2Navigating Latin America’s AI competitiveness While some sectors are making progress on adopting AI, much more must be done to reap the benefits. This chapter examines the region’s progress across the Blueprint’s three layers, highlighting stand-out examples and using survey results and stakeholder interviews to identify gaps that must be closed if AI adoption is to be accelerated. Latin America has made some progress on the foundational components of AI competitiveness, including establishing data centres and high- performance computing infrastructure that harness sustainable energy sources, although the growing resource needs to power AI innovation could create new challenges. Connectivity is improving in the region, but it remains a challenge that threatens equitable access, particularly in rural and underserved areas. AI Investment is reaching Latin America through venture capitalists, development financiers and hyperscalers, but fragmented regulation continues to restrict scale and regional integration. Some organizations aim to monitor and increase visibility of this landscape. Globally, the IMF developed the AI Preparedness Index which analyses countries’ digital infrastructure, human capital and labour market policies, innovation and economic integration, and regulation and ethics.9 More regionally, the Latin American Index of Artificial Intelligence is an annual report led by Chile’s National Center for Artificial Intelligence (CENIA) that evaluates infrastructure, data and human talent as part of benchmarking AI progress across the region.10 By utilizing indices such as these, Latin American economies can identify AI competitiveness gaps and develop targeted actions. Build sustainable AI infrastructure If AI is to be scaled, there must be electricity to power it and water to cool it. Some countries in Latin America are well positioned in this regard, but many lack clean energy and robust national grids to distribute it. Argentina, Brazil, Chile and Paraguay have developed a clean-energy sector which can help support sustainable AI infrastructure. As of 2024, 88% of Brazil’s electricity comes from renewable sources,11 meaning data centres and supercomputers could largely run on clean electricity – if the country’s grid can connect the places where energy is generated with the locations chosen for digital hubs. However, many Latin American countries do not have the infrastructure or reserve capacity to supply this demand only with clean energy. In Colombia, droughts affect reliability of hydropower generation. In the Caribbean, many countries still rely on diesel or fuel oil. In addition, AI infrastructure requires water and land for solar panels, wind turbines and new energy plants. While data centres bring investment, it is important that the consequences do not create new inequity challenges for local communities. According to UNDP , even small data centres can require an annual volume of water equivalent to that used by approximately 300,000 people.12 Keeping up with these demands is a challenge in some Latin American countries, especially those prone to drought. This creates public concern. For example, in Querétaro, Mexico, there have been public discussions about the water use of data centres.13 Cross-border collaboration to help balance these issues is still incipient; while examples are emerging, latency issues could limit these efforts. Chile and the Dominican Republic have signed a memorandum with The Development Bank of Latin America and the Caribbean, formally establishing a feasibility study to create a regional network of high-performance computing centres designed to address their shared digital infrastructure gap.14 This network aims to empower the essential large-scale data processing capabilities required to scale AI utilization across the region.2.1 Building the foundations for intelligent economies Latin America in the Intelligent Age: A New Path for Growth 11

Latin America Intelligent Age