Kenya could be set for a major new investment in artificial intelligence infrastructure, with Greek firm Amaco Energy Group proposing a $1.5 billion (about KSh194 billion) offshore AI data centre in Mombasa. The project is still at the proposal stage, but its defining feature is clear: rather than relying entirely on Kenya’s national grid, the facility would generate much of its own electricity through an offshore LNG-powered platform. The concept is built around Amaco’s HERCULES platform, which combines power generation, cooling and data-centre infrastructure in a single offshore system.
The proposal comes as access to reliable electricity emerges as one of the biggest constraints on Africa’s ambitions to build hyperscale AI infrastructure. Large AI data centres consume enormous amounts of electricity, placing pressure on national grids that are often already struggling to meet existing demand. In Kenya, the proposed 100 MW Microsoft-G42 data centre at Olkaria has highlighted this challenge, with concerns over whether the grid can accommodate such a large additional load. Amaco’s model seeks to sidestep that problem by building dedicated generation alongside the computing infrastructure, allowing the data centre to operate with substantially less dependence on the national transmission network.
The HERCULES concept is centred on LNG-fired generation, although Amaco says its offshore smart-power platforms can incorporate renewable energy and eventually transition towards hydrogen. The company has broader ambitions to develop as much as 4 GW of offshore power capacity for AI data centres by 2030. If the Mombasa project progresses, it would therefore serve not only as a Kenyan investment but also as a potential demonstration of a different model for Africa’s digital infrastructure: instead of waiting for national grids to become capable of supporting hyperscale computing, developers could build dedicated energy systems specifically for data centres.
Read Also: Nigeria Revives 6,000MW Power Target as Gas and Grid Constraints Persist
The model, however, presents a significant trade-off. LNG could provide the reliability and energy security required by an AI facility, but using fossil-fuel generation would raise questions about emissions and its compatibility with Kenya’s climate ambitions. The project would also need to demonstrate that an offshore facility can generate meaningful local economic benefits through jobs, skills, supply chains and tax revenues. Before construction can begin, Amaco will still need to secure regulatory approvals, financing, technical partners and anchor customers. For Kenya, the proposal illustrates a broader reality of the AI race: the countries that can solve the power problem will have a major advantage in attracting the next generation of data centres.
