Computing plan of $860 million a year: Egypt charts a regional inference path and avoids a costly training race

Egypt is putting forward an investment plan that aims to close the existing gap between the abundance of engineering talent and submarine cable landing sites on the one hand, and the lack of locally available advanced computing capabilities on the other. According to a document issued by the Ministry of Communications and Information Technology, the plan targets annual spending of between half a billion and 860 million dollars on data centres, computing infrastructure and artificial-intelligence applications, starting in fiscal year 2026/2027 and running through the end of the current decade. Although this figure represents roughly a tenth of a percentage point of gross domestic product, it nonetheless constitutes a financially significant decision given the fiscal tightening affecting the general budget.

The plan’s formulation through three ministries, the Ministry of Communications, the Ministry of Electricity and the Ministry of Investment, reveals the governing equation: data centres are primarily energy projects.The electricity grid, which underwent load-shedding during recent summers, forces the energy sector to be at the core of the planning, while the involvement of the investment ministry reflects the state’s recognition that it cannot fund these facilities on its own. The targeted spending is allocated to building new capacities, enhancing data security and privacy, and introducing artificial-intelligence tools in a selected group of state-owned companies, which is the less glamorous but more effective route in practice.

In parallel with government funding, the National Telecommunications Regulatory Authority issued ten data-centre licences over the past two years, including a first-phase licence for a project belonging to the Hassan Alam Holding Group valued at roughly 400 million dollars. This project serves as a concrete test of the market’s ability to generate sufficient local and regional demand to operate a privately funded facility. On the hardware side, Huawei submitted a proposal that includes about 1,400 Ascend chips for training tasks and 600 chips for model inference, in a package exceeding 2,000 processors sold as an integrated national computing block, akin to a power-generation plant.

The most prominent planning dilemma lies in the absence of a publicly announced national register that tracks the actual volume of AI computing available, its leasing costs and utilisation rates.Thus, building a strategy that aims to raise the technology sector’s contribution to about eight percent of GDP by 2030, without an accurate database of existing capacities, leaves spending decisions without a real measurement benchmark. Creating a periodic register of computing capabilities and their usage fees would spare the budget the cost of guesswork and precisely clarify what companies and researchers in the market need.

The facts show that the common comparison with Gulf spending paths misleads planning; the Gulf funds data-centres from sovereign budgets that do not demand a return on investment for five years, whereas Egypt pays for each processor in foreign currency and allocates every megawatt from a grid that faces other vital demands. Egypt’s realistic advantage lies in the intersection of international telecom cables between Europe and Asia, targeted renewable-energy projects, and the low cost of engineering talent. These factors make the market suitable for inference computing serving Africa and the Eastern Mediterranean, and for data-processing and model-preparation work, rather than for building massive complexes to train foundational models that consume budgets without parallel returns.

For those building in Egypt and the region, the practical impact lies in providing computing and inference capacity that developers and start-ups can rent in Egyptian pounds without draining their foreign-currency resources on external cloud platforms. The approach also opens a path to employ local engineering talent in adapting models, running them and preparing data instead of relying solely on imported ready-made solutions. The success of this shift will depend on the financing line appearing clearly in the 2026/2027 budget, on new projects’ ability to attract major tenants before launch, and on publishing a transparent inventory of the computing capabilities Egypt already possesses.