Ghana has the blueprint. The question now is whether it has the machinery to execute it. Three days after President John Dramani Mahama formally launched the National Artificial Intelligence (AI) Strategy 2025 to 2035 in Accra, attention is already shifting from the ambition of the document to the credibility of its delivery.

The strategy sets a headline target of GHC 500 billion in AI-driven economic value by 2035, roughly $45 billion at current exchange rates, for an economy that recorded a gross domestic product (GDP) of approximately $76 billion in 2023. It commits $270 million in government funding, proposes a Responsible AI Authority (RAA) to coordinate implementation across ministries, and targets one million AI-ready Ghanaians by 2033. On paper, the architecture is sound. In practice, the distance between design and delivery in emerging markets has historically been wide.

The Funding Gap
The government’s initial $270 million commitment, while significant, is modest against the strategy’s full ambition. Of that figure, $250 million is earmarked for a national AI computing centre, leaving $20 million for short-to-medium-term implementation across eight strategic pillars. The plan depends on attracting GHC 200 billion in combined local and foreign private investment by 2035. That scale of capital typically follows demonstrated execution, not policy announcements alone. Investors look for functioning institutions, successful pilot projects, and regulatory predictability before committing at that level.

Connectivity remains Ghana’s structural challenge. Urban centers approach near-universal 4G coverage, but rural penetration lags considerably, creating a divide that limits both data generation and service access. The strategy calls for fifth-generation (5G) rollout, rural broadband expansion, and distributed data infrastructure including solar-powered edge facilities. These are the right priorities. They are also capital-intensive and slow-moving. Across the continent, comparable infrastructure projects have repeatedly stalled at the intersection of procurement delays, financing gaps, and shifting political priorities.

The proposed Natural Language Processing (NLP) Centre of Excellence is one of the more strategically differentiated elements. Building large-scale datasets in Ghanaian languages could position the country as a regional hub for AI development, particularly as global technology companies seek more diverse training data. The strategy targets one trillion tokens of Ghanaian datasets by 2030. In a field where data quality increasingly defines competitive advantage, this is one area where late movers can still carve out meaningful ground.

Talent Without Retention Is a Subsidy for Others
Ghana’s universities produce capable graduates in engineering and computer science. Retaining them is the harder challenge. Skilled researchers and engineers are drawn to global hubs offering deeper salaries, computing resources, and research networks. The strategy proposes diaspora fellowships and repatriation incentives, echoing policies used in countries including India and Rwanda. Evidence suggests such programs can work, but only with sustained funding and long-term policy consistency across administrations.

The One Million Coders Programme targets training at least 300,000 Ghanaians in digital and AI skills within the current year. Over 100,000 applications have already been processed. The scale is commendable, but training infrastructure and quality assurance at that volume will require close monitoring to ensure output matches market demand.

Governance Is the Decisive Variable
The planned RAA is designed to coordinate across ministries, standardize implementation, and provide ethical oversight. Its practical authority will matter more than its formal mandate. Whether it is adequately funded, insulated from political turnover, and empowered to compel cooperation across agencies will determine whether the strategy’s eight pillars advance together or drift independently. Even in advanced economies, interagency coordination remains a persistent obstacle. In markets with more pronounced institutional fragmentation, it can become a defining constraint.

What Ghana has built is a credible starting point, not a guarantee. The country has identified the right problems and proposed solutions that reflect lessons from more mature AI ecosystems. It has also set targets that will require sustained discipline across multiple administrations and economic cycles. The next phase will unfold not in launch events but in budget negotiations, regulatory decisions, and the slow work of institutional alignment. That is where comparable strategies have most often faltered, and where Ghana’s will ultimately be judged.