Figure 8-3 Data center distribution in 2024
TWh of electricity, accounting for 1.5% of total countries, 95% of STEM graduates in frontrun- global consumption. This is projected to rise to 945 TWh by 2030, representing approximately 1 fields, whereas this proportion is only 50% and 3% of total global consumption (IEA, 2024a). Among these, a single hyperscale AI data cen- tively (IDC et al., 2024). Insufficient digital in- (IEA, 2024b). However, the energy supply sys- tems in developing countries remain too weak graduates locally, resulting in low employ- ment conversion rates and severe brain drain. to support their computing power deployment. Statistics show that over 600 million people in Additionally, weak infrastructure and poor SSA still lack access to electricity, accounting for 60% of the region's total population. The re- gion's share of global energy investments is less solutions, further increasing operational and than 1% (IEA, 2024c). As global computing in-delivery costs and restricting the localization frastructure expands rapidly, insufficient energy nologies such as AI (Kamanzi, 2025; Satariano veloping countries? intelligent transformation. nologies in developing countries. Although In recent years, a handful of developed econ- the proportion of STEM (science, technology, omies, out of strategic competition consider- engineering, and mathematics) graduates in ations, have generalized security concepts and