The Hidden Costs of the Generative AI Boom: How Overinvestment Exacerbates the Digital Divide in Developing Countries

Abstract

Between 2024 and 2025, a global investment frenzy in generative artificial intelligence (AI) has emerged, with governments and enterprises worldwide competing to deploy large language models, AI chips, and intelligent applications. However, in developing countries with limited resources, such a “technological leapfrogging” strategy may entail severe structural risks. This paper argues that blind pursuit of cutting-edge AI not only diverts investment from critical basic digital infrastructure (including broadband networks, power supply, and digital literacy education) but also risks widening the digital divide across urban-rural, regional, and social groups. By analyzing policy practices and practical dilemmas in India, Vietnam, selected African nations, and central and western regions of China, this study reveals that the lack of a technology strategy aligned with developmental stages can easily lead to a dual predicament: “underutilization of advanced technologies” coexisting with “collapse of basic capabilities”. The paper advocates for a “stratified AI strategy”, prioritizing the consolidation of digital foundations before selectively developing lightweight or high social-return AI applications, to achieve inclusive technological progress.