The cloud may seem light, but it lives in protected buildings that are heavy, expensive, and dependent on a brutal, constant energy load.

The Google Cloud region in Johannesburg, opening in 2024, and AWS's presence in Cape Town show that Africa has entered the physical map of major digital platforms, with services closer to banks, governments, and businesses, without, by itself, resolving technological dependence.

This approach has concrete advantages. The shorter distance between servers and users can reduce delays, improve digital services, and facilitate the adoption of artificial intelligence tools.

For African companies, the cloud can lower initial costs and provide access to technical capabilities previously reserved for large international groups with their own teams and larger budgets.

But infrastructure is never neutral. Whoever owns the data centers sets the prices, conditions, technical standards, and access methods. When a state or company depends on platforms it doesn't control, its decision-making margin diminishes, and digital sovereignty becomes dependent on contracts written far from the continent, even when national law remains valid.

The risk is that Africa will become merely a consumer of artificial intelligence sold in pre-packaged goods. The continent's economic history knows this scenario well: exporting the raw materials, importing the finished product, and paying dearly for the embedded knowledge. In the digital economy, the raw material is the data produced by millions of citizens every day.

Therefore, the arrival of the cloud must be approached with caution. Foreign investment can be beneficial, but it requires transparent contracts, data protection rules, local training obligations, and space for African suppliers. Without these, the continent will gain technical speed but lose political power over the machine that accelerates it.

Artificial intelligence learns from data. It learns from mobile payments, medical exams, agricultural images, public records, urban maps, local languages, and consumer behavior.

Each record can feed into models that then sell predictions, automate decisions, and reorganize markets. Data has ceased to be an administrative waste product and has become a political asset, both within and beyond African national borders.

The dispute begins when data collection is faster than the law. Many African countries still have weak institutions to oversee platforms, protect citizens, and compel companies to explain how they use information. Without strong regulations, innovation may advance first, and accountability may come too late, without clear redress.

The African Union's Continental Artificial Intelligence Strategy is important because it recognizes that no country can negotiate effectively alone with the tech giants. The continental scale allows for discussions on common rules, ethical standards, data protection, scientific training, and cooperation among states, with the political capacity to implement decisions.

There is also a cultural dispute. Models trained far from Africa may fail to read the languages, names, customs, religious practices, and local forms of authority. When the machine does not understand the people it analyzes, the technical decision may reproduce error, exclusion, and injustice under the guise of neutrality.

African data should not only feed foreign companies. It should serve African universities, hospitals, farmers, schools, courts, and small businesses.

The essential question is simple: if African digital life generates value, does some of that value go back into improving the concrete lives of the people who produced it, with clear public consent and recognizable benefits in the communities?

Artificial intelligence consumes electricity, water, cooling, and security. A data center doesn't just exist in the clean vocabulary of innovation. It exists in a concrete electrical grid, often strained by power outages, high tariffs, and communities that still live without a regular supply.

The cloud always depends on a physical territory, with often hidden social costs. The investment announced by Microsoft and G42 in Kenya exposed this contradiction. The project promised a data center linked to geothermal energy in Olkaria and a new cloud region for East Africa.

Then the question arose about the electrical capacity needed to sustain the ambition without weakening the public grid and the local economy. The issue is not against investment. It is against the illusion that artificial intelligence arrives without material costs.

Servers require continuous electricity, cooling, land, cabling, and security. When these resources are scarce, every decision about the cloud also becomes a decision about the country's real economy and the tariffs paid by citizens.

Africa has significant energy advantages. The sun, wind, rivers, and geothermal energy can support a portion of the digital infrastructure. But this advantage only translates into sovereignty when contracts ensure public benefits, network expansion, technical training, local access, and clear environmental costs, not just energy for foreign machines.

If clean energy powers data centers first and communities only later, modernization will be unfair. Artificial intelligence needs light, but so does the population. The question about cloud energy is, fundamentally, a question about who has priority when the future arrives plugged into the wall socket, in the neighborhood, at school, and in the hospital.

Technological sovereignty doesn't depend solely on states. It also depends on African companies capable of building data centers, developing software, training engineers, and selling solutions tailored to the continent.

Without its own business fabric, Africa will continue to buy artificial intelligence like someone buying a closed box, without knowing what's inside or who controls the key components.

There are signs of change. African telecommunications, fiber optic, and data center companies are seeking to occupy space in a chain dominated by foreign capital. The objective should not be to reject partnerships, but to enter them with their own capabilities. Those who do not control the infrastructure always negotiate from a weak position, even when the contract promises balance.

Knowledge is the deciding factor. Chips, models, servers, and platforms require engineers, universities, research, mathematics, linguistics, law, and data science.

Artificial intelligence doesn't just come from expensive machines. It comes from trained people, funded laboratories, and institutions that transform talent into national capacity, with public continuity and African scientific memory. Small African businesses cannot be left out of this race. Technology will be of little use if it only serves banks, ministries, and multinationals.

The real leap forward will happen when farmers, clinics, schools, transporters, cooperatives, and breeders are able to use accessible, auditable tools designed for concrete African problems in the countryside, cities, and neighborhoods.

The race for artificial intelligence is also a race for youth. The continent has human energy, creativity, and a need for skilled jobs. If this youth only uses imported applications, the promise will be short-lived.

If technology is the author of the models, platforms, and companies, it can become an instrument of sovereignty if training is included in public budgets.

Digital sovereignty begins with laws, but it doesn't end with legislation. A country can approve a good data protection standard and still lack the power to enforce it. Enforcement requires independent authorities, well-prepared courts, capable technicians, and sanctions that don't exist merely on paper against companies larger than many states and with the means to investigate.

Therefore, the continental strategy needs to move beyond conferences and into public contracts, government procurement, university curricula, and agreements with platforms.

AI in Africa will not be sovereign if each country negotiates in isolation, without common standards, without knowledge transfer clauses, and without effective protection of citizens, with public monitoring and accountability. Data protection cannot be presented as an obstacle to development.

On the contrary, it is a condition for innovation to gain public trust. When citizens don't know who collects their information, where it is stored, and what it will be used for, technology becomes an opaque box in the eyes of society, even when it promises faster services.

African states also need to defend local languages ​​and national archives. Artificial intelligence that ignores the names, accents, memories, community boundaries, and ways of life of the continent can automate long-standing exclusions with a modern veneer. Technology without context turns difference into error, especially in automated public services.

Regulation, when done seriously, does not stifle innovation. It compels it to respond to the public interest. AI in Africa needs rules that protect citizens, strengthen local businesses, value African research, and prevent digital modernization from repeating, under a different name, the continent's old extractive logic in finance, health, and education.

Artificial intelligence can help Africa produce better, diagnose earlier, educate more broadly, and organize public services more efficiently. But none of these promises eliminates the central question of AI in Africa: who controls the data that feeds the machine, and who receives the value it creates?

The continent doesn't need to reject the cloud, data centers, or foreign investment. It needs to negotiate firmly, legislate intelligently, and build its own capacity. Technological sovereignty is born when states protect their citizens, African companies gain ground, and universities cease to be mere consumers of foreign technology.

If Africa enters this race merely as a market, it will once again lose its central role in decision-making. If it enters as the owner of data, energy, knowledge, and rules, artificial intelligence can cease to be a dependency and become a source of power.