The narrative that developing nations can simply 'leapfrog' into the AI era—bypassing traditional phases of development just as they skipped landlines for mobile phones—is currently being challenged by the very institutions meant to fund such progress. World Bank President Ajay Banga recently injected a sobering dose of reality into the conversation, questioning the feasibility of this trajectory for lower-income countries.

Unlike the mobile revolution, which thrived on low-cost adoption and lightweight infrastructure, modern artificial intelligence is structurally demanding. It requires massive, reliable computing power and significant electrical capacity, resources that remain concentrated among a handful of tech giants in the United States and China.

As the article notes, this concentration creates an uneven playing field. When nations attempt to build their own sovereign AI—essentially creating proprietary versions of large-scale models—they often face significant hurdles, including export restrictions on high-end semiconductors and reliance on foreign cloud services. Trying to compete in the foundational model space is, for most, a race they are structurally ill-equipped to win.

However, this does not mean the door to progress is closed. Instead of focusing on the 'infrastructure layer'—building the underlying compute and base models—the real leapfrog opportunity lies in the 'governance layer.' By investing in robust Digital Public Infrastructure (DPI) today, developing nations can set up systems without the 'technical debt' that legacy institutions in wealthy countries are forced to manage.

This approach allows governments to design interoperable identity, payment, and data systems from scratch, ensuring they are built for their specific social and political context. The goal shifts from trying to out-compute the hyperscalers to becoming architects of the rules that dictate how these tools function.

Ultimately, the window for this strategy is not defined by technology roadmaps, but by the governance calendar. As global bodies begin drafting the frameworks that will regulate AI, nations that show up with coherent, modern digital governance structures will have the leverage to shape these standards, rather than simply having them imposed from above.

The narrative that developing nations can simply 'leapfrog' into the AI era—bypassing traditional phases of development just as they skipped landlines for mobile phones—is currently being challenged by the very institutions meant to fund such progress. World Bank President Ajay Banga recently injected a sobering dose of reality into the conversation, questioning the feasibility of this trajectory for lower-income countries.

Unlike the mobile revolution, which thrived on low-cost adoption and lightweight infrastructure, modern artificial intelligence is structurally demanding. It requires massive, reliable computing power and significant electrical capacity, resources that remain concentrated among a handful of tech giants in the United States and China.

As the article notes, this concentration creates an uneven playing field. When nations attempt to build their own sovereign AI—essentially creating proprietary versions of large-scale models—they often face significant hurdles, including export restrictions on high-end semiconductors and reliance on foreign cloud services. Trying to compete in the foundational model space is, for most, a race they are structurally ill-equipped to win.

However, this does not mean the door to progress is closed. Instead of focusing on the 'infrastructure layer'—building the underlying compute and base models—the real leapfrog opportunity lies in the 'governance layer.' By investing in robust Digital Public Infrastructure (DPI) today, developing nations can set up systems without the 'technical debt' that legacy institutions in wealthy countries are forced to manage.

This approach allows governments to design interoperable identity, payment, and data systems from scratch, ensuring they are built for their specific social and political context. The goal shifts from trying to out-compute the hyperscalers to becoming architects of the rules that dictate how these tools function.

Ultimately, the window for this strategy is not defined by technology roadmaps, but by the governance calendar. As global bodies begin drafting the frameworks that will regulate AI, nations that show up with coherent, modern digital governance structures will have the leverage to shape these standards, rather than simply having them imposed from above.