How the Global South is reimagining the future of AI

As India gears up to host the AI Impact Summit, one of the Global South’s largest AI convenings to date, it signals more than a diplomatic milestone – it marks a deeper reimagining of what AI can look like when shaped outside the traditional Western, big tech-dominated ecosystem.

Despite the digital divide that places much of the world at a disadvantage, countries across the Global South are charting their own pathways of AI development, grounded in cultural context, digital public infrastructure and inclusive innovation. Here are four trends that capture this shift.

Decolonizing AI begins with a fundamental question: Who gets to speak and who gets to be understood by machines? In the big tech-led AI ecosystem, linguistic representation is uneven, with the majority of large language models trained on English. This reinforces the existing power structures within the next generation of technology. A wave of language-first AI initiatives from the Global South is, however, bringing an alternative to the fore.

In Africa, for example, Lelapa AI has built InkubaLM, a small language model for five African languages, namely Swahili, Yoruba, IsiXhosa, Hausa and IsiZulu, which have over 350 million speakers. In Peru, Huqariq offers digital archives for the Quechua and Aymara languages, which are spoken by millions. Karya has built a similar voice data repository in Indian languages.

Beyond language, indigenous knowledge systems are beginning to inform how AI reasons and is governed. This challenges the long-standing assumption of Western epistemologies as the blueprint of intelligent systems. Innovators in the Global South are bringing to AI development insights from their philosophical traditions and community governance models.

In Africa, for instance, the Masakhane initiative embeds the cultural and historical context of Africa, including community-centred values such as Ubuntu, into AI development, ensuring that the models reflect the local ways of knowing. In India, startups such as Mokx are experimenting with Vedas-inspired AI models that embed principles from Indian knowledge systems into logical frameworks and ethical reasoning.

Countries in the Global South such as India and Brazil have played a pioneering role in building digital public infrastructure (DPI) rails, most notably identity systems (India’s Aadhaar) and real‑time payment networks (India’s UPI and Brazil’s PIX). As they mature, the integration of AI with DPI is increasingly viewed not as a convenient add‑on, but as a core pillar of national capability-building and digital AI sovereignty. DPI provides public datasets, transparent governance mechanisms, and shared technical standards that make AI systems more interoperable, trustworthy and locally relevant.

This foundation is now enabling new possibilities, such as public AI agents that can interact with citizens to streamline government workflows, for example, simplifying agricultural subsidy approvals. In India, AI is already being integrated with DPI to strengthen digital identity matching. The Open Healthcare Network, a UN‑recognized digital public good, has developed voice‑to‑text and language‑translation AI tools that help doctors and nurses generate electronic medical records without language barriers. Similarly, Hello! UPI now enables AI‑powered voice payments, making digital transactions more inclusive for users.

The geography of AI is skewed, and that is largely a function of compute and chip geopolitics, with an overwhelming majority concentrated in the US (75%) and China (15%). The Global South is beginning to rebalance this by investing in publicly governed compute infrastructure that reduces concentration and expands equitable access.

In Mexico, a major public investment into AI is the development of Coatlicue, a national supercomputer with a capacity of 314 petaflops and integrating a 14,480 GPU architecture. Similarly, the IndiaAI mission has already deployed a national “common compute” pool of more than 34,000 publicly funded GPUs. This common pool infrastructure democratizes access, supporting startups, researchers and public-sector innovators by lowering the cost of training advanced models and developing India‑specific AI.

Together, the Global South’s emerging AI ecosystems – rooted in linguistic diversity, indigenous epistemologies, digital public infrastructure and shared compute – are catalyzing AI-powered entrepreneurship shaped by local values, needs and lived realities. In India, Bharat Intelligence is using voice‑based AI to organize rural labour markets in regional languages. In Brazil, NuBank is leveraging the country’s public digital payment rail, PIX, to power AI‑enhanced risk processing and fraud detection.

While the trends signal a structural rebalancing of the AI economy, sustained investment in inclusive AI foundations remains essential to enable all regions, not just the historic tech hubs, to shape the AI future.