Microsoft and G42’s ambitious plan to build a $1 billion-plus Azure data center in Kenya has hit a roadblock after the companies asked the Kenyan government to commit to guaranteed annual cloud-capacity payments—a request that has complicated what was meant to be a definitive May agreement.

Multiple sources confirmed to Windows News that negotiations have slowed significantly over the issue, threatening a flagship African sovereign cloud project that was announced with fanfare in early 2025. The demand for capacity guarantees, coupled with unresolved power supply assurances, has forced both sides back to the drawing board as Nairobi scrutinizes the long-term fiscal implications.

A $1 Billion Bet on African Cloud

The Kenya data center is a cornerstone of Microsoft’s partnership with G42, the Abu Dhabi-based artificial intelligence and cloud computing firm. Together, the two companies have committed to investing billions in digital infrastructure across the Global South, with Africa as a primary theater. The Kenyan facility was slated to be one of the continent’s most advanced hyperscale data centers, designed to deliver Azure services—including AI workloads—with the sovereignty, latency, and data residency that local governments and enterprises demand.

Microsoft and G42 outlined the project as part of a broader “Digital Kenya” initiative, promising not just cloud capacity but also skilling programs and a local ecosystem of partners. The data center would sit in a special economic zone and be powered, at least partially, by renewable energy—a key requirement for both Kenya’s green ambitions and the tech giants’ sustainability targets.

The Capacity Guarantee Impasse

At the heart of the delay is a relatively standard but politically sensitive ask: Microsoft and G42 want the Kenyan government to underwrite a baseline level of cloud consumption over several years. In practice, this means Nairobi would pledge to purchase—or ensure the purchase of—a set amount of cloud capacity annually, irrespective of whether government agencies or local businesses actually consume the full allocation.

For Microsoft, such commercial models are not unprecedented. Large-scale sovereign clouds, like those delivered to governments in Europe and Asia, often involve upfront commitments that give the provider financial certainty to justify the massive capital expenditure. But for Kenya, an emerging market with competing budget priorities, guaranteeing tens or even hundreds of millions of dollars annually for cloud services represents a significant fiscal risk. Public debt already stands at over 70% of GDP, and the government is under IMF pressure to rein in spending.

“The government is not against the data center,” a person familiar with the talks said. “But agreeing to capacity guarantees effectively shifts demand risk from the vendor to the taxpayer. That’s a hard sell domestically, especially when core services are still underfunded.”

The May timeline—likely tied to a memorandum of understanding that was due to convert into a binding agreement—has now slipped. Negotiators are scrambling to find a middle ground, perhaps a phased approach where commitments ramp up gradually, or a co-investment model that brings in development finance from institutions like the World Bank’s IFC or the African Development Bank.

Power Worries Add Another Layer

Alongside the commercial dispute, reliable electricity supply remains a major concern. Kenya’s grid, while among Africa’s most developed, still suffers from frequent outages and high transmission losses. A hyperscale data center in Olkaria or a similar site would draw huge, uninterrupted power—ideally from geothermal sources—but dedicated transmission lines and backup generation require yet more investment. The companies have sought sovereign guarantees that power will be available at a predictable tariff over the life of the facility, adding another dimension to the government’s liability exposure.

“Power guarantees are a red line for lenders,” an infrastructure analyst told Windows News. “No one finances a billion-dollar data center without firm assurance that the lights will stay on. If the government can’t provide that, the risk premium goes through the roof, and the whole business case unravels.”

Microsoft has been unwilling to comment publicly on the state of negotiations, while G42 referred queries to its partner. The Kenyan Ministry of Information, Communications and the Digital Economy did not immediately respond to a request for comment. But officials have previously emphasized the transformative potential of the project and the need to balance enthusiasm with prudence.

Sovereign Cloud Ambition vs. Fiscal Realities

The impasse underscores the tension inherent in sovereign cloud deals. Governments see them as pathways to leapfrog into the AI age, modernize public services, and stimulate local tech sectors. Providers like Microsoft see them as beachheads for market share and long-term revenue in fast-growing regions. Yet the mismatch between political horizons—where leaders want ribbon-cutting ceremonies within an election cycle—and infrastructure horizons, where hypercalers think in decades, is increasingly stark.

For Kenya, a sovereign cloud promises to keep sensitive data onshore, satisfying legal and security requirements that may otherwise block cloud adoption in government, banking, and telecoms. The data center would also create high-tech jobs and, supporters argue, attract fintech and agritech startups that need world-class infrastructure. President William Ruto’s administration has heavily promoted the deal as an emblem of a new, investment-friendly Kenya.

But guaranteeing cloud demand is not a trivial matter. Government agencies remain at very different stages of digitization. Many still run legacy systems that are not ready for the cloud. Forcing them to migrate quickly—and to pay for reserved capacity they may not fully use—could waste money and invite political backlash. Moreover, Kenyan enterprises and startups, while growing, cannot be compelled to use Azure; market forces and competition from AWS, Google Cloud, or nimble local providers will shape actual consumption.

What Could Break the Deadlock?

Several compromise scenarios are on the table, according to insiders. One is a “government anchor tenant” model where Nairobi agrees to migrate a critical mass of ministry workloads to the new Azure region within a set timeframe, without hard financial penalties for underconsumption. Microsoft and G42 might then use that commitment to raise project finance on their own balance sheets, reducing reliance on sovereign backing.

Another idea is to involve the African Union or regional bodies, turning the facility into a multi-country resource. If neighboring nations—like Uganda, Tanzania, or Rwanda—also pledge a minimum off-take, the aggregated demand could reach a level that satisfies both the vendors and the host government without overly burdening any single treasury.

A third path involves development finance. The World Bank’s private sector arm, the IFC, has an existing partnership with G42 on digital infrastructure and might be willing to provide partial risk guarantees that unlock commercial lending. This could ease the pressure on Nairobi for direct fiscal guarantees.

Still, none of these mechanisms can entirely sidestep the core issue: Microsoft and G42 want assurance that when the data hall lights come on, there will be customers paying for the power and capacity. In the absence of a mature regional cloud market, that assurance, for now, must come from governments.

Geopolitical Overtones

The delay also unfolds against a complex geopolitical backdrop. G42’s deep ties to the UAE and its role as a vehicle for Gulf capital in African tech have drawn scrutiny from Western regulators, especially in the United States. Microsoft itself has been navigating US export controls and semiconductor restrictions related to AI infrastructure in countries with Chinese influence. Kenya, while broadly aligned with the West, has also deepened relations with Beijing through Belt and Road investments. The data center, if built, would anchor a US-UAE digital alliance in East Africa at a time when China’s Huawei and ZTE are vying for telecom and cloud contracts across the continent.

Some analysts speculate that Washington’s quiet encouragement of the Microsoft-G42 partnership is partly about countering Chinese tech expansion. A faltering project could therefore dent US strategic interests, potentially prompting diplomatic intervention. So far, however, the Biden administration has been reluctant to directly fund or guarantee foreign commercial cloud ventures, leaving the parties to sort out their own financing.

Implications for Africa’s Cloud Race

For the broader African cloud market, the Kenya impasse sends a cautionary signal. Hyperscale data centers are capital-intensive monsters that demand sustained utilization. In regions where enterprise IT spending is still nascent, the economic viability of multiple regional Azure, AWS, or Google Cloud zones often depends on government anchor tenancy. If Kenya—which has one of the more diversified and tech-forward economies in sub-Saharan Africa—cannot make the numbers work, it may dampen enthusiasm for similar projects in Nigeria, Ghana, or Senegal.

On the other hand, some African governments may learn from the negotiations and structure more creative public-private partnerships. The Kenyan government’s insistence on sound fiscal vetting, while frustrating to proponents, could set a precedent for greater scrutiny. In the long term, that might lead to more sustainable cloud investments, even if the short-term headlines look like delay.

Local business leaders have expressed frustration. The Kenya Private Sector Alliance had hoped the data center would accelerate digital transformation among members, lowering latency for banking applications, mobile money, and AI-driven services like crop insurance. “Every month of delay is a missed opportunity,” one tech executive said. “But we also understand that the government can’t just sign a blank cheque.”

What Happens Next

Both sides have incentives to rescue the deal. Microsoft is under pressure to show progress on its “Digital Kenya” pledges and to expand Azure’s footprint in a region where AWS is opening a South Africa region and Google is eyeing Lagos. G42’s expansion into Africa depends on demonstrating replicable models with anchor clients. Kenya, meanwhile, wants the jobs, the FDI, and the symbolic victory of hosting one of Africa’s first sovereign Azure regions.

Negotiators are expected to continue intensive talks through the summer, with a revised agreement possible by the end of the third quarter. In an ideal scenario, the government could announce a “phased cloud adoption” plan in its next budget, incorporating modest but credible commitments that reassure Microsoft and G42 without triggering a parliamentary standoff over fiscal guarantees.

But there is no time to waste. Construction delays in hyperscale projects compound quickly—equipment lead times, land acquisition, and regulatory approvals all have expiration windows. If the deadlock persists into 2026, the companies might be forced to re-evaluate gate sizes, capex envelopes, or even site alternatives in friendlier jurisdictions.

The world will be watching. Africa’s cloud potential is real, but the gap between ambition and on-the-ground execution remains wide. The Kenya data center saga is a vivid reminder that even when the technology is ready and the capital is committed, the economics and politics of cloud can be as complex as any piece of silicon.