 2026 or beyond, as the government and Microsoft work through the guarantee frameworks. Mediation may come from development finance institutions like the World Bank’s IFC, which has experience in de-risking digital infrastructure projects in emerging markets.

One possible outcome is a phased approach: an initial smaller data center with limited power and expandable capacity, combined with a pay-as-you-go government commitment model. Another is for Microsoft to shift focus to renewable microgrids, supplying its own power via solar and battery storage, reducing dependence on the grid—though that raises capital costs even further.

For now, the cloud over Kenya’s cloud is one of uncertainty. The coming months will test the commitment of both Microsoft and the Kenyan government to bridge the gap between ambition and execution. East Africa’s digital future hangs in the balance.