Compute Is Sovereignty

Why the nations that cede their digital stack will cede their economic future

In 2024, Microsoft and the Abu Dhabi-based G42 announced a $1.5 billion partnership to expand AI and cloud infrastructure across Africa. The deal was structured as a development investment: G42, backed by UAE sovereign capital, would deploy compute infrastructure in key African markets, partnering with Microsoft’s Azure to provide cloud services and AI tooling to governments and enterprises across the continent. Kenyan officials welcomed the announcement. So did several others across East Africa.

The concern that follows this kind of deal is not hypothetical. Africa accounts for less than one percent of global data center capacity despite being home to 18 percent of the world’s population. The continent’s digital infrastructure — the servers on which government records are stored, the pipes through which financial transactions flow, the compute that will eventually underpin AI-driven public services — is overwhelmingly owned by entities headquartered in the United States, China, and, now, the Gulf. That is not an investment gap. It is a sovereignty gap.

The concept of digital sovereignty is sometimes treated as an abstract policy concern — the kind of thing that occupies white papers and intergovernmental meetings without touching the daily operations of businesses and citizens. That framing is wrong. Digital sovereignty, or the lack of it, is the single most consequential structural variable in determining whether African nations capture value from the technology transition or merely facilitate it for foreign owners.

The Infrastructure Deficit in Numbers

The scale of Africa’s data center gap is only recently becoming widely understood. Africa’s digital connectivity stood at 34 percent in 2024, meaning fewer than four in ten Africans had internet access. By 2025, demand for digital services was projected to grow 18 percent year-on-year, with an estimated 680 million smartphone connections active by year-end. These are demand curves that presuppose infrastructure — and the infrastructure, in the form of owned, operated, and sovereign compute capacity, does not yet exist at the required scale.

African governments spend, on average, seven times more on debt service than on infrastructure. Median debt service grew from 3.5 percent of government expenditure in 2010 to 12 percent in 2024. An annual infrastructure investment of $155 billion would add $2.83 trillion to Africa’s GDP by 2040 — more than doubling the $2.80 trillion recorded in 2024. But that investment requires fiscal space that most African sovereigns do not currently possess, creating a dependency on foreign capital that structurally replicates the conditions of earlier resource extraction cycles.

The subsea cable revolution of the last five years offers an instructive parallel. The landing of Equiano, 2Africa, and PEACE cables along Africa’s coastlines has transformed available bandwidth potential dramatically. Theoretical connectivity capacity has increased by multiples. But the majority of these cable assets are owned by foreign entities — Google, Meta, a consortium of global telecoms — and the critical middle-mile infrastructure that would distribute that bandwidth inland has been slower and more patchwork in its development. Access without ownership replicates a familiar dynamic: Africa provides the market; the infrastructure owners capture the margin.

The Sovereign Cloud Problem

A 2025 assessment found that many African countries lack national strategies or policies for establishing a sovereign cloud — a domestically owned and controlled computational environment for sensitive government data and critical national infrastructure. This absence is not merely a technical gap. It is a governance failure with compounding consequences.

When a government stores citizens’ tax records on AWS servers in Dublin, it has implicitly accepted that the terms of that storage — the security standards, the data retention policies, the conditions under which data can be accessed by third parties — are determined by the infrastructure owner’s home jurisdiction and corporate governance. The European Union’s experience with GDPR compliance across US-headquartered cloud providers illustrates how even the world’s most powerful regulatory bloc struggles to assert sovereignty over foreign-owned infrastructure. African governments, with far less regulatory leverage, are in a structurally weaker position.

Digital identity infrastructure compounds this vulnerability. In 2024, Uganda’s National Digital ID system became a cautionary study in what happens when identity infrastructure is built without adequate consideration for operational resilience and inclusive access. Rigid biometric requirements, technical failures, and insufficient recourse mechanisms denied millions of citizens — predominantly elderly, rural, and women — access to healthcare, education, and social protection services for which a digital ID was a prerequisite. The system was domestically built and operated, which limited some foreign sovereignty risks, but exposed a different failure mode: the false security of infrastructure that exists without adequate operational governance.

Contrast this with Mauritania, where biometric ID coverage exceeds 94 percent, and Benin, where it approaches 98 percent. These are not wealthy countries. What distinguishes them is the quality of institutional commitment and implementation governance, not the scale of external funding. The lesson is that sovereign infrastructure requires not just ownership but operational capacity — the human capital, the institutional design, and the continuity of governance attention necessary to make infrastructure serve citizens rather than merely exist on a spreadsheet.

The African Stack as Strategic Imperative

The concept of an African Stack — an interoperable set of digital identity, payment, and data exchange systems that form the substrate of a continental digital economy — has moved from aspiration to active policy within the last three years. The AfCFTA’s Digital Trade Protocol, adopted in 2024, establishes harmonized rules on e-commerce, cross-border data flows, and online consumer protection across the 54 signatories of the African Continental Free Trade Area. The Smart Africa Alliance, representing the majority of African governments, has committed to a Single Digital Market by 2030, with a broadband target of 51 percent penetration — up from 34 percent in 2019.

These commitments are consequential only if they are backed by infrastructure that African institutions own and control. A digital trade protocol that facilitates cross-border commerce through platforms owned by foreign entities is not an exercise in African digital sovereignty — it is an exercise in African digital consumption. The distinction matters for the same reason it matters in commodities: the entity that owns the extraction infrastructure captures the majority of the value, regardless of where the resources originated.

The analogy is not rhetorical overreach. In 2023, Africa processed over $800 billion in mobile money transactions — more than half the global total, representing a financial infrastructure achievement of genuine global significance. The majority of the switching and settlement infrastructure underlying those transactions is owned by foreign-headquartered firms. M-Pesa, the most celebrated mobile money system in the world, is a Safaricom product — and Safaricom is majority-owned by Vodafone and the Government of Kenya. The government stake is sovereign; the controlling economic interest is not.

Compute Policy as Industrial Policy

What does sovereign digital infrastructure actually require? The debate often collapses into a false binary between full domestic ownership, which most African economies cannot finance, and open foreign investment, which offers capital but extracts control. The more useful frame is policy architecture: what rules govern data residency, what requirements attach to critical infrastructure ownership, what technical standards ensure interoperability rather than lock-in, and what procurement decisions privilege locally operated systems.

Rwanda’s approach is instructive. The country has pursued a deliberate policy of building domestic cloud capacity — Rwanda Cloud sits within the Kigali Innovation City development — while simultaneously engaging foreign providers under terms that require meaningful technology transfer and local capacity building. The result is imperfect and incomplete, but it represents a governance posture that treats compute as a strategic asset rather than a procurement category.

Kenya’s Konza Technopolis project, now more than a decade old and still substantially incomplete, represents the opposite failure mode: the designation of compute infrastructure as a national priority without the sustained governance attention, institutional coordination, and public investment necessary to move from announcement to operation. The lesson is not that public investment in sovereign digital infrastructure cannot work in Africa. It is that it requires the same governance intensity as any other strategic infrastructure program — and that intensity has not always been present.

The AI layer raises the stakes considerably. Nations without domestic compute capacity will, by structural necessity, develop AI capabilities on foreign infrastructure, generating training data that flows to foreign models, and building dependencies on foreign APIs for the digital services their citizens use. As of 2025, Africa accounts for less than one percent of global data center capacity. By 2030, if current trends persist, the continent will be running its governments, its healthcare systems, its financial infrastructure, and its AI applications on hardware owned and operated outside its borders. The strategic consequences of that scenario extend well beyond technology policy.

The Window Is Narrow

Infrastructure investments are characterized by long lead times, high switching costs, and path dependency. The data centers, fiber networks, and power infrastructure that will define Africa’s digital landscape for the next thirty years are being built now. The ownership structures that will determine who captures the value from those assets are being negotiated in the current cycle of investment. The policy frameworks that will govern data residency, infrastructure access, and technology transfer are being drafted — or neglected — in the present legislative environment.

African governments and development institutions that treat digital infrastructure as a passive output of foreign direct investment are making a thirty-year mistake. Compute is not a utility to be procured. It is the foundational layer of economic and political sovereignty in the digital era. Nations that own their stack will own their future. Those that do not will find, as previous generations found with physical infrastructure, that the terms of dependency are set by the infrastructure owner — and renegotiated only at considerable cost.