Latin and Central America risk turning artificial intelligence into another source of economic inequality unless governments move beyond national strategies and businesses gain the computing capacity and skills needed to use it productively.
That is the challenge emerging from the Latin American Artificial Intelligence Index 2026, which finds that connectivity remains essential but is no longer enough to build competitive AI economies. The larger obstacles now involve computing infrastructure, business adoption and the ability to turn policy commitments into practical programmes.
The fourth index examines 20 regional economies. Its findings expose a widening challenge: access to AI does not automatically deliver productivity, innovation or better public services.
Brazil, Chile and Uruguay form the index’s pioneering group, each scoring above 60 points despite their different economic structures. Costa Rica joins Colombia and Argentina among ten adopters scoring between 35 and 60. Seven countries remain explorers, scoring below 35.
For Central America, Costa Rica’s position offers evidence that advanced capabilities need not depend exclusively on the region’s largest economies. However, the wider findings underline the importance of matching investment to national needs, productive strengths and institutional capacity.
The sharpest imbalance concerns computing power. Brazil accounts for 87.7% of the graphics processing units and 92% of the regional processing capacity covered by the study. That concentration highlights the constraints facing countries seeking to develop AI capabilities with much smaller resources.
Government ambition also runs ahead of implementation. While 85% of countries have an AI strategy in place or under development and 75% have formally designated a lead institution, only Brazil and Colombia explicitly allocate budgets within their national instruments.
Without funding, strategic commitments risk remaining statements of intent. The challenge is to connect infrastructure, data, research and public administration with measurable improvements in economic performance and people’s lives.
Businesses face a similarly uneven transition. AI adoption reaches 27.6% among large companies, compared with 12.8% among medium businesses and just 6.8% among small firms. Unless that gap narrows, productivity gains could become concentrated among companies already best equipped to capture them.
Education provides a stronger foundation, although skills shortages persist. All 20 countries offer postgraduate programmes, 65% have specialised research centres and more than half have incorporated AI into school curricula. These advances have yet to translate into sufficient specialist talent across labour markets.
Regional cooperation could help overcome limitations that individual countries cannot easily resolve. Shared computing resources, collaborative research, advanced training and coordinated participation in international standards offer ways to expand capabilities beyond national scale.
ECLAC identifies its regional digital agenda, eLAC, as a platform for that cooperation. For Latin and Central America, the priorities are: fund implementation, widen business adoption and ensure AI strengthens local capabilities.
The measure of success will be whether technology delivers broader economic opportunity and more effective services, rather than simply becoming another imported tool available mainly to those with the greatest resources.
