Bold ambitions, fragile foundations: Is Ghana ready for the AI and Data Centre risks it is inviting?

Ghana is investing heavily in artificial intelligence (AI) and data centre infrastructure. In 2026, President John Dramani Mahama announced a $250 million investment for a world-class AI Computing Centre and $20 million for the National AI Strategy. International data centre operators, including Digital Realty and Equinix, are setting up shop in Accra.

MTN Ghana has pledged $1 billion over three years to bolster digital infrastructure, including data centres and cloud computing. The goal is to position Ghana as West Africa’s digital gateway and a continental leader in responsible AI.

However, ambition and preparedness are not synonymous. As AI workloads become more compute-intensive, they will put pressure on already fragile power grids, require robust cybersecurity measures, and test regulatory frameworks still in development. The question is whether Ghana has constructed the necessary foundations to handle the risks associated with this technological evolution.

Ghana's regulatory landscape on AI is relatively advanced for the region. The National AI Strategy 2023–2033, approved in February 2026, covers eight pillars: AI education, digital infrastructure, data governance, ecosystem coordination, sectoral AI adoption, applied research, ethical safeguards, and public-sector use. A Responsible AI (RAI) Office will oversee implementation and monitor progress quarterly.

An Emerging Technologies Bill is also being drafted to create binding regulations around AI, blockchain, and other frontier technologies. The Data Protection Act, Cybersecurity Act, and Electronic Transactions Act provide a legal foundation, but none were designed for generative AI, autonomous decision-making, or massive data centres.

Ghana currently has five major data centre facilities near the capital. Digital Realty, Africa Data Centres, PAIX Data Centres, and Equinix are all expanding their presence. The country's strategic advantages include political stability, prime location for subsea cables, and incentives like VAT and duty exemptions for data centre equipment.

However, the reality of power supply is less attractive. Industrial electricity rates are between $0.15 and $0.22 per kilowatt-hour, and the grid is not designed for the high-density, dynamic demand of AI workloads. A single AI training run can temporarily increase power consumption by ten times the baseline. Without reliable power, operating costs will rise, and the case for hosting sensitive AI workloads in Ghana will weaken.

The risks Ghana must address include power and energy, water consumption, brain drain, and data sovereignty. Power is the most critical infrastructure bottleneck. Cooling requirements for data centres are substantial, consuming up to 25.5 million litres of water annually per megawatt. In a nation already grappling with water access issues, expanding data centre capacity without promoting water-efficient cooling technologies would be irresponsible.

Additionally, there is a risk of brain drain as Ghana's most talented technology professionals may leave in pursuit of better opportunities abroad. Finally, data sovereignty is a significant concern, as hosting locally ensures security and compliance with domestic regulations.