In a significant gathering held in Abuja from 21st to 23rd October 2026, the Nigeria AI Collective, operating under the Federal Ministry of Communications, Innovation and Digital Economy, hosted the BUILD Conference. The event, themed "Building Africa’s Shared Compute and AI Infrastructure," brought together policymakers, students, researchers, journalists, innovators, and investors. Discussions at the conference centred on how Africa can develop, share, and govern the computing power essential for its AI future, whilst also exploring strategies to dismantle the shackles of digital imperialism.
The poignant warning from Kwame Nkrumah in his 1965 publication, "Neo-Colonialism: The Last Stage of Imperialism," resonates powerfully today. Even as newly independent African nations celebrated their sovereignty, Nkrumah cautioned that empire had merely adopted a new guise, changing its accounting methods. Were Nkrumah alive in the present era, he would undoubtedly recognise these same patterns manifesting within the digital realm, now largely residing "in the cloud."
The New Face of Imperialism: Digital Control
Traditional imperialism employed brutal directness: siphoning local resources to metropolitan centres, suppressing indigenous industries, locking colonies into inequitable trade agreements, and arbitrarily redrawing borders, much like the Berlin Conference of 1884 with its rulers and maps. Today, this process is perpetuated through terms of service, cloud subscriptions, and the control of training data.
Consider the foundational infrastructure that underpins the contemporary world. Industry estimates reveal that three American corporations – Amazon, Microsoft, and Google – collectively command over 65 per cent of global cloud capacity. In stark contrast, Africa, home to 18 per cent of the world’s population, controls less than one per cent. This disparity starkly illustrates who truly owns the fundamental infrastructure of the digital economy and who is merely a tenant.
Across the continent, news organisations are increasingly reliant on global tech platforms, including social media and search engines, for content distribution. This dependence renders them susceptible to external algorithms and rule-setting frameworks.
This evolving manifestation of neo-imperialism is not abstract; it represents an extractive business model that imperils economic self-determination and is actively undermining one of Africa’s most vital democratic pillars: its independent press.
Three Waves of Digital Assault
The assault on Africa's digital autonomy has unfolded in three distinct waves:
- Distribution Capture: The initial wave saw foreign tech conglomerates monopolising digital advertising. According to the Reuters Institute, these entities captured over 80 per cent of local advertising expenditure in key African markets. Newsrooms, which had historically weathered military decrees and fuel shortages, suddenly found their audiences dictated by algorithms originating in Silicon Valley. When these platforms subsequently deprioritised news content to maximise screen time, the economic foundation of many news outlets crumbled. The CEO of the Newspaper Proprietors’ Association of Nigeria recounted instances of publishers reducing their newsroom staff by 40 per cent in two years, not due to a decline in readership, but because the platforms controlling audience access shifted their priorities. The audience persists, but the viable business model does not.
- Data Colonialism (Generative AI): The second wave has been swifter and more extractive, transforming distribution capture into data colonialism. Decades of African journalism – including investigations conducted under challenging military regimes, during contentious elections, and amidst public health crises, often at considerable personal risk – are being exploited without licence, credit, or compensation. This data is being used to train large language models in places like California. These models then return African work in a form that undermines the original, creating "zero-click summaries." This begs the question: why would a user visit a Nigerian newspaper's website when a chatbot can condense its exclusive investigation into a single paragraph, offering no traffic, no attribution, and no revenue to the newsroom that funded its creation? Compounding this issue, African content constitutes less than two per cent of global AI training data, as indicated by several AI governance audits. Consequently, these summaries frequently default to crude Western stereotypes. When AI fails to comprehend Africa, it fabricates it, resulting in an erasure through statistics.
- Stealth AI: The third wave presents a particularly alarming threat. Termed "Stealth AI," this phenomenon transcends mere chatbot hallucinations. It represents an industrial-scale production of synthetic deceptions: artificial reporters, deepfake audio of political candidates released on election eve, and autonomous bot swarms generating artificial consensus on platforms like WhatsApp and Facebook. Threat researchers recently exposed international firms that entirely fabricated a French-Ghanaian investigative journalist, complete with name, byline, and image, to orchestrate coordinated smear campaigns in West African media. Over 60 per cent of African journalists now report difficulty in distinguishing genuine sources from synthetic fakes, according to surveys by the Centre for Journalism Innovation and Research. This deliberate confusion serves a clear purpose. It is noteworthy that the Federal Competition and Consumer Protection Commission (FCCPC) has initiated formal inquiries into these practices, yet the data scraping continues daily.
Addressing Economic Surrender and Paving the Way for Dominance
The rise of Stealth AI is fundamentally enabled by a deeper economic surrender. Africa, with its 1.5 billion inhabitants, 60 per cent of whom are under 25, speaking over 2,000 languages, generates billions of valuable data points daily – from mobile money transactions and agricultural telemetry to Nollywood scripts and voice notes in Amharic. This is the raw material essential for artificial intelligence. Yet, where is this data processed? Northern Virginia. Where is it stored? Dublin and Singapore. Who ultimately owns and monetises the resulting models? Companies headquartered in San Francisco, Seattle, and Hangzhou.
Africa exports its raw data freely across platforms it does not own and through subsea cables it does not control. This raw material is then imported back as an expensive finished product: a chatbot that struggles with Hausa, a credit scoring model oblivious to informal markets, or an educational tool unversed in the literary works of Achebe.
Therefore, the discourse in Abuja must evolve from mere declaration to a pursuit of digital dominance. "Compute" – encompassing reliable energy, data centre capacity, fibre-optic networks, and specialised processors – is the equivalent of the railways and seaports of this century. True digital sovereignty is unattainable without energy sovereignty. Whosoever controls compute and power will ultimately control commerce.
Practical Steps Towards Digital Sovereignty:
- Fair Compensation Framework: Following the examples set by Australia and Canada, which compelled tech platforms to negotiate with publishers, Africa urgently requires a continental AI and data bargaining code. If an AI model has been trained using African archives, it must be licensed and paid for. When a news organisation invests resources to create content, it is entitled to a fair share of revenue generated from its consumption. This is not merely a commercial consideration; the press is an indispensable national and civic infrastructure. As Thomas Jefferson famously stated, "Were it left to me to decide whether we should have a government without newspapers or newspapers without a government, I should not hesitate a moment to prefer the latter." This vital national infrastructure must be zealously protected and defended at all times.
- Sovereign Storage Mandate: A mandate is needed for all public data, government archives, and data generated with public funds to be stored on cloud infrastructure physically located within Africa and subject to African legal frameworks. True sovereignty necessitates physical location.
- Concrete Compute Target: The African Union, leveraging the African Continental Free Trade Area, should establish and fund an ambitious goal: to achieve five per cent of global data centre capacity on African soil by 2030, powered by sustainable and reliable energy sources. While ambitious, this is not an impossible feat, given that the continent already attracts over $2 billion annually in data centre investment. What is now required is policy intervention to guide and scale this investment effectively.
The fundamental question facing Africa is whether it will choose to be a tenant or a landlord in the unfolding digital era.
Adediran is the Chief Executive Officer of the Newspaper Proprietors’ Association of Nigeria (NPAN).