Criticism of Brazil's AI strategy: data centralization and supercomputers instead of edge computing and education
An analysis on Tech Policy Press criticizes Brazil's AI strategy: the government is investing 2.3 billion reais in supercomputers and data centralization with Huawei and iFlytek, while only 27% of primary schools have computer labs and technical education remains neglected.
An analytical piece on Tech Policy Press (author Yasodara Cordova) criticizes the direction of Brazil's AI strategy, which, according to the author, focuses on data centralization and supercomputers while neglecting edge computing and technical education. In August, the Brazilian government announced an investment of 2.3 billion reais ($444.2 million) in two AI projects — a supercomputing partnership with the Chinese companies Huawei and iFlytek, and a tender for a second supercomputer. This capacity is meant to power the centralized "Base de Dados Brasil" initiative, which, according to Esther Dweck, Minister for Management and Innovation in Public Services, the government is preparing in order to centralize, integrate, and protect citizens' data for training AI.
The author contrasts this investment with the actual situation in education: according to the 2024 school census, only 27% of public primary schools and 73% of public secondary schools have computer labs, while camera systems, according to a 2025 Cetic.br survey, are present in 79% of schools in state capitals. According to the Ministry of Labor, 84% of employed young Brazilians work in positions without specific qualifications, and only 1 in 7 obtains a technical or university-level position. Only 12% of high school students want to study engineering, partly due to weak confidence in mathematics (5.2 out of 10, according to a survey by CIEE and Instituto Locomotiva), and Brazil again scored below the OECD average in mathematics in PISA 2025.
The author argues that instead of chasing frontier models and hyperscale data centers, Brazil should invest in edge AI — local, low-power data processing for industrial, agricultural, and manufacturing applications — following the model of China, where the state links strategic direction with regional industrial clusters (the East Data, West Computing program). As an argument regarding jobs, she cites a July 2026 study by the Getulio Vargas Foundation (FGV), according to which, for a model 100 MW data center, 85% of jobs are in the construction phase and only about 4% are permanent operations and maintenance roles, with an estimated 15% permanent jobs overall when including spillover effects in other sectors.
Why it matters
The piece is a contribution to the policy debate on which direction a developing economy's national AI strategy should take — whether to invest in centralized infrastructure and frontier models, or in local edge data processing and technical education, which, according to the author, creates more lasting jobs. For readers outside Brazil, it is relevant as an example of the broader dispute over where to direct public AI investment and what balance between infrastructure and human capital actually generates economic benefit.
Check the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.